HSPA t o LTE- Advanced, Rysavy Research / 3G Americas, Sept ember 2009 Page 2

TABLE OF CONTENTS
I NTRODUCTI ON ...................................................................................................... 4  
BROADBAND DEVELOPMENTS ................................................................................ 6 
Wireless versus Wireline . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 7 
Bandwidt h Management . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 8 
WI RELESS DATA MARKET ..................................................................................... 10  
Trends . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 10 
EDGE/ HSPA/ HSPA+ Deployment . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 13 
St at ist ics . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 13 
WI RELESS TECHNOLOGY EVOLUTI ON AND MI GRATI ON ....................................... 14 
1G t o 4G . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 14 
3GPP Evolut ionary Approach . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 16 
Spect rum . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 19 
Core- Net work Evolut ion . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 22 
Service Evolut ion . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 23 
Device I nnovat ion . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 24 
Net work I nt erfaces for Applicat ions . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 25 
Mobile Applicat ion Archit ect ures . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 26 
Broadband- Wireless Deployment Considerat ions . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 27 
Feat ure and Net work Roadmap . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 28 
Deployment Scenarios . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 32 
COMPETI NG TECHNOLOGI ES ................................................................................ 33  
CDMA2000 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 33 
WiMAX . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 35 
I EEE 802. 20 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 39 
Wi- Fi and Municipal Wi- Fi Syst ems . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 39 
COMPARI SON OF WI RELESS TECHNOLOGI ES ....................................................... 40  
Dat a Throughput . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 40 
HSDPA Throughput in Represent at ive Scenarios . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 43 
Release 99 and HSUPA Uplink Performance . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 45 
LTE Throughput . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 46 
Lat ency . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 49 
Spect ral Efficiency . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 50 
Cost , Volume and Market Comparison . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 58 
Compet it ive Summary . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 60 

HSPA t o LTE- Advanced, Rysavy Research / 3G Americas, Sept ember 2009 Page 3
CONCLUSI ON ........................................................................................................ 62  
APPENDI X: TECHNOLOGY DETAI LS ...................................................................... 64 
EDGE/ EGPRS . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 65 
Evolved EDGE . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 67 
UMTS- HSPA Technology . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 75 
UMTS Release 99 Dat a Capabilit ies . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 77 
HSDPA . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 77 
HSUPA . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 81 
Evolut ion of HSPA ( HSPA+ ) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 82 
HSPA Voice Support . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 90 
3GPP LTE . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 92 
4G, I MT- Advanced and LTE- Advanced . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 97 
UMTS TDD . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 101 
TD- SCDMA . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 102 
I MS . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 102 
Broadcast / Mult icast Services . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 103 
EPC/ SAE. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 104 
Whit e Space . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 106 
ACRONYMS ......................................................................................................... 108  
ADDI TI ONAL I NFORMATI ON .............................................................................. 112 
REFERENCES ...................................................................................................... 112  

HSPA t o LTE- Advanced, Rysavy Research / 3G Americas, Sept ember 2009 Page 4
Introduction
We are at t he dawn of a new decade t hat will bring t o mass market t he mobile broadband
innovat ions int roduced over t he last several years. 3G t echnology has shown us t he power
and pot ent ial of always- on, everyplace net work connect ivit y and has ignit ed a massive wave
of indust ry innovat ion t hat spans devices, applicat ions, I nt ernet int egrat ion, and new
business models. Already used by hundreds of millions of people, mobile broadband
connect ivit y is on t he verge of becoming ubiquit ous. I t will do so on a powerful foundat ion
of net working t echnologies, including GSM wit h EDGE, HSPA, and LTE. LTE in a fort hcoming
release will be one of t he first t echnologies t o meet t he requirement s of I nt ernat ional Mobile
Telephone ( I MT) Advanced, a proj ect of t he I nt ernat ional Telecommunicat ions Union ( I TU)
t hat t his year defined official “ 4G” requirement s.
Through const ant innovat ion, Universal Mobile Telecommunicat ions Syst em ( UMTS) wit h
High Speed Packet Access ( HSPA) t echnology has est ablished it self as t he global, mobile-
broadband solut ion. Building on t he phenomenal success of Global Syst em for Mobile
Communicat ions ( GSM) , t he GSM- HSPA ecosyst em has become t he most successful
communicat ions t echnology family ever. Through a process of const ant improvement , t he
GSM family of t echnologies has not only mat ched or exceeded t he capabilit ies of all
compet ing approaches, but has significant ly ext ended t he life of each of it s member
t echnologies.
UMTS- HSPA, in part icular, has many key t echnical and business advant ages over ot her
mobile wireless t echnologies. Operat ors worldwide are now deploying bot h High Speed
Downlink Packet Access ( HSDPA) and High Speed Uplink Packet Access ( HSUPA) , t he
combinat ion of t he t wo t echnologies called simply HSPA. HSPA is t he most capable cellular-
dat a t echnology ever developed and deployed. HSPA, already widely available, follows t he
successful deployment of UMTS net works around t he world and is now a st andard feat ure.
HSPA is st rongly posit ioned t o be t he dominant mobile- dat a t echnology for t he next five t o
t en years. To leverage operat or invest ment s in HSPA, t he 3GPP ( Third Generat ion
Part nership Proj ect ) st andards body has developed a series of enhancement s t o creat e
“ HSPA Evolut ion, ” also referred t o as “ HSPA+ . ” HSPA Evolut ion represent s a logical
development of t he Wideband Code Division Mult iple Access ( WCDMA) approach, and it is
t he st epping st one t o an ent irely new Third Generat ion Part nership Proj ect ( 3GPP) radio
plat form called 3GPP Long Term Evolut ion ( LTE) . LTE, which uses Ort hogonal Frequency
Division Mult iple Access ( OFDMA) , will be ready for deployment in 2010. Simult aneously,
3GPP — recognizing t he significant worldwide invest ment s in GSM net works—has defined
enhancement s t hat will significant ly increase EDGE dat a capabilit ies t hrough an effort called
Evolved EDGE.
Combined wit h t hese improvement s in radio- access t echnology, 3GPP has also spearheaded
t he development of maj or core- net work archit ect ure enhancement s such as t he I P
Mult imedia Subsyst em ( I MS) and t he Evolved Packet Core ( EPC) , previously called Syst em
Archit ect ure Evolut ion or SAE, as well as development s in Fixed Mobile Convergence ( FMC) .
These development s will facilit at e new t ypes of services, t he int egrat ion of legacy and new
net works, t he convergence bet ween fixed and wireless syst ems, and t he t ransit ion from
circuit - swit ched approaches for voice t raffic t o a fully packet - swit ched model.
The result is a balanced port folio of complement ary t echnologies t hat covers bot h radio-
access and core net works, provides operat ors maximum flexibilit y in how t hey enhance t heir
net works over t ime, and support s bot h voice and dat a services.
This paper discusses t he evolut ion of EDGE, HSPA enhancement s, 3GPP LTE, t he capabilit ies
of t hese t echnologies, and t heir posit ion relat ive t o ot her primary compet ing t echnologies. I t

HSPA t o LTE- Advanced, Rysavy Research / 3G Americas, Sept ember 2009 Page 5
explains how t hese t echnologies fit int o t he I TU roadmap t hat leads t o I MT- Advanced. The
following are some of t he import ant observat ions and conclusions of t his paper:
 The wireless t echnology roadmap now ext ends t o I MT- Advanced wit h LTE- Advanced
being one of t he first t echnologies defined t o meet I MT- Advanced requirement s. LTE-
Advanced will be capable of peak t hroughput rat es t hat exceed 1 gigabit per second
( Gbps) .
 Persist ent innovat ion creat ed EDGE, which was a significant advance over GPRS;
HSPA and HSPA+ , which are bringing UMTS t o it s full pot ent ial; and is now delivering
LTE, t he most powerful, wide- area wireless t echnology ever developed.
 GSM- HSPA
1
has an overwhelming global posit ion in t erms of subscribers,
deployment , and services. I t s success will cont inue t o marginalize ot her wide- area
wireless t echnologies.
 I n current deployment s, HSPA users regularly experience t hroughput rat es well in
excess of 1 megabit per second ( Mbps) under favorable condit ions, on bot h
downlinks and uplinks, wit h 4 Mbps downlink speed commonly being measured.
Planned enhancement s such as dual- carrier operat ion will double peak user-
achievable t hroughput rat es.
 HSPA Evolut ion provides a st rat egic performance roadmap advant age for incumbent
GSM- HSPA operat ors. Feat ures such as dual- carrier operat ion, MI MO, and higher-
order modulat ion offer operat ors mult iple opt ions for upgrading t heir net works, wit h
many of t hese feat ures ( e. g. , dual- carrier, higher- order modulat ion) being available
as net work soft ware upgrades.
 HSPA+ wit h 2x2 MI MO, successive int erference cancellat ion, and 64 Quadrat ure
Amplit ude Modulat ion ( QAM) is more spect rally efficient t han compet ing t echnologies
including Worldwide I nt eroperabilit y for Microwave Access ( WiMAX) Release 1. 0.
 The LTE Radio Access Net work t echnical specificat ion was approved in 2008 for 3GPP
Release 8, which was fully rat ified in March, 2009. I nit ial deployment s will occur in
2010 and will expand rapidly t hereaft er.
 The 3GPP OFDMA approach used in LTE mat ches or exceeds t he capabilit ies of any
ot her OFDMA syst em. Peak t heoret ical downlink rat es are 326 Mbps in a 20 MHz
channel bandwidt h. LTE assumes a full I nt ernet Prot ocol ( I P) net work archit ect ure,
and it is designed t o support voice in t he packet domain.
 LTE has become t he t echnology plat form of choice as GSM- UMTS and CDMA/ EV- DO
operat ors are making st rat egic, long- t erm decisions on t heir next - generat ion
plat forms. I n June of 2008, aft er ext ensive evaluat ion, LTE was t he first and t hus far
only t echnology recognized by t he Next Generat ion Mobile Net work alliance t o meet
it s broad requirement s.
 GSM- HSPA will comprise t he overwhelming maj orit y of subscribers over t he next five
t o t en years, even as new wireless t echnologies are adopt ed. The deployment of LTE
and it s coexist ence wit h UMTS- HSPA will be analogous t o t he deployment of UMTS-
HSPA and it s coexist ence wit h GSM.
 3GPP has made significant progress on how t o enhance LTE t o meet t he
requirement s of I MT- Advanced in a proj ect called LTE- Advanced. LTE- Advanced is
expect ed t o be t he first t rue “ 4G” syst em available.

1
This paper ’s use of t he t erm “ GSM- HSPA” includes GSM, EDGE, UMTS, HSPA and HSPA+ .
“ UMTS- HSPA” refers t o UMTS t echnology deployed in conj unct ion wit h HSPA capabilit y.

HSPA t o LTE- Advanced, Rysavy Research / 3G Americas, Sept ember 2009 Page 6
 HSPA- LTE has significant economic advant ages over ot her wireless t echnologies.
 WiMAX has developed an ecosyst em support ed by many companies, but it will st ill
only represent a very small percent age of wireless subscribers over t he next five
years.
 EDGE t echnology has proven ext remely successful and is widely deployed on GSM
net works globally. Advanced capabilit ies wit h Evolved EDGE can double and
event ually quadruple current EDGE t hroughput rat es, halve lat ency and increase
spect ral efficiency.
 Wit h a UMTS mult i- radio net work, a common core net work can efficient ly support
GSM, WCDMA, and HSPA access net works and offer high efficiency for bot h high and
low dat a rat es, as well as for bot h high- and low- t raffic densit y configurat ions. I n t he
fut ure, EPC/ SAE will provide a new core net work t hat support s bot h LTE and
int eroperabilit y wit h legacy GSM- UMTS radio- access net works.
 I nnovat ions such as EPC/ SAE and UMTS one- t unnel archit ect ure will “ flat t en” t he
net work, simplifying deployment and reducing lat ency.
 Circuit - swit ched, voice- over HSPA, t hen moving t o voice over I nt ernet Prot ocol
( VoI P) over HSPA will add t o voice capacit y and reduce infrast ruct ure cost s. I n t he
meant ime, UMTS- HSPA enj oys high circuit - swit ched voice spect ral efficiency, and it
can combine voice and dat a on t he same radio channel.
This paper begins wit h an overview of t he market , looking at t rends, EDGE and UMTS- HSPA
deployment s, and market st at ist ics. I t t hen examines t he evolut ion of wireless t echnology,
part icularly 3GPP t echnologies, including spect rum considerat ions, core- net work evolut ion,
broadband- wireless deployment considerat ions, and a feat ure and net work roadmap. Next ,
t he paper discusses ot her wireless t echnologies, including Code Division Mult iple Access
2000 ( CDMA2000) and WiMAX. Finally, it compares t he different wireless t echnologies
t echnically, based on feat ures such as performance and spect ral efficiency.
The appendix explains in det ail t he capabilit ies and workings of t he different t echnologies
including EDGE, Evolved EDGE, WCDMA
2
, HSPA, HSPA Evolut ion ( HSPA+ ) , LTE, LTE-
Advanced, I MS, and SAE.
Broadband Developments
As wireless t echnology represent s an increasing port ion of t he global communicat ions
infrast ruct ure, it is import ant t o underst and overall broadband t rends and t he role bet ween
wireless and wireline t echnologies, as well as I nt ernet t rends. Somet imes wireless and
wireline t echnologies compet e wit h each ot her, but in most inst ances, t hey are
complement ary. For t he most part , backhaul t ransport and core infrast ruct ure for wireless
net works are based on wireline approaches, whet her opt ical or copper. This applies as
readily t o Wi- Fi net works as it does t o cellular net works.
Trends show explosive bandwidt h growt h of t he I nt ernet at large and for mobile broadband
net works in part icular. Cisco proj ect s global I P t raffic as nearly doubling every t wo years

2
Alt hough many use t he t erms “ UMTS” and “ WCDMA” int erchangeably, in t his paper we use “ WCDMA”
when referring t o t he r adio int erface t echnology used wit hin UMTS and “ UMTS” t o refer t o t he
complet e syst em. HSPA is an enhancement t o WCDMA.

HSPA t o LTE- Advanced, Rysavy Research / 3G Americas, Sept ember 2009 Page 7
t hrough 2012
3
, and mobile broadband t raffic growing at a CAGR of 131 percent bet ween
2008 and 2013, reaching 2 exabyt es
4
per mont h by 2013.
5

Wit h declining voice revenue, but increasing dat a revenue, cellular operat ors face a
t remendous opport unit y t o develop a mobile broadband business. Successful execut ion,
however, means more t han j ust providing high speed net works. I t also means nurt uring an
applicat ion ecosyst em, providing complement ary services, and supplying at t ract ive devices.
These are all areas in which t he indust ry has done well. An emerging challenge, however, is
managing bandwidt h, which will require a number of different approaches.
Wireless versus Wireline
Wireless t echnology is playing a profound role in net working and communicat ions, even
t hough wireline t echnology, such as fiber links, has inherent capacit y advant ages.
The overwhelming global success of mobile t elephony, and now t he growing adopt ion of
mobile dat a, conclusively demonst rat es t he desire for mobile- orient ed communicat ions.
Mobile broadband combines compelling high- speed dat a services wit h mobilit y. Thus, t he
opport unit ies are limit less when considering t he many diverse market s mobile
broadband can successfully address. Developed count ries cont inue t o show t remendous
upt ake of mobile broadband services. Addit ionally, in developing count ries, t here is no
doubt t hat 3G t echnology will cat er t o bot h ent erprises and t heir high- end mobile
workers and consumers, for whom 3G can be a cost - effect ive opt ion, compet ing wit h
digit al subscriber line ( DSL) for home use.
Relat ive t o wireless net works, wireline net works have always had great er capacit y, and
hist orically have delivered fast er t hroughput rat es. Figure 1 shows advances in t ypical
user t hroughput rat es wit h a consist ent 10x advant age of wireline t echnologies over
wireless t echnologies.

3
Source: Cisco, “ Approaching t he Zet t abyt e Era, ” June 16, 2008
4
One gigabyt e is 10
9
byt es. 1 t erabyt e is 10
12
byt es. 1 exabyt e is 10
15
byt es. 1 zet t abyt e is 10
18

byt es.
5
Source: Cisco Visual Net working I ndex: Global Mobile Dat a Traffic, Forecast Updat e, January 29,
2009

HSPA t o LTE- Advanced, Rysavy Research / 3G Americas, Sept ember 2009 Page 8
Fi gur e 1: Wi r el i ne and Wi r el ess Advances


The quest ion is whet her some of t he limit at ions of wireless t echnology, relat ive t o
wireline t echnology, may limit it s appeal and usage.
Bandwidth Management
Alt hough it is t rue t hat most 3G syst ems are now offering t hroughput s of about 1
Mbps—which is comparable t o what many users experience wit h a basic DSL or cable-
modem service—t he overall capacit y of wireless syst ems is generally lower t han it is
wit h wireline syst ems. This is especially t rue when wireless is compared t o opt ical fiber,
which some operat ors in developed count ries ( such as t he U. S. ) are now deploying t o
people’s homes. Wit h wireline operat ors looking t o provide 50 t o 100 Mbps t o eit her
people’s homes or businesses via next - generat ion cable- modem services, very high-
speed DSL ( VDSL) or fiber—especially for services such as high- definit ion I P Television
( I PTV) —t he quest ion becomes, “ I s it possible t o mat ch t hese rat es using wireless
approaches?” The answer is yes from a purely t echnical perspect ive, but it is no from a
pract ical point of view. I t is only possible t o achieve t hese rat es by using large amount s
of spect rum, generally more t han is available for current 3G syst ems, and by using
relat ively small cell sizes. Ot herwise, it simply will not be possible t o deliver t he
hundreds of gigabyt es per mont h t hat users will event ually be consuming over t heir
broadband connect ions wit h wide- area wireless net works. Consider t oday’s high
definit ion ( HD) t elevision cont ent t hat demands 6 t o 9 Mbps of cont inuous connect ivit y,
wherein one subscriber could essent ially consume t he ent ire capacit y of a WiMAX or
HSPA cell sect or.
Thus, operat ors are bot h deploying and considering mult iple approaches for managing
bandwidt h. These include:
2010 2000 2005
100 kbps
10 kbps
1 Mbps
10 Mbps
100 Mbps
GPRS 40 kbps
UMTS 350 kbps
HSDPA 1 Mbps
HSPA+ 5 Mbps
LTE 10 Mbps
EDGE 100 kbps
ADSL 1 Mbps
ISDN
128 kbps
ADSL 3 to 5 Mbps
ADSL2+ 25 Mbps
FTTH 100 Mbps

HSPA t o LTE- Advanced, Rysavy Research / 3G Americas, Sept ember 2009 Page 9
- Mor e spect r um. Spect rum correlat es direct ly t o capacit y, and more spect rum is
becoming available globally for mobile broadband.
- I ncr eased spect r al ef f i ci ency . Newer t echnologies are spect rally more
efficient , meaning great er t hroughput in t he same amount of spect rum.
- Mor e cel l si t es. Smaller cell sizes result in more capacit y per subscriber.
- Femt ocel l s. Femt o cells can significant ly offload t he macro net work. Pricing
plans can encourage users t o move high- bandwidt h act ivit ies ( e. g. , movie
downloads t o femt ocell connect ions.
- Wi - Fi . Wi- Fi net works offer anot her means of offloading heavy t raffic.
- Of f - peak hour s. Operat ors can offer lower rat es or perhaps fewer rest rict ions on
large dat a t ransfers t hat occur at off- peak hours such as overnight .
- Qual i t y of ser v i ce. By priorit izing t raffic, large downloads can occur wit h lower
priorit y, t hus not affect ing ot her act ive users.
I t will t ake a creat ive blend of all of t he above as well as ot her measures t o make t he
mobile broadband market successful and t o enable it t o exist as a complement ary
solut ion t o wired broadband.
Table 1 summarizes t he st rengt hs and weaknesses of wireless versus wireline broadband
approaches.
Tabl e 1: St r engt hs and Weak ness of Br oadband Appr oaches
St r engt h Weak ness
Mobi l e br oadband
( EDGE, HSPA, LTE)
Const ant connect ivit y
Broadband capabilit y
across ext remely wide
areas
Good access solut ion for
areas lacking wireline
infrast ruct ure
Capacit y enhancement via
FMC
Excellent voice
communicat ions
Lower capacit y t han
wireline approaches
I nabilit y t o serve high-
bandwidt h applicat ions
such as I P TV
Wi r el i ne br oadband
( e.g., DSL, DOCSI S,
FTTH)
High- capacit y broadband
at very high dat a rat es
Evolut ion t o ext remely
high t hroughput rat es
Expensive t o deploy new
net works, especially in
developing economies
lacking infrast ruct ure

3GPP t echnologies clearly address proven market needs; hence t heir overwhelming success.
The 3GPP roadmap, which ant icipat es cont inual performance and capacit y improvement s,
provides t he t echnical means t o deliver on proven business models. As t he applicat ions for
mobile broadband cont inue t o expand, HSPA, HSPA+ , LTE and LTE- Advanced will cont inue
t o provide a compet it ive plat form for t omorrow’s new business opport unit ies.

HSPA t o LTE- Advanced, Rysavy Research / 3G Americas, Sept ember 2009 Page 10


Wireless Data Market
By May 2009, more t han 3. 7 billion subscribers were using GSM- HSPA
6
—approaching an
ast onishing 50 percent of t he world’s t ot al 6. 8 billion populat ion.
7
By t he end of 2013, t he
global 3G wireless market is expect ed t o include more t han 2 billion subscribers, of which
1. 6 billion will use 3GPP t echnologies, represent ing 80% market share.
8
I n 2007, 3G
Americas President Chris Pearson st at ed, “ This level of wireless t echnology growt h exceeds
t hat of almost all ot her lifest yle- changing innovat ions.”
9
This growt h cont inues. Clearly,
GSM- HSPA has est ablished global dominance. Alt hough voice st ill const it ut es most cellular
t raffic, wireless dat a worldwide now comprises a significant percent age of revenue per user
( ARPU) . I n t he Unit ed St at es, wireless dat a is now more t han 26 percent of ARPU, and is
proj ect ed t o hit 30% by t he end of 2009.
10

This sect ion examines t rends and deployment , and t hen provides market dat a t hat
demonst rat es t he rapid growt h of wireless dat a.
Trends
As st at ed in a Rysavy Research report for CTI A on mobile broadband spect rum demand,
” We are at a unique and pivot al t ime in hist ory, in which t echnology capabilit y,
consumer awareness and comfort wit h emerging wireless t echnology, and indust ry
innovat ion are converging t o creat e mass- market accept ance of mobile broadband. ”
11

The market fact ors cont ribut ing t o t he surging growt h in t his market are shown in t he
following figure.






6
Source: I nforma Telecoms & Media, May 2009.
7
Source: US Census Bureau, ht t p: / / www. census. gov/ ipc/ www/ idb/ worldpopinfo. ht ml
8
Source: Subscriber Dat a - I nforma Telecoms and Media, World Cellular I nformat ion Service, March
2009
9
Source: 3G Americas press release of June 5, 2007.
10
Chet an Sharma, US Wireless Dat a Market Updat e - Q1 2009.
11
Source: Rysavy Research, “ Mobile Broadband Spect rum Demand, ” December 2008.

HSPA t o LTE- Advanced, Rysavy Research / 3G Americas, Sept ember 2009 Page 11

Fi gur e 2: Mar k et Fact or s Cont r i but i ng t o Gr ow t h of Mobi l e Br oadband

As dat a const it ut es a rising percent age of t ot al cellular t raffic, it is essent ial t hat
operat ors deploy spect rally efficient dat a t echnologies t hat meet cust omer requirement s
for performance—especially because dat a applicat ions can demand significant ly more
net work resources t han t radit ional voice services. Operat ors have a huge invest ment in
spect rum and in t heir net works; dat a services must leverage t hese invest ment s. I t is
only a mat t er of t ime before t oday’s more t han 4 billion cellular cust omers st art t aking
full advant age of dat a capabilit ies. This adopt ion will offer t remendous opport unit ies and
t he associat ed risks t o operat ors as t hey choose t he most commercially viable
evolut ionary pat h for migrat ing t heir cust omers. The EDGE/ HSPA/ LTE evolut ionary pat hs
provide dat a capabilit ies t hat address market needs and deliver ever- higher dat a
t hroughput s, lower lat ency, and increased spect ral efficiency.
As a consequence, t his rich net work and device environment is spawning t he availabilit y
of a wide range of wireless applicat ions and cont ent . Because of it s growing size—and it s
unassailable pot ent ial—applicat ion and cont ent developers are making t he wireless
market a high priorit y. For example, t here are now more t han 50, 000 applicat ions for
t he Apple iPhone.
12

Based on one leading UMTS- HSPA infrast ruct ure vendor’s st at ist ics, Figure 3 compares
t he rapid growt h in wireless dat a t raffic compared t o voice t raffic across mult iple
operat ors. By mid 2009, in HSPA coverage areas worldwide, t he volume of dat a t raffic
significant ly exceeded voice t raffic. Operat ors t hat are t he most aggressive wit h mobile
broadband services are experiencing dat a growt h rat es even higher t han t hese average
values. Traffic has cont inued t o increase since.

12
Source: ht t p: / / www. apple. com/ pr/ library/ 2009/ 06/ 22iphone. ht ml .

HSPA t o LTE- Advanced, Rysavy Research / 3G Americas, Sept ember 2009 Page 12
Fi gur e 3: UMTS- HSPA Voi ce and Dat a Tr af f i c
13


Over t ime, dat a demands are expect ed t o grow significant ly. Figure 4 shows a proj ect ion
by Chet an Sharma of mobile dat a growt h in t he US t hrough 2014.
Fi gur e 4: Mobi l e Dat a Gr ow t h i n t he Uni t ed St at es
14



13
Based on leading UMTS- HSPA infrast ruct ure vendor st at ist ics.
14
Source: “ Managing Growt h and Profit s in t he Yot t abyt e Era” , Chet an Sharma, July 2009. One
Terabyt e is 1000 gigabyt es.
1
5
11
3
7
9
13
15
17
19
65
95
25
55
85
5
45
75
05
35
Jan
07
Mar
07
May
07
Jul
07
Sep
07
Nov
07
Jan
08
Mar
08
May
08
July
08
Sep
08
Nov
08
Jan
09
Mar
09
May
09
Relative Network Load
Packet data
Voice
~ 2x
~ 18x

HSPA t o LTE- Advanced, Rysavy Research / 3G Americas, Sept ember 2009 Page 13

The key for operat ors is enhancing t heir net works t o support t he demands of consumer
and business applicat ions as t hey grow, along wit h offering complement ary capabilit ies
such as I P- based mult imedia. This is where t he GSM family of wireless- dat a t echnologies
is t he undisput ed leader. Not only does it provide a plat form for cont inual improvement s
in capabilit ies, but it does so over huge coverage areas and on a global basis.
EDGE/HSPA/HSPA+ Deployment
Three quart ers of GSM net works t oday support EDGE, represent ing more t han 345
net works in approximat ely 170 count ries.
15

Because of t he very low increment al cost of including EDGE capabilit y in GSM net work
deployment s, virt ually all new GSM infrast ruct ure deployment s are also EDGE- capable
and nearly all new mid- t o high- level GSM devices include EDGE radio t echnology.
Meanwhile, UMTS has est ablished it self globally. Nearly all WCDMA handset s are also
GSM handset s, so WCDMA users can access t he wide base of GSM net works and
services. There are more t han 336 million UMTS- HSPA cust omers worldwide spanning
283 commercial net works. Two hundred and sixt y four operat ors in 114 count ries offer
HSDPA, and 77 of t hese have HSUPA deployed.
16
Almost all UMTS operat ors are
deploying HSPA for t wo reasons: first , t he increment al cost of HSPA is relat ively low and
second, HSPA makes such efficient use of spect rum for dat a t hat it result s in a much
lower overall cost per megabyt e of dat a delivered. Already, t here are more t han 1375
commercial HSPA devices available worldwide from 135 suppliers.
17
Devices include
handset s, dat a cards, modems, rout ers, lapt ops, media players and cameras.
Operat ors have begun deploying evolved HSPA feat ures and HSPA+ launches include:
Telst ra ( Aust ralia) , Mobilkom ( Aust ria) , CSL Limit ed ( Hong Kong) , St arhub ( Singapore) .
As t he t echnology mat ures, upgrading t o HSPA+ will likely represent a minimal
invest ment for operat ors in order t o significant ly boost net work performance.
Statistics
A variet y of st at ist ics show t he rapid growt h in wireless dat a. Chet an Sharma report ed
t hat in Q1 2009, t he US wireless dat a market grew 32% over Q1 of 2008 t o reach $10B
in mobile dat a revenues, t he first t ime t he US market has crossed t he $10B milest one.
He also st at es t hat 62% of US subscribers were using some form of dat a service.
18

Berg I nsight report ed t hat in t he European Union, 11. 6% of broadband links at t he end
of 2008 were based on HSPA in bot h mobile and home environment s.
19
Pyramid
Research proj ect s t he number of European mobile broadband users t o reach 117 million
in 2014, up from 24 million in 2008.
20


15
Source: “ World Cellular I nformat ion Service, ” I nforma Telecoms & Media, June 2009.
16
I bid.
17
Source: GSMA.
18
Source: Chet an Sharma, US Wireless Dat a Market Updat e - Q1 2009.
19
Source: Berg I nsight , ht t p: / / www. berginsight . com/ News. aspx?m_m= 6&s_m= 1.
20
Source: Pyramid Resear ch, “ Europe t o See Huge Growt h in Mobile Broadband Services despit e
Recession, ” 2009.

HSPA t o LTE- Advanced, Rysavy Research / 3G Americas, Sept ember 2009 Page 14
Though most mobile broadband growt h t oday is based on HSPA ( wit h some EV- DO) , LTE
should see relat ively rapid adopt ion as it becomes deployed st art ing in 2010. Pyramid
Research expect s LTE net works t o grow more quickly t han prior 3G net works, reaching
100 million subscribers in j ust four years from init ial 2010 deployment s.
21
According t o
Juniper Research, t here are already in excess of 30 net work operat or commit ment s t o
LTE.
22
According t o 3G Americas ( www. 3gamericas. org) , t here are more t han 100
operat ors t hat have commit t ed or expressed int ent ions t o commit t o LTE.
From a device perspect ive, I nforma WCI S proj ect ed in June 2009 t he following sales
growt h rat e for WCDMA handset s:
23

2009: 307 million
2010: 416 million
2011: 564 million
2012: 736 million
2013: 927 million
I t is clear t hat bot h EDGE and UMTS/ HSDPA are dominant wireless t echnologies. And
powerful dat a capabilit ies and global presence mean t hese t echnologies will likely
cont inue t o capt ure most of t he available wireless- dat a market .
Wireless Technology Evolution and Migration
This sect ion discusses 1G t o 4G designat ions, t he evolut ion and migrat ion of wireless- dat a
t echnologies from EDGE t o LTE, as well as t he evolut ion of underlying wireless approaches.
Progress in 3GPP has occurred in mult iple phases, first wit h EDGE, and t hen UMTS, followed
by t oday’s enhanced 3G capabilit ies such as HSPA, HSPA+ and now, LTE, which it self is
evolving t o LTE- Advanced. Meanwhile, underlying approaches have evolved from Time
Division Mult iple Access ( TDMA) t o CDMA, and now from CDMA t o OFDMA, which is t he
basis of LTE.

1G to 4G
There is some confusion in t he indust ry as t o what t echnology falls int o which cellular
generat ion. 1G refers t o analog cellular t echnologies and became available in t he 1980s.
2G denot es init ial digit al syst ems, int roducing services such as short messaging and
lower speed dat a. CDMA2000 1xRTT and GSM are t he primary 2G t echnologies, alt hough
CDMA2000 1xRTT is somet imes called a 3G t echnology because it meet s t he 144 kbps
mobile t hroughput requirement . EDGE, however, also meet s t his requirement . 2G
t echnologies became available in t he 1990s.
3G requirement s were specified by t he I TU as part of t he I nt ernat ional Mobile Telephone
2000 ( I MT- 2000) proj ect , for which digit al net works had t o provide 144 kbps of
t hroughput at mobile speeds, 384 kbps at pedest rian speeds, and 2 Mbps in indoor

21
Source: Global Telecom I nsider repor t , "LTE’s Five- Year Global Forecast : Poised t o Grow Fast er t han
3G, " 2009.
22
Source: Juniper Research, LTE Report , July 2009.
23
Source: “ World Cellular I nformat ion Service, ” I nforma Telecoms & Media, June 2009.

HSPA t o LTE- Advanced, Rysavy Research / 3G Americas, Sept ember 2009 Page 15
environment s. UMTS- HSPA and CDMA2000 EV- DO are t he primary 3G t echnologies,
alt hough recent ly WiMAX was also designat ed as an official 3G t echnology.
The I TU has recent ly issued requirement s for I MT- Advanced, which const it ut es t he
official definit ion of 4G. Requirement s include operat ion in up t o 40 MHz radio channels
and ext remely high spect ral efficiency. The I TU recommends operat ion in up t o 100 MHz
radio channels and peak spect ral efficiency of 15 bps/ Hz, result ing in a t heoret ical
t hroughput rat e of 1. 5 Gbps. Previous t o t he publicat ion of t he requirement s, 1 Gbps
was frequent ly cit ed as a 4G goal.
No t echnology meet s t hese requirement s yet ; none is even close. I t will require new
t echnologies such as LTE- Advanced ( wit h work already underway) and I EEE 802. 16m.
Some have t ried t o label current versions of WiMAX and LTE as “ 4G” , but t his is only
accurat e t o t he ext ent t hat such designat ion refers t o t he general approach or plat form
t hat will be enhanced t o meet t he 4G requirement s.
Wit h WiMAX and HSPA significant ly out performing 3G requirement s, calling t hese
t echnologies 3G clearly does not give t hem full credit as t hey are a generat ion beyond
current t echnologies in capabilit y. But calling t hem 4G is not correct . Unfort unat ely, t he
generat ional labels do not properly capt ure t he scope of available t echnologies and have
result ed in some amount of market confusion. Some people have even called
t echnologies such as HSPA 3. 5G and LTE 3. 9G, alt hough t hese are not official
designat ions.
The following t able summarizes t he generat ions.
Tabl e 2: 1G t o 4G
Gener at i on Requi r ement s Comment s
1G No official requirement s.
Analog t echnology.
Deployed in t he 1980s.
2G No official requirement s.
Digit al Technology.
First digit al syst ems.
Deployed in t he 1990s.
New services such as SMS
and low- rat e dat a.
Primary t echnologies
include CDMA2000 1xRTT
and GSM.
3G I TU’s I MT- 2000 required 144
kbps mobile, 384 kbps
pedest rian, 2 Mbps indoors
Primary t echnologies
include CDMA2000 EV- DO
and UMTS- HSPA.
WiMAX now an official 3G
t echnology.
4G I TU’s I MT- Advanced
requirement s include abilit y t o
operat e in up t o 40 MHz radio
channels and wit h very high
spect ral efficiency.
No t echnology meet s
requirement s t oday.
I EEE 802. 16m and LTE
Advanced being designed
t o meet requirement s.


HSPA t o LTE- Advanced, Rysavy Research / 3G Americas, Sept ember 2009 Page 16
3GPP Evolutionary Approach
Rat her t han emphasizing any one wireless approach, 3GPP’s evolut ionary plan is t o
recognize t he st rengt hs and weaknesses of every t echnology and t o exploit t he unique
capabilit ies of each one accordingly. GSM, based on a TDMA approach, is mat ure and
broadly deployed. Already ext remely efficient , t here are nevert heless opport unit ies for
addit ional opt imizat ions and enhancement s. St andards bodies have already defined
“ Evolved EDGE, ” which will be available for deployment in t he 2009 t o 2010 t imeframe.
Evolved EDGE more t han doubles t hroughput over current EDGE syst ems, halves
lat ency, and increases spect ral efficiency. By t he end of t he decade, because of sheer
market moment um, t he maj orit y of worldwide subscribers will st ill be using GSM/ EDGE
t echnologies.
Meanwhile, CDMA was chosen as t he basis of 3G t echnologies including WCDMA for t he
frequency division duplex ( FDD) mode of UMTS and Time Division CDMA ( TD- CDMA) for
t he t ime division duplex ( TDD) mode of UMTS. The evolved dat a syst ems for UMTS, such
as HSPA and HSPA+ , int roduce enhancement s and simplificat ions t hat help CDMA- based
syst ems mat ch t he capabilit ies of compet ing syst ems, especially in 5 MHz spect rum
allocat ions.
I nnovat ions such as dual- carrier HSPA, explained in det ail in t he appendix sect ion
“ Evolut ion of HSPA ( HSPA+ ) , ” coordinat e t he operat ion of HSPA on t wo adj acent 5 MHz
carriers for higher t hroughput rat es. I n combinat ion wit h MI MO, dual- carrier HSPA will
achieve peak net work speeds of 84 Mbps.
Given some of t he advant ages of an OFDM approach, 3GPP has specified OFDMA as t he
basis of it s Long Term Evolut ion
24
effort . LTE incorporat es best - of- breed radio
t echniques t o achieve performance levels beyond what will be pract ical wit h CDMA
approaches, part icularly in larger channel bandwidt hs. I n t he same way t hat 3G coexist s
wit h Second Generat ion ( 2G) syst ems in int egrat ed net works, LTE syst ems will coexist
wit h bot h 3G syst ems and 2G syst ems. Mult imode devices will funct ion across LTE/ 3G or
even LTE/ 3G/ 2G, depending on market circumst ances. Beyond radio t echnology,
EPC/ SAE provides a new core archit ect ure t hat enables bot h flat t er archit ect ures and
int egrat ion of LTE wit h bot h legacy GSM- HSPA net works, as well as ot her wireless
t echnologies. The combinat ion of EPC and LTE is referred t o as t he Evolved Packet
Syst em ( EPS) .
LTE is of crucial import ance t o operat ors since it provides t he efficiencies and capabilit ies
being demanded by t he quickly growing mobile broadband market . The cost for
operat ors t o deliver dat a ( e. g. , cost per Mbyt e) is almost direct ly proport ional t o t he
spect ral efficiency of t he t echnologies. LTE has t he highest spect ral efficiency of any
specified t echnology, making it an essent ial t echnology as t he market mat ures.
LTE is available in bot h FDD and TDD modes. Many deployment s will be based on FDD in
paired spect rum. The TDD mode, however, will be import ant in enabling deployment s
where paired spect rum is unavailable.
To address I TU’s I MT- Advanced requirement s, 3GPP is developing LTE- Advanced, a
t echnology t hat will have peak rat es of more t han 1 Gbps. See t he appendix sect ion “ 4G,
I MT- Advanced and LTE- Advanced” for a det ailed explanat ion.
Alt hough lat er sect ions quant ify performance and t he appendix of t he whit e paper
present s funct ional det ails of t he different t echnologies, t his sect ion provides a summary

24
3GPP also refers t o LTE as Enhanced UMTS Terrest rial Radio Access Net work ( E- UTRAN) .

HSPA t o LTE- Advanced, Rysavy Research / 3G Americas, Sept ember 2009 Page 17
int ended t o provide a frame of reference for t he subsequent discussion. Table 3
summarizes t he key 3GPP t echnologies and t heir charact erist ics.

Tabl e 3: Char act er i st i cs of 3GPP Technol ogi es
Technol ogy
Name
Ty pe Char act er i st i cs Ty pi cal
Dow nl i nk
Speed
Ty pi cal Upl i nk
Speed
GSM TDMA Most widely deployed
cellular t echnology in t he
world. Provides voice and
dat a service via
GPRS/ EDGE.

EDGE TDMA Dat a service for GSM
net works. An enhancement
t o original GSM dat a service
called GPRS.
70 kbps
t o 135 kbps
70 kbps
t o 135 kbps
Evolved
EDGE
TDMA Advanced version of EDGE
t hat can double and
event ually quadruple
t hroughput rat es, halve
lat ency and increase
spect ral efficiency.
175 kbps
t o 350 kbps
expect ed
( Single Carrier)
350 kbps
t o 700 kbps
expect ed ( Dual
Carrier)

150 kbps
t o 300 kbps
expect ed
UMTS CDMA 3G t echnology providing
voice and dat a capabilit ies.
Current deployment s
implement HSPA for dat a
service.
200 t o 300
kbps
200 t o 300
kbps
HSPA CDMA Dat a service for UMTS
net works. An enhancement
t o original UMTS dat a
service.
1 Mbps t o
4 Mbps
500 kbps
t o 2 Mbps
HSPA+ CDMA Evolut ion of HSPA in
various st ages t o increase
t hroughput and capacit y
and t o lower lat ency.
1. 5 Mbps t o
7 Mbps
1 Mbps t o
4 Mbps
LTE OFDMA New radio int erface t hat
can use wide radio channels
and deliver ext remely high
t hroughput rat es. All
communicat ions handled in
I P domain.
4 Mbps t o
24 Mbps
( in 2 x 20 MHz)

LTE-
Advanced
OFDMA Advanced version of LTE
designed t o meet I MT-
Advanced requirement s.


HSPA t o LTE- Advanced, Rysavy Research / 3G Americas, Sept ember 2009 Page 18

User achievable rat es and great er det ails on t ypical rat es are covered in Table 6 in t he
sect ion “ Dat a Throughput ” lat er in t his paper. Figure 5 shows t he evolut ion of t he
different wireless t echnologies and t heir peak net work performance capabilit ies.
Fi gur e 5: Ev ol ut i on of TDMA, CDMA, and OFDMA Sy st ems

The development of GSM and UMTS- HSPA happens in st ages referred t o as 3GPP
releases, and equipment vendors produce hardware t hat support s part icular versions of
each specificat ion. I t is import ant t o realize t hat t he 3GPP releases address mult iple
t echnologies. For example, Release 7 opt imized VoI P for HSPA, but also significant ly
enhanced GSM dat a funct ionalit y wit h Evolved EDGE. A summary of t he different 3GPP
releases is as follows:
25

 Rel ease 99: Complet ed. First deployable version of UMTS. Enhancement s t o
GSM dat a ( EDGE) . Maj orit y of deployment s t oday are based on Release 99.
Provides support for GSM/ EDGE/ GPRS/ WCDMA radio- access net works.
 Rel ease 4: Complet ed. Mult imedia messaging support . First st eps t oward using
I P t ransport in t he core net work.

25
Aft er Release 99, release versions went t o a numerical designat ion inst ead of designat ion by year.
EDGE
DL: 474 kbps
UL: 474 kbps
Evolved
EDGE
DL: 1.89 Mbps
UL: 947 kbps
2008 2009 2010 2012 2013
Notes: Throughput rates are peak theoretical network rates. Radio channel bandwidths indicated.
Dates refer to expected initial commercial network deployment except 2008, which shows available technologies that year.
2011
Fixed WiMAX
HSPA
DL: 14.4 Mbps
UL: 5.76 Mbps
In 5 MHz
Rel 7 HSPA+
DL: 28 Mbps
UL: 11.5 Mbps
In 5 MHz
LTE
DL: 326 Mbps
UL: 86 Mbps
In 20 MHz
EV-DO Rev A
DL: 3.1 Mbps
UL: 1.8 Mbps
In 1.25 MHz
EV-DO Rev B
DL: 14.7 Mbps
UL: 4.9 Mbps
In 5 MHz
Rel 8 HSPA+
DL: 42 Mbps
UL: 11.5 Mbps
In 5 MHz
LTE
(Rel 9)
Release 1.0
DL: 46 Mbps
UL: 4 Mbps
10 MHz 3:1 TDD
Rel 1.5 IEEE 802.16m
LTE Advanced
DL: > 1 Gbps
Rel 9 HSPA+
DL: 84 Mbps
UL: 23 Mbps
In 10 MHz

HSPA t o LTE- Advanced, Rysavy Research / 3G Americas, Sept ember 2009 Page 19
 Rel ease 5: Complet ed. HSDPA. First phase of I MS. Full abilit y t o use I P- based
t ransport inst ead of j ust Asynchronous Transfer Mode ( ATM) in t he core net work.
 Rel ease 6: Complet ed. HSUPA. Enhanced mult imedia support t hrough
Mult imedia Broadcast / Mult icast Services ( MBMS) . Performance specificat ions for
advanced receivers. WLAN int egrat ion opt ion. I MS enhancement s. I nit ial VoI P
capabilit y.
 Rel ease 7: Complet ed. Provides enhanced GSM dat a funct ionalit y wit h Evolved
EDGE. Specifies HSPA Evolut ion ( HSPA+ ) , which includes higher order modulat ion
and MI MO. Provides fine- t uning and increment al improvement s of feat ures from
previous releases. Result s include performance enhancement s, improved spect ral
efficiency, increased capacit y, and bet t er resist ance t o int erference. Cont inuous
Packet Connect ivit y ( CPC) enables efficient “ always- on” service and enhanced
uplink UL VoI P capacit y, as well as reduct ions in call set - up delay for PoC. Radio
enhancement s t o HSPA include 64 QAM in t he downlink DL and 16 QAM in t he
uplink. Also includes opt imizat ion of MBMS capabilit ies t hrough t he
mult icast / broadcast , single- frequency net work ( MBSFN) funct ion.
 Rel ease 8: Complet ed. Comprises furt her HSPA Evolut ion feat ures such as
simult aneous use of MI MO and 64 QAM. I ncludes work it em for dual- carrier HSPA
( DC- HSPA) wherein t wo WCDMA radio channels can be combined for a doubling
of t hroughput performance. Specifies OFDMA- based 3GPP LTE. Defines EPC.
 Rel ease 9: Under development . Likely 2010. Will include HSPA and LTE
enhancement s including HSPA mult i- carrier operat ion.
 Rel ease 10: Under development . Likely 2011. Will specify LTE- Advanced t hat
meet s t he requirement s set by I TU’s I MT- Advanced proj ect .
Whereas operat ors and vendors act ively involved in t he development of wireless
t echnology are heavily focused on 3GPP release versions, most users of t he
t echnology are more int erest ed in part icular feat ures and capabilit ies such as
whet her a device support s HSDPA. For t his reason, t he det ailed discussion of t he
t echnologies in t his paper emphasizes feat ures as opposed t o 3GPP releases.
Spectrum
Anot her import ant aspect of UMTS- HSPA deployment is t he expanding number of
available radio bands, as shown in Figure 6, and t he corresponding support from
infrast ruct ure and mobile- equipment vendors. The fundament al syst em design and
net working prot ocols remain t he same for each band; only t he frequency- dependent
port ions of t he radios have t o change.
As ot her frequency bands become available for deployment , st andards bodies are
adapt ing UMTS for t hese bands as well. This includes 450 and 700 MHz. UMTS- TDD
equipment is already available for 450 MHz. The 1710- 1770 uplink was mat ched wit h
2110- 2170 downlink t o allow for addit ional global harmonizat ion of t he 1. 7/ 2. 1GHz
band. Meanwhile, t he Federal Communicat ions Commission auct ioned t he 700 MHz band
in t he Unit ed St at es in January 2008. The availabilit y of t his band, t he Advanced
Wireless Services ( AWS) band at 1710- 1755 MHz wit h/ 2110- 2155 MHz in t he US, and
t he fort hcoming 2. 6 GHz frequency band in Europe are providing operat ors wit h wider
deployment opt ions. An increasing number of operat ors are also deploying UMTS at 900
MHz, a t radit ional GSM band.

HSPA t o LTE- Advanced, Rysavy Research / 3G Americas, Sept ember 2009 Page 20
As t he t ot al amount of available spect rum increases and as t echnologies simult aneously
become spect rally more efficient , t ot al capacit y rises rapidly, support ing more
subscribers and making many new t ypes of applicat ions feasible.
The following figure shows t he FDD bands defined for 3GPP t echnologies.
Fi gur e 6: FDD Bands f or 3GPP Technol ogi es
26



I t should be not ed t hat alt hough t he support of a new frequency band may be
int roduced in a part icular release, t he 3GPP st andard also specifies ways t o implement
devices and infrast ruct ure operat ing on any frequency band, according t o release
ant erior t o t he int roduct ion of t hat part icular frequency band. For example, alt hough
band 5 ( US Cellular Band) was int roduced in Release 6, t he first devices operat ing on
t his band were compliant wit h t he release 5 of t he st andard.
Figure 7 shows TDD bands defined for 3GPP Technologies.

26
Source: A 3G Americas’ member company.
850 MHz
1800 MHz
1900 MHz
2.1 GHz
2.6 GHz
1.7/2.1 GHz
900 MHz
Band 1
Band 2
Band 3
Band 4
Band 5
Band 7
Band 8
1700 MHz Band 9
800 MHz Band 6
2x25 MHz
2x75 MHz
2x60 MHz
2x60 MHz
2x70 MHz
2x45 MHz
2x35 MHz
2x35 MHz
2x10 MHz
824-849
1710-1785
1850-1910
1920-1980
2500-2570
1710-1755
880-915
1749.9-1784.9
830-840
Operating
band
Band name
Total
spectrum
Uplink [MHz]
869-894
1805-1880
1930-1990
2110-2170
2620-2690
2110-2155
925-960
1844.9-1879.9
875-885
Downlink [MHz]
Ext 1.7/2.1MHz Band 10 2x60 MHz 1710-1770 2110-2170
1500 MHz Band 11 2x25 MHz 1427.9 - 1452.9 1475.9 - 1500.9
Lower 700 MHz Band 12 2x18 MHz 698-716 728-746
Upper 700 MHz Band 13 2x10 MHz 777-787 746-756
Upper 700 MHz,
public safety/private
Band 14 2x10 MHz 788-798 758-768

HSPA t o LTE- Advanced, Rysavy Research / 3G Americas, Sept ember 2009 Page 21
Fi gur e 7: TDD Bands f or 3GPP Technol ogi es
27



Different count ries have regulat ed spect rum more loosely t han ot hers. For example,
operat ors in t he Unit ed St at es can use eit her 2G or 3G t echnologies in cellular, Personal
Communicat ions Service ( PCS) , and 3G bands, whereas in Europe t here are great er
rest rict ions—alt hough effort s are underway t hat are result ing in great er flexibilit y
including t he use of 3G t echnologies in current 2G bands.
Wit h t he proj ect ed increase in t he use of mobile- broadband t echnologies, t he amount of
spect rum required by t he next generat ion of wireless t echnology ( t hat is, aft er 3GPP LTE
in proj ect s such as I nt ernat ional Mobile Telecommunicat ions ( I MT) Advanced) could be
subst ant ial given t he desire t o operat e radio channels as wide as 100 MHz. I deally, t his
spect rum would fall below 5 GHz. This search for new spect rum is a long- t erm
undert aking, and it may be well int o t he next decade before any such new spect rum
becomes available. Given t he expanding size and economic significance of t he mobile-
comput ing indust ry, however, decisions made on new spect rum—especially wit h respect
t o global harmonizat ion—will have profound consequences.
As regulat ors make more spect rum available, it is import ant t hat such spect rum be:
1. Harmonized on a regional or global basis.
2. Unencumbered by spect rum caps and ot her legacy voice- cent ric spect rum
policies.
3. Made available in as wide radio channels as possible ( i. e. , 10 MHz, 20 MHz and
more) .
4. Ut ilized efficient ly wit hout causing int erference t o exist ing spect rum holders.
Emerging t echnologies such as LTE benefit from wider radio channels. These wider radio
channels are not only spect rally more efficient , but offer great er capacit y, an essent ial
at t ribut e because t ypical broadband usage cont ribut es t o a much higher load t han a

27
Source: A 3G Americas’ member company.
Operating
band
Total
spectrum
Frequencies [MHz]
Band 39
Band 40
40 MHz
100 MHz
1880-1920
2300-2400
Band 38 50 MHz 2570-2620
Band 33
Band 34
Band 35
Band 36
Band 37 20 MHz
60 MHz
15 MHz
20 MHz
60 MHz
1910-1930
1850-1910
2010-2025
1900-1920
1930-1990

HSPA t o LTE- Advanced, Rysavy Research / 3G Americas, Sept ember 2009 Page 22
voice user. For inst ance, wat ching a YouTube video consumes 100 t imes as many bit s
per second on t he downlink as a voice call.
Figure 8 shows increasing LTE spect ral efficiency obt ained wit h wider radio channels,
wit h 20 MHz showing t he most efficient configurat ion.
Fi gur e 8: LTE Spect r al Ef f i ci ency as Funct i on of Radi o Channel Si ze
28


Of some concern in t his regard is t hat spect rum for LTE is becoming available in different
frequency bands in different count ries. For inst ance, init ial US deployment s will be at
700 MHz, in Japan at 1500 MHz and in Europe at 2. 6 GHz. Thus, wit h some many
varying spect rum bands, it will most likely necessit at e t hat roaming operat ion be based
on GSM or HSPA on common regional or global bands.
Core-Network Evolution
3GPP is defining a series of enhancement s t o t he core net work t o improve net work
performance and t he range of services provided, and t o enable a shift t o all- I P
archit ect ures.
One way t o improve core- net work performance is by using flat t er archit ect ures. The
more hierarchical a net work, t he more easily it can be managed cent rally; t he t radeoff,
however, is reduced performance, especially for dat a communicat ions, because packet s
must t raverse and be processed by mult iple nodes in t he net work. To improve dat a
performance and, in part icular, t o reduce lat ency ( delays) , 3GPP has defined a number
of enhancement s in Release 7 and Release 8 t hat reduce t he number of processing
nodes and result in a flat t er archit ect ure.
I n Release 7, an opt ion called one- t unnel archit ect ure allows operat ors t o configure t heir
net works so t hat user dat a bypasses a serving node and t ravels direct ly via a gat eway

28
Source: 3G Americas’ member company analysis.
0
10
20
30
40
50
60
70
80
90
100
1.4 3 5 10 20
%
 
E
f
f
i
c
i
e
n
c
y
 
R
e
l
a
t
i
v
e
 
t
o
 
2
0
 
M
H
z
MHz

HSPA t o LTE- Advanced, Rysavy Research / 3G Americas, Sept ember 2009 Page 23
node. There is also an opt ion t o int egrat e t he funct ionalit y of t he radio- net work
cont roller direct ly int o t he base st at ion.
For Release 8, 3GPP has defined an ent irely new core net work, called t he Evolved Packet
Core, previously called Syst em Archit ect ure Evolut ion. The key feat ures and capabilit ies
of EPC/ SAE include:
 Reduced lat ency and higher dat a performance t hrough a flat t er archit ect ure.
 Support for bot h LTE radio- access net works and int erworking wit h GSM- HSPA
radio- access net works.
 The abilit y t o int egrat e non- 3GPP net works such as WiMAX.
 Opt imizat ion for all services provided via I P.
This paper provides furt her det ails in t he sect ions on HSPA Evolut ion ( HSPA+ ) and
EPC/ SAE.
Service Evolution
Not only do 3GPP t echnologies provide cont inual improvement s in capacit y and dat a
performance, t hey also evolve capabilit ies t hat expand t he services available t o
subscribers. Key service advances include Fixed Mobile Convergence ( FMC) , I MS, and
broadcast ing t echnologies. This sect ion provides an overview of t hese t opics, and t he
appendix provides great er det ail on each of t hese it ems.
FMC refers t o t he int egrat ion of fixed services ( such as t elephony provided by wireline or
Wi- Fi) wit h mobile cellular- based services. Though FMC is st ill in it s early st ages of
deployment by operat ors, it promises t o provide significant benefit s t o bot h users and
operat ors. For users, FMC will simplify how t hey communicat e making it possible for
t hem t o use one device ( for example, a cell phone) at work and at home where it might
connect via a Wi- Fi net work or a femt o cell. When mobile, users connect via a cellular
net work. Users will also benefit from single voice mailboxes and single phone numbers,
as well as t he abilit y t o cont rol how and wit h whom t hey communicat e. For operat ors,
FMC allows t he consolidat ion of core services across mult iple- access net works. For
inst ance, an operat or could offer complet e VoI P- based voice service t hat support s access
via DSL, Wi- Fi, or 3G. FMC also offloads t he macro net work from dat a- int ensive
applicat ions such as movie downloads.
There are various approaches for FMC, including Generic Access Net work ( GAN) ,
formerly known as Unlicensed Mobile Access ( UMA) , femt ocells, and I MS. Wit h GAN,
GSM- HSPA devices can connect via Wi- Fi or cellular connect ions for bot h voice and dat a.
UMA/ GAN is a 3GPP t echnology, and it has been deployed by a number of operat ors,
including T- Mobile in t he Unit ed St at es. An alt ernat ive t o using Wi- Fi for t he “ fixed”
port ion of FMC is femt ocells. These are t iny base st at ions t hat cost lit t le more t han a Wi-
Fi access point and, like Wi- Fi, femt ocells leverage a subscriber' s exist ing wireline-
broadband connect ion ( for example, DSL) . I nst ead of operat ing on unlicensed bands,
femt ocells use t he operat or’s licensed bands at very low power levels. The key
advant age of t he femt ocell approach is t hat any single- mode, mobile- communicat ions
device a user has can now operat e using t he femt ocells.
I MS is anot her key t echnology for convergence. I t allows access t o core services and
applicat ions via mult iple- access net works. I MS is more powerful t han GAN, because it
support s not only FMC, but also a much broader range of pot ent ial applicat ions. I n t he
Unit ed St at es, AT&T has commit t ed t o an I MS approach and has already deployed an
I MS- based video sharing service. Alt hough defined by 3GPP, t he Third Generat ion

HSPA t o LTE- Advanced, Rysavy Research / 3G Americas, Sept ember 2009 Page 24
Part nership Proj ect 2 ( 3GPP2) , CableLabs and WiMAX have adopt ed I MS. I MS is how
VoI P will ( or could) be deployed in CDMA 2000 EV- DO, WiMAX, HSPA and LTE net works.
I MS allows t he creat ive blending of different t ypes of communicat ions and informat ion,
including voice, video, I M, presence informat ion, locat ion, and document s. I t provides
applicat ion developers t he means t o creat e applicat ions t hat have never before been
possible, and it allows people t o communicat e in ent irely new ways by dynamically using
mult iple services. For example, during an int eract ive chat session, a user could launch a
voice call. Or during a voice call, a user could suddenly est ablish a simult aneous video
connect ion or st art t ransferring files. While browsing t he Web, a user could decide t o
speak t o a cust omer- service represent at ive. I MS will be a key plat form for all- I P
archit ect ures for bot h HSPA and LTE.
A new init iat ive called Rich Communicat ions Suit e ( RCS) , support ed by many operat ors
and vendors, builds upon I MS t echnology t o provide a consist ent feat ure set , as well as
implement at ion guidelines, use cases, and reference implement at ions. RCS uses exist ing
st andards and specificat ions from 3GPP, OMA and GSMA.
Core feat ures include:
- An enhanced phone book ( device and/ or net work based) t hat includes service
capabilit ies and presence- enhanced cont act informat ion.
- Enhanced messaging ( support ing t ext , inst ant messaging and mult imedia) wit h
chat and messaging hist ory.
- Enriched calls t hat include mult imedia cont ent ( e. g. , video sharing) during voice
calls.
Anot her import ant new service is support for mobile TV t hrough what is called mult icast
or broadcast funct ions. 3GPP has defined mult icast / broadcast capabilit ies for bot h HSPA
and LTE.

Device Innovation
Comput ing it self is becoming more mobile, and not ebook comput ers and smart phones
are now prevalent . I n fact , all mobile phones are becoming “ smart , ” wit h some form of
dat a capabilit y, and leading not ebook vendors are now offering comput ers wit h
int egrat ed 3G ( e. g. , HSPA) capabilit ies. Modems are available in mult iple format s
including USB devices, PC Cards and Express cards.
Comput er manufact urers are also delivering new form fact ors such as net books, mobile
I nt ernet devices ( MI D) and smart books. The movement t o open net works also allows a
great er number of companies t o develop product s t hat use wireless net works in bot h
vert ical- market and horizont al- market scenarios. According t o a recent report by
Forward Concept s, t he global MI D- only market is expect ed t o grow from 305, 000
shipment s in 2008 t o 40 million in 2012.
29

Cellular t elephones are becoming more powerful and feat ure large color t ouch displays,
graphics viewers, st ill cameras, movie cameras, MP3 players, I M client s, e- mail client s,
Push- t o- Talk over Cellular ( PoC) , downloadable execut able cont ent capabilit ies, and ever
more powerful browsers. All t hese capabilit ies consume dat a.

29
Source: Forward Concept s, Mobile I nt ernet Device and Chip Market Opport unit ies, June 2008.

HSPA t o LTE- Advanced, Rysavy Research / 3G Americas, Sept ember 2009 Page 25
Meanwhile, smart phones, which emphasize a rich comput ing environment on a phone,
represent t he convergence of t he personal digit al assist ant , a fully capable mobile
comput er, and a phone, all in a device t hat is only slight ly larger t han t he average
cellular t elephone. Many users would prefer t o carry one device t hat “ does it all. ”
Smart phones, originally t arget ed for t he high end of t he market , are now available at
much lower price point s and t hus affordable t o a much larger market segment . Ovum
predict s t hat smart phones will const it ut e 29% of phones by 2014.
30
The success of t he
iPhone demonst rat es t he pot ent ial of t his market .

Network Interfaces for Applications
Anot her import ant development relat ed t o service evolut ion is operat ors making
int erfaces available t o ext ernal applicat ions for informat ion and cont rol. Two widely
deployed capabilit ies t oday include locat ion queries and short message service. Wit h
locat ion, mobile devices or ext ernal applicat ions ( e. g. , applicat ions operat ing on
comput ers out side of t he net work) can query t he locat ion of a user, subj ect t o privacy
rest rict ions. This can significant ly enhance many applicat ions including navigat ion,
supplying locat ion of nearby dest inat ions ( e. g. , rest aurant s, st ores) , locat ion of friends
for social net working, and worker dispat ch. Wit h SMS, ext ernal applicat ions can send
user request ed cont ent such as flight updat es.
Unt il now, t he int erfaces for such funct ions have eit her been propriet ary, or specific t o
t hat funct ion. However, t here are now int erfaces t hat span mult iple funct ions using a
consist ent set of programming met hods. One set is t he Parlay X Web Services, a set of
funct ions specified t hrough a j oint proj ect of t he Parlay Group, t he European
Telecommunicat ions St andards I nst it ut e ( ETSI ) and 3GPP. The Open Mobile Alliance
( OMA) now manages t he Parlay X specificat ions. Parlay X Web Services include support
for locat ion and SMS, as well as many ot her funct ions wit h which developers will be able
t o build innovat ive applicat ions.


30
Source: Ovum Comment , Adam Leach, Devices principal analyst , “ Smart phones: t he silver lining of
t he declining handset mar ket , ” June 2009.

HSPA t o LTE- Advanced, Rysavy Research / 3G Americas, Sept ember 2009 Page 26
Table 4 summarizes t he available Parlay X specificat ions.
31
Operat ors are beginning t o
select ively deploy t hese funct ions. The advant age of t his approach is t hat developers can
build applicat ions t hat are compat ible wit h mult iple operat or net works.



31
See ht t p: / / www. parlay. org/ en/ specificat ions/ pxws.asp for act ual specificat ions.

HSPA t o LTE- Advanced, Rysavy Research / 3G Americas, Sept ember 2009 Page 27
Tabl e 4: Par l ay X Speci f i cat i ons
Par t Ti t l e Funct i ons
1 Common Definit ions common across Parlay X specificat ions
2 Third Part y Call Creat es and manages calls
3 Call Not ificat ion Management of calls init iat ed by a subscriber
4 Short Messaging Send and receive of SMS including delivery receipt s
5 Mult imedia Messaging Send and receive of mult imedia messages
6 Payment Pre- paid and post - paid payment s and payment
reservat ions
7 Account Management Management of account s of prepaid cust omers
8 Terminal St at us Obt ain st at us such as reachable, unreachable or busy
9 Terminal Locat ion Obt ain locat ion of t erminal
10 Call Handling Cont rol by applicat ion for call handling of specific
numbers
11 Audio Call Cont rol for media t o be added/ dropped during call
12 Mult imedia Conference Creat e mult imedia conferences including dynamic
management of part icipant s
13 Address List
Management
Manage subscriber groups
14 Presence Provide presence informat ion
15 Message Broadcast Send messages t o all users in specified area
16 Geocoding Obt ain locat ion address of subscriber
17 Applicat ion- driven QoS Cont rol qualit y of service of end- user connect ion
18 Devices Capabilit ies
and Configurat ion
Obt ain device capabilit y informat ion and be able t o
push device configurat ion t o device
19 Mult imedia St reaming
Cont rol
Cont rol mult imedia st reaming t o device
20 Mult imedia Mult icast
Session Management
Cont rol mult icast sessions, members, mult imedia
st ream and obt ain channel presence informat ion

A relat ed proj ect is GSMA OneAPI , a GSM Associat ion proj ect t o also define net work
int erfaces, but t hat priorit izes implement at ion based on expect ed market demand.
OneAPI defines a simplified Web service for most funct ions t hat is essent ially a subset of
t he relat ed Parlay X Web service.
32
I t also defines a REST ( Represent at ional St at e
Transfer) int erface for most funct ions as an alt ernat ive t o using t he Web service.
RESTful int erfaces are simpler for developers t o work wit h and experiment wit h t han
Web services.
Regardless of whet her operat ors deploy wit h Parlay X or OneAPI , t hese are mainst ream
int erfaces t hat will open wireless net works t o t housands of I nt ernet programmers who
will be able t o build applicat ions t hat leverage t he lat ent informat ion and capabilit ies of
wireless net works.
Mobile Application Architectures
Many applicat ions used over wireless connect ions will be t he same as t hose used over
t he I nt ernet wit h deskt op/ lapt op PCs. An increasing number of applicat ions, however,
will be developed specifically for mobile devices. This can be a challenge for developers,

32
See ht t p: / / oneapi. aepona. com/ port al/ t ws_gsma/ Resources for more informat ion about OneAPI .

HSPA t o LTE- Advanced, Rysavy Research / 3G Americas, Sept ember 2009 Page 28
because t here are a number of different mobile plat forms now available including
Android, Apple iPhone, LiMo, Palm Pre, RI M BlackBerry, Symbian and Windows Mobile.
Unlike t he deskt op market , t he mobile device market has become quit e fragment ed.
Each of t he device plat forms comes wit h it s own applicat ion development environment ,
and developers must face a learning curve t o become adept at programming for any
specific plat form. Some developers may be cont ent t arget ing specific plat forms. Ot hers,
however, may need t heir applicat ions t o operat e across mult iple plat forms.
Fort unat ely, t here are various development s t hat address t he fragment at ion challenge.
These include:
- Mobile Middleware. These are soft ware infrast ruct ures t hat consist of a client
component t hat operat es on t he mobile device, and a server component t hat act s
as a proxy for t he client . Vendors provide t ools wit h which developers can
develop an applicat ion in a plat form- neut ral manner, and which t hen operat es on
mult iple device t ypes. Mobile middleware is most ly used for business
applicat ions.
- Mobile Web 2. 0. Mobile browsers are adopt ing many of t he same sophist icat ed
capabilit ies as deskt op browsers. Combined wit h net works t hat have higher
t hroughput s and lower lat ency, an increasing number of applicat ions can be Web
host ed, making t he applicat ions available from diverse plat forms. Mobile Web 2. 0
t echnologies include it ems such as Aj ax, offline operat ion, video capabilit ies, fast
JavaScript execut ion, and mashups ( combining dat a from mult iple Web sources) .
Cloud comput ing, enabled by Mobile Web 2. 0, will play as import ant a role for
mobile syst ems as for deskt ops.
- Java Development s. Though Java it self has present ed a challenge t hrough
inconsist ent implement at ion on devices, t here are new capabilit ies t hat will result
in more consist ent , as well as more powerful, device execut ion environment s.
Examples include Mobile Service Archit ect ure ( MSA) for predict able capabilit y and
Mobile I nformat ion Device Profile 3 for mult i- t asking.
Broadband-Wireless Deployment Considerations
Much of t he debat e in t he wireless indust ry is on t he merit s of different radio
t echnologies, yet ot her fact ors are equally import ant in det ermining t he services and
capabilit ies of a wireless net work. These fact ors include t he amount of spect rum
available, backhaul, and net work t opology.
Spect rum has always been a maj or considerat ion for deploying any wireless net work,
but it is part icularly import ant when looking at high- performance broadband syst ems.
HSPA and HSPA+ can deliver high t hroughput rat es on t he downlink and uplink wit h low
lat ency in 5 MHz channels when deployed in single frequency ( 1/ 1) reuse. By t his, we
mean t hat every cell sect or ( t ypically t hree per cell) in every cell uses t he same radio
channel( s) .
As previously discussed, an OFDMA approach in a 5 MHz radio channel yields only a
small performance advant age. To achieve higher dat a rat es requires wider radio
channels, such as 10 or 20 MHz wide channels, in combinat ion wit h emerging OFDMA
radio t echnologies. Very few operat ors t oday, however, have access t o t his much
spect rum. I t was challenging enough for GSM operat ors t o obt ain UMTS spect rum. I f
delivering very high dat a rat es is t he obj ect ive, t hen t he syst em must minimize
int erference. This result is best achieved by employing looser reuse, such as having
every sect or use only one- t hird of t he available radio channels ( 1/ 3 reuse) . The 10 MHz
radio channel could now demand as much as 30 MHz of available spect rum.

HSPA t o LTE- Advanced, Rysavy Research / 3G Americas, Sept ember 2009 Page 29
Backhaul is anot her fact or. As t he t hroughput of t he radio link increases, t he circuit s
connect ing t he cell sit es t o t he core net work must be able t o handle t he increased load.
Wit h many cell sit es t oday serviced by j ust a small number of T1/ E1 circuit s, each able
t o carry only 1. 5/ 2. 0 Mbps, operat ors will have t o significant ly upgrade backhaul
capacit y t o obt ain t he full benefit of next - generat ion wireless t echnologies. An OFDMA
syst em wit h 1. 5 bps per hert z ( Hz) of spect ral efficiency in 10 MHz on t hree sect ors has
up t o 45 Mbps average cell t hroughput .
Addit ionally, any t echnology’s abilit y t o reach it s peak spect rum efficiency is somewhat
cont ingent on t he syst em’s abilit y t o reach t he inst ant aneous peak dat a rat es allowed by
t hat t echnology. For example, a syst em claiming spect rum efficiency of 1. 5 bps/ Hz ( as
described above) might rely on t he abilit y t o reach 100 Mbps inst ant aneously t o achieve
t his level of spect rum efficiency. Any const raint on t he t ransport syst em below 100 Mbps
will rest rict t he range of achievable t hroughput and, in t urn, impact t he spect ral
efficiency of t he syst em.
The mismat ch bet ween backhaul capabilit ies and radio performance in some net works is
one reason t hat user rat es on some 3G syst ems are lower t han t heoret ical rat es.
Operat ors are act ively enhancing t heir backhaul approaches, and t here are many
available and emerging wireline t echnologies such as VDSL and opt ical Et hernet , as well
as compet it ive point - t o- point microwave syst ems t hat make t his possible.
Finally, t he overall net work t opology also plays an import ant role, especially wit h respect
t o lat ency. Low lat ency is crit ical t o achieving very high dat a rat es, because of t he way it
affect s TCP/ I P t raffic. How t raffic rout es t hrough t he core net work—how many hops and
nodes it must pass t hrough—can influence t he overall performance of t he net work. One
way t o increase performance is by using flat t er archit ect ures, meaning a less
hierarchical net work wit h more direct rout ing from mobile device t o end syst em. The
core EPC/ SAE net work for 3GPP LTE emphasizes j ust such a flat t er archit ect ure.
I n summary, it can be misleading t o say t hat one wireless t echnology out performs
anot her wit hout a full underst anding of how t hat t echnology will be deployed in a
complet e syst em t hat also t akes spect rum int o account .
Feature and Network Roadmap
GSM operat ors first enhanced t heir net works t o support dat a capabilit y t hrough t he
addit ion of GPRS infrast ruct ure wit h t he abilit y t o use exist ing cell sit es, t ransceivers,
and int erconnect ion facilit ies. Since inst alling GPRS, GSM operat ors have largely
upgraded dat a service t o EDGE, and any new GSM net work includes EDGE capabilit y.
Operat ors have deployed UMTS- HSPA worldwide. Alt hough UMTS involves a new radio-
access net work, several fact ors facilit at e deployment . First , most UMTS cell sit es can be
collocat ed in GSM cell sit es enabled by mult i- radio cabinet s t hat can accommodat e
GSM/ EDGE, as well as UMTS equipment . Second, much of t he GSM/ GPRS core net work
can be used. This means t hat all core- net work element s above t he Serving GPRS
Support Node ( SGSN) and Mobile Swit ching Cent er ( MSC) —t he Gat eway GPRS Support
Node ( GGSN) , t he Home Locat ion Regist er ( HLR) , billing and subscriber administ rat ion
syst ems, service plat forms, and so fort h—need, at most , a soft ware upgrade t o support
3G UMTS- HSPA. And while early 3G deployment used separat e 2G/ 3G SGSNs and MSCs,
all- new MSC and/ or SGSN product s are capable of support ing bot h GSM and UMTS- HSPA
radio- access net works. Similarly, new HSPA equipment will be upgradeable t o LTE
t hrough a soft ware upgrade.
New feat ures such as HSDPA, HSUPA, and MBMS are being designed so t hat t he same
upgraded UMTS radio channel can support a mixt ure of t erminals including t hose based

HSPA t o LTE- Advanced, Rysavy Research / 3G Americas, Sept ember 2009 Page 30
on 3GPP Release 99, Release 5, and Release 6. I n ot her words, a net work support ing
Release 5 feat ures ( for example, HSDPA) can support Release 99, Release 5, and
Release 6 t erminals ( for example, HSUPA) operat ing in a Release 5 mode. Alt ernat ively,
a net work support ing Release 6 feat ures can support Release 99, Release 5, and Release
6 t erminals. This flexibilit y assures t he maximum degree of forward- and backward-
compat ibilit y. Not e also t hat most UMTS t erminals t oday support GSM, t hus facilit at ing
use across large coverage areas and mult iple net works.
Once deployed, operat ors can minimize t he cost s of managing GSM/ EDGE and UMTS
net works, because t hese net works share many of t he same aspect s including:
 Packet - dat a archit ect ure
 Cell sit es
 Ant enna syst ems
 Backhaul circuit s
 Subscriber account management
 Service plat forms
Users largely don’t even need t o know t o what t ype of net work t hey are connect ed,
because t heir mult imode GSM- HSPA devices can seamlessly hand off bet ween net works.
The changes being planned for t he core net work are anot her aspect of evolut ion. Here,
t he int ent is t o reduce t he number of nodes t hat packet s must t raverse. This will result
in bot h reduced deployment cost s and reduced lat ency. The key enabling t echnology is
EPC/ SAE, which is described in det ail lat er in t his paper.
The upgrade t o LTE will be relat ively st raight forward, wit h new LTE infrast ruct ure having
t he abilit y t o reuse a significant amount of t he UMTS- HSPA cell sit e and base st at ion
including using t he same shelt er, t ower, ant ennas, power supply and climat e cont rol.
Different vendors have different so- called “ zero- foot print ” solut ions allowing operat ors t o
use empt y space t o enable re- use of exist ing sit es wit hout t he need for any new floor
space.
An operat or can add LTE capabilit y simply by adding an LTE baseband card. New mult i-
st andard radio unit s ( HSPA and LTE) , as well as LTE- only baseband cards, are
mechanically compat ible wit h older building pract ices, so t hat operat ors can use empt y
space in an old base st at ion for LTE baseband cards, t hus enabling re- use of exist ing
sit es wit hout t he need for any new floor space, as ment ioned previously.
Base st at ion equipment is available for many bands including t he 1. 7/ 2. 1 GHz AWS band
and t he recent ly auct ioned 700 MHz bands in t he US. Vendors and operat ors are
planning LTE commercial deployment s beginning in 2010.
On t he device side, mult i- mode chipset s will enable devices t o easily operat e across
UMTS and LTE net works. For example, one chipset vendor has announced a series of
chips t hat support t he following combinat ion of t echnologies: UMTS, HSPA+ and LTE;
EV- DO Rev B; and UMTS, HSPA+ , EV- DO Rev B and LTE.
33

One import ant and int erest ing aspect of t echnology deployment is t hat an advanced
t echnology such as LTE enables operat ors t o upgrade prior t echnologies, such as HSPA.
Examples include:

33
ht t p: / / www. qualcomm. com/ press/ r eleases/ 2008/ 080207_Qualcomm_t o_Ship. ht ml.

HSPA t o LTE- Advanced, Rysavy Research / 3G Americas, Sept ember 2009 Page 31
 VoI P for HSPA. Since LTE uses an I P core, once it is deployed, support ing voice
on HSPA via VoI P will be a much simpler t ask as it can share t he same core I P
net work as LTE.
 Device processing power. Support ing t he high t hroughput rat es wit h LTE ( e. g.,
50 Mbps or higher) will provide sufficient processing in t he device t o also support
very high HSPA rat es ( e. g. , 30 Mbps or higher) .
Table 5 shows t he rollout of EDGE/ HSPA/ LTE feat ures over t ime.
Tabl e 5: Expect ed UMTS/ LTE Feat ure and Capabilit y Availabilit y
Year Feat ur es
2009 Net works and devices capable of Release 7 HSPA+ , including MI MO,
boost ing HSPA peak speeds t o 28 Mbps

Enhanced I MS- based services ( for example, int egrat ed
voice/ mult imedia/ presence/ locat ion)

2010 Evolved EDGE capabilit ies available t o significant ly increase EDGE
t hroughput rat es
HSPA+ peak speeds furt her increased t o peak rat es of 42 Mbps based on
Release 8
LTE int roduced for next - generat ion t hroughput performance using 2X2
MI MO
Advanced core archit ect ures available t hrough EPC/ SAE, primarily for LTE
but also for HSPA+ , providing benefit s such as int egrat ion of mult iple
net work t ypes and flat t er archit ect ures for bet t er lat ency performance
Most new services implement ed in t he packet domain over HSPA+ and LTE
2011 and
lat er
LTE enhancement s such as 4X2 MI MO and 4X4 MI MO
LTE- Advanced specificat ions complet ed
2012 LTE- Advanced pot ent ially deployed in init ial st ages

Over t ime, t he separat e GSM/ EDGE Access Net work ( GERAN) , UMTS Access Net work
( UTRAN) , and core- infrast ruct ure element s will undergo consolidat ion, t hus lowering
t ot al net work cost and improving int egrat ed operat ion of t he separat e access net works.
For act ual users wit h mult imode devices, t he net works t hey access will be largely
t ransparent . Today, nearly all UMTS phones and modems support GSM / EDGE.
Figure 9 present s t he cont inuing advances in HSPA and LTE, plot t ed over t ime, showing
an approximat e doubling of t hroughput per year.

HSPA t o LTE- Advanced, Rysavy Research / 3G Americas, Sept ember 2009 Page 32
Fi gur e 9: Peak Rat es f or Dow nl i nk and Upl i nk ov er Ti me
34


Despit e rapid UMTS deployment , market moment um means t hat even by t he end of t he
decade, most worldwide subscribers will st ill be using GSM. By t hen, however, most new
subscribers will be t aking advant age of UMTS. Only over many years, as subscribers
upgrade t heir equipment , will most net work usage migrat e t o UMTS. Similarly, even as
operat ors st art t o deploy LTE net works at t he end of t his decade and t he beginning of
t he next , it will probably be t he middle of t he next decade before a large percent age of
subscribers are act ually using LTE. During t hese years, most net works and devices will
be t ri- mode—support ing GSM, UMTS, and LTE. The hist ory of wireless- net work
deployment provides a useful perspect ive. GSM, which in 2009 is st ill growing it s
subscriber base, was specified in 1990 wit h init ial net works deployed in 1991. The UMTS
Task Force est ablished it self in 1995, Release 99 specificat ions were complet ed in 2000,
and HSPA+ specificat ions were complet ed in 2007. Alt hough it ’s been more t han a
decade since work began on t he t echnology, only now is UMTS deployment and adopt ion
st art ing t o surge.
Figure 10 shows t he relat ive adopt ion of t echnologies over a mult i- decadal period and
t he lengt h of t ime it t akes for any new t echnology t o be adopt ed widely on a global
basis.

34
Source: A 3G Americas’ member company.
DL R’99-384k
HSDPA 1.8M
HSDPA 3.6M
HSDPA 7.2M
HSDPA 14.4M
MIMO 2x2 28M
MIMO/64QAM 42M
DL LTE(20MHz) 140M
DL LTE(20MHz) 300M
100 kbps
1 M|to
10 Mbps
20 Mbps
100 kbps
1 Mbps
10 Mbps
100 Mbps
2004 2005 2006 2007 2008 2009 2010 2011 2012 2013
UL R’99 384k
HSUPA 1.5M
HSUPA 5.6M
HSUPA/16QAM 11M
UL LTE (10MHz) 25M
UL LTE (10MHz) 50M
• HSPA DL and UL peak throughputs expected
to double every year on average
• Limitations not induced by the technology itself
but time frames required to upgrade
infrastructure and transport networks, obtain
devices with corresponding capabilities and
interoperability tests
Downlink Speeds
Uplink Speeds

HSPA t o LTE- Advanced, Rysavy Research / 3G Americas, Sept ember 2009 Page 33
Fi gur e 10: Rel at i ve Adopt i on of Technol ogi es
35



One opt ion for GSM operat ors t hat have not yet commit t ed t o UMTS, and do not have an
immediat e pressing need t o do so, is t o migrat e direct ly from GSM/ EDGE or Evolved
EDGE t o LTE wit h net works and devices support ing dual- mode GSM- EDGE/ LTE
operat ion.
Deployment Scenarios
There are many different scenarios t hat operat ors will use t o migrat e from t heir current
net works t o fut ure t echnologies such as LTE. Figure 11 present s various scenarios
including operat ors who t oday are using CDMA2000, UMTS, GSM and WiMAX. For
example, as shown in t he first bar, a CMDA2000 operat or in scenario A could in t he
medium t erm deploy a combinat ion of 1xRTT, EV- DO Rev A/ B and LTE and, in t he long
t erm, could migrat e EV- DO dat a t raffic t o LTE. I n scenario B, a CDMA2000 operat or wit h
j ust 1xRTT could int roduce LTE as a broadband service and, in t he long t erm, could
migrat e 1xRTT users t o LTE including voice service.

35
Source: Rysavy Research proj ect ion based on hist orical dat a.
1990 2000 2020 2010
LTE
UMTS/HSPA
R
e
l
a
t
i
v
e

S
u
b
s
c
r
i
p
t
i
o
n
s
GSM/EDGE
2030

HSPA t o LTE- Advanced, Rysavy Research / 3G Americas, Sept ember 2009 Page 34
Fi gur e 11: Di f f er ent Depl oy ment Scenar i os f or LTE
36



3GPP and 3GPP2 bot h have specified det ailed migrat ion opt ions for current 3G syst ems
( UMTS- HSPA and EV- DO) t o LTE. Due t o economies of scale for infrast ruct ure and
devices, 3GPP operat ors are likely t o have a compet it ive cost advant age over 3GPP2
operat ors.
Competing Technologies
Alt hough GSM- HSPA net works are dominat ing global cellular- t echnology deployment s,
operat ors are deploying ot her wireless t echnologies t o serve bot h wide and local areas. This
sect ion of t he paper looks at t he relat ionship bet ween GSM/ UMTS/ LTE and some of t hese
ot her t echnologies.
CDMA2000
CDMA2000, consist ing principally of 1xRTT and One Carrier- Evolved, Dat a- Opt imized
( 1xEV- DO) versions, is t he ot her maj or cellular t echnology deployed in many part s of
t he world. 1xRTT is current ly t he most widely deployed CDMA2000 version. A number of
operat ors have deployed or are deploying 1xEV- DO where a radio carrier is dedicat ed t o
high- speed dat a funct ions. I n June 2009, t here were 106 EV- DO Rel. 0 net works and 62
EV- DO Rev. A net works deployed worldwide.
37

Current ly deployed net work versions are based on eit her Rel. 0 or Rev. A radio- int erface
specificat ions. EV- DO Rev. A incorporat es a more efficient uplink, which has spect ral

36
Source: A 3G Americas’ member company.
37
Source: www. cdg. org, June 5, 2009.

HSPA t o LTE- Advanced, Rysavy Research / 3G Americas, Sept ember 2009 Page 35
efficiency similar t o t hat of HSUPA. Operat ors st art ed t o make EV- DO Rev. A
commercially available in 2007.
EV- DO uses many of t he same t echniques for opt imizing spect ral efficiency as HSPA
including higher order modulat ion, efficient scheduling, t urbo- coding, and adapt ive
modulat ion and coding. For t hese reasons, it achieves spect ral efficiency t hat is virt ually
t he same as HSPA. The 1x t echnologies operat e in t he 1. 25 MHz radio channels,
compared t o t he 5 MHz channels UMTS uses, result ing in lower t heoret ical peak rat es,
alt hough average t hroughput s for high level net work loading are similar. Under low- t o
medium- load condit ions, because of t he lower peak achievable dat a rat es, EV- DO or EV-
DO Rev. A achieves a lower t ypical performance level t han HSPA. Operat ors have quot ed
400 t o 700 kilobit s per second ( kbps) t ypical downlink t hroughput for EV- DO Rev. 0
38

and bet ween 600 kbps and 1. 4 Mbps for EV- DO Rev. A.
39

One challenge for EV- DO operat ors is t hat t hey cannot dynamically allocat e t heir ent ire
spect ral resources bet ween voice and high- speed dat a funct ions. The EV- DO channel is
not available for circuit - swit ched voice, and t he 1xRTT channels offer only medium-
speed dat a. I n t he current st age of t he market , in which dat a only const it ut es a small
percent age of t ot al net work t raffic, t his is not a key issue. But as dat a usage expands,
t his limit at ion will cause subopt imal use of radio resources. Figure 12 illust rat es t his
severe limit at ion.
Fi gur e 12: Radi o Resour ce Management 1x RTT/ 1x EV- DO v er sus UMTS- HSPA

Anot her limit at ion of using a separat e channel for EV- DO dat a services is t hat it
current ly prevent s users from engaging in simult aneous voice and high- speed dat a

38
Source: Verizon Broadband Access Web page, July 29, 2005.
39
Source: Sprint press release January 30, 2007.
T
h
r
e
e

1
.
2
5

M
H
z

C
h
a
n
n
e
l
s
O
n
e

5

M
H
z

C
h
a
n
n
e
l

HSPA t o LTE- Advanced, Rysavy Research / 3G Americas, Sept ember 2009 Page 36
services, whereas t his is possible wit h UMTS and HSPA. Many users enj oy having a
t et hered dat a connect ion from t heir lapt op—by using Bluet oot h, for example—and being
able t o init iat e and receive phone calls while maint aining t heir dat a sessions.
EV- DO will event ually provide voice service using VoI P prot ocols t hrough EV- DO Rev. A,
which includes a higher speed uplink, QoS mechanisms in t he net work, and prot ocol
opt imizat ions t o reduce packet overhead, as well as addressing problems such as j it t er.
Even t hen, however, operat ors will face difficult choices: How many radio channels at
each base st at ion should be made available for 1xRTT t o support legacy t erminals versus
how many radio channels should be allocat ed t o EV- DO. I n cont rast , UMTS allows bot h
circuit - swit ched and packet - swit ched t raffic t o occupy t he same radio channel, where t he
amount of power each uses can be dynamically adj ust ed. This makes it simple t o
migrat e users over t ime from circuit - swit ched voice t o packet - swit ched voice.
Beyond Rev. A, 3GPP2 has defined EV- DO Rev. B, which can combine up t o 15 1. 25 MHz
radio channels in 20 MHz—significant ly boost ing peak t heoret ical rat es t o 73. 5 Mbps.
More likely, an operat or would combine t hree radio channels in 5 MHz. Such an
approach by it self does not necessarily increase overall capacit y, but it does offer users
higher peak- dat a rat es. Ot her enhancement s are planned for EV- DO, including femt ocell
support , MI MO and 64 QAM in t he downlink, and 16 QAM in t he uplink. There are also a
number of planned improvement s for CDMA2000 1xRTT t hat will result in increased
voice capacit y.
3GPP2 had defined an OFDM- based t echnology called Ult ra Mobile Broadband ( UMB) ,
wit h performance charact erist ics similar t o LTE. This work it em, however, has been
t erminat ed as t he st andard had no commercial support , and many CDMA2000 operat ors
including Verizon have announced t heir int ent ions t o migrat e t o LTE. 3GPP2 has defined
t echnical means t o int egrat e CDMA2000 net works wit h LTE along t wo available
approaches:
1. Loose coupling. This involves lit t le or no int er- syst em funct ionalit y, and resources
are released in t he source syst em prior t o handover execut ion.
2. Tight coupling. The t wo syst ems int ercommunicat e wit h net work- cont rolled
make- before- break handovers. Tight coupling allows maint enance of dat a
sessions wit h t he same I P address. This will likely involve a more complex
implement at ion t han loose coupling.
CDMA2000 is clearly a viable and effect ive wireless t echnology and, t o it s credit , many
of it s innovat ions have been brought t o market ahead of compet ing t echnologies.
WiMAX
WiMAX has emerged as a pot ent ial alt ernat ive t o cellular t echnology for wide- area
wireless net works. Based on OFDMA and recent ly accept ed by t he I nt ernat ional
Telecommunicat ions Union ( I TU) as an I MT- 2000 ( 3G t echnology) under t he name
OFDMA TDD WMAN ( Wireless Met ropolit an Area Net work) , WiMAX is t rying t o challenge
exist ing wireless t echnologies—promising great er capabilit ies and great er efficiencies
t han alt ernat ive approaches such as HSPA. But as WiMAX, part icularly mobile WiMAX,
has come closer t o realit y, vendors have cont inued t o enhance HSPA, and perceived
WiMAX advant ages are no longer apparent . I nst ead, WiMAX has gained t he great est
t ract ion in developing count ries as an alt ernat ive t o wireline deployment . I n t he Unit ed
St at es, Clearwire, Sprint Next el and ot hers ( I nt el, Google, Comcast , Time Warner Cable,
and Bright House Net works) have creat ed a j oint vent ure t o deploy a nat ionwide WiMAX
net work. I n June 2009, t his net work was available in At lant a, Balt imore and Port land,

HSPA t o LTE- Advanced, Rysavy Research / 3G Americas, Sept ember 2009 Page 37
OR. Clearwire announced on August 3, 2009, t hat it would launch 10 addit ional market s
on Sept ember 1, 2009.
40

Like GSM- HSPA, WiMAX is not a single t echnology; it is a family of int eroperable
t echnologies. The original specificat ion, I EEE 802. 16, was complet ed in 2001 and
int ended primarily for t elecom backhaul applicat ions in point - t o- point , line- of- sight
configurat ions using spect rum above 10 GHz. This original version of I EEE 802. 16 uses a
radio int erface based on a single- carrier waveform.
The next maj or st ep in t he evolut ion of I EEE 802. 16 occurred in 2004 wit h t he release of
t he I EEE 802. 16- 2004 st andard. I t added mult iple radio int erfaces, including one based
on OFDM- 256 and one based on OFDMA. I EEE 802. 16- 2004 also support s point - t o-
mult ipoint communicat ions, sub- 10 GHz operat ion, and non- line- of- sight
communicat ions. Like t he original version of t he st andard, operat ion is fixed, meaning
t hat subscriber st at ions are t ypically immobile. Pot ent ial applicat ions include wireless
I nt ernet Service Provider ( I SP) service, local t elephony bypass ( as an alt ernat ive t o
cable modem or DSL service) , and cellular backhaul for connect ions from cellular base
st at ions t o operat or infrast ruct ure net works. Vendors can design equipment for eit her
licensed or unlicensed bands.
Vendors are now delivering I EEE 802. 16- 2004- cert ified equipment . This st andard does
not compet e direct ly wit h cellular- dat a and privat e Wi- Fi net works; t hus, it can provide
complement ary services. I n addit ion t o operat or- host ed access solut ions, privat e ent it ies
such as municipal government s, universit ies, and corporat ions will be able t o use t his
version of WiMAX in unlicensed bands ( for example, 5. 8 GHz) for local connect ivit y,
alt hough t here has been lit t le or no development in t his area.
The I EEE has also complet ed a mobile- broadband st andard—I EEE 802. 16e- 2005—t hat
adds mobilit y capabilit ies including support for radio operat ion while mobile, handovers
across base st at ions, and handovers across operat ors. Unlike I EEE 802. 16- 2004, which
operat es in bot h licensed and unlicensed bands, I EEE 802. 16e- 2005 ( referred t o as
mobile WiMAX) makes t he most sense in licensed bands. Operat ors have begun limit ed
mobile WiMAX net work deployment s in 2009. Current WiMAX profiles emphasize TDD
operat ion. Mobile WiMAX net works are not backward- compat ible wit h I EEE 802. 16- 2004
net works.
I nit ial mobile WiMAX net works will be deployed using 2X2 MI MO, TDD and 10 MHz radio
channels in a profile defined by t he WiMAX Forum known as WiMAX Wave 2 or, more
formally, as WiMAX Syst em Profile 1. 0. Beyond Release 1. 0, t he WiMAX Forum has
defined a new profile called WiMAX Release 1.5 wit h product cert ificat ion expect ed by
t he end of 2009. Mobile WiMAX release 1. 5 includes various refinement s int ended t o
improve efficiency and performance and could be available for deployment in a similar
t imeframe as LTE.
Release 1. 5 enhancement s include MAC overhead reduct ions for VoI P ( persist ent
scheduling) , handover opt imizat ions, load balancing, locat ion- based services support ,
FDD operat ion, 64 QAM in t he uplink, downlink adapt ive modulat ion and coding, closed-
loop MI MO ( FDD mode only) , and uplink MI MO.
A subsequent version, Mobile WiMAX 2. 0, will be designed t o address t he performance
requirement s being developed in t he I TU I MT- Advanced Proj ect and will be st andardized

40
Source: Clearwire Press Release, “ Clearwire t o Officially Launch CLEAR 4G Service in 10 Market s on
Sept ember 1, 2009” , August 3rd, 2009

HSPA t o LTE- Advanced, Rysavy Research / 3G Americas, Sept ember 2009 Page 38
in a new I EEE st andard, I EEE 802. 16m. According t o Sprint Next el, I EEE 802. 16m will be
available in 2011.
41

WiMAX employs many of t he same mechanisms as HSPA t o maximize t hroughput and
spect ral efficiency, including high- order modulat ion, efficient coding, adapt ive
modulat ion and coding, and Hybrid Aut omat ic Repeat Request ( HARQ) . The principal
difference from HSPA is I EEE 802. 16e- 2005’s use of OFDMA. As discussed in t he sect ion
“ Technical Approaches ( TDMA, CDMA, OFDMA) ” above, OFDM provides a pot ent ial
implement at ion advant age for wide radio channels ( for example, 10 t o 20 MHz) . I n 5 t o
10 MHz radio channels, t here is no evidence indicat ing t hat WiMAX will have any
performance advant age compared t o HSPA+ .
I t should be not ed, however, t hat I EEE 802. 16e- 2005 cont ains some aspect s t hat may
limit it s performance, part icularly in scenarios in which a sect or cont ains a large number
of mobile users. The performance of t he MAC layer is inefficient when scheduling large
numbers of users, and some aspect s—such as power cont rol of t he mobile st at ion—are
provided using MAC signaling messages rat her t han t he fast power cont rol used in
WCDMA and ot her t echnologies. Thus, while WiMAX uses OFDMA, t he performance will
likely be somewhat less t han HSPA due t o increased overhead and ot her design issues.
Relat ive t o LTE, WiMAX has t he following t echnical disadvant ages: 5 msec frames
inst ead of 1 msec frames, Chase combining inst ead of increment al redundancy, coarser
granularit y for modulat ion and coding schemes and vert ical coding inst ead of horizont al
coding.
42
One deployment considerat ion is t hat TDD requires net work synchronizat ion. I t
is not possible for one cell sit e t o be t ransmit t ing and an adj acent cell sit e t o be
receiving at t he same t ime. Different operat ors in t he same band must eit her coordinat e
t heir net works or have guard bands t o ensure t hat t hey don’t int erfere wit h each ot her.
This may int roduce problems as more operat ors int roduce net works in t he same
spect rum band; for example, t he 2. 5 GHz band in t he Unit ed St at es may be used for
bot h TDD and FDD operat ion.
Alt hough I EEE 802. 16e exploit s significant radio innovat ions similar t o HSPA+ and LTE,
it faces challenges such as economies of scale and t echnology mat urit y. Very few
operat ors t oday have access t o spect rum for WiMAX t hat would permit t hem t o provide
widespread coverage.
I n reference t o economies of scale, GSM- HSPA subscribers number in t he billions. Even
over t he next five years, t he number of WiMAX subscribers is likely t o be quit e low. For
example, I nforma proj ect s 82. 1 million by 2013
43
while Maravedis predict s a lower 75
million WiMAX subscribers by t he end of 2014
44
.
Finally, from a t echnology st andpoint , mobile WiMAX on paper may be slight ly more
capable t han t oday’s available versions of HSPA. But by t he t ime it becomes widely
available, mobile WiMAX will act ually have t o compet e against evolved HSPA syst ems
t hat will offer largely similar capabilit ies. Furt her, by t hen, LTE will not be t hat far from
deployment .

41
Ali Tabassi, Sprint Next el, Fierce Wireless Webcast , “ WiMAX: Mobilizing t he I nt ernet ” , March 5,
2008.
42
I EEE I nt ernat ional Symposium on Personal, I ndoor and Mobile Radio Communicat ions: Anders
Furuskär et al “ The LTE Radio I nt erface – Key Charact erist ics and Performance” , 2008.
43
Source: I nforma WiMAX proj ect ion, June 2009, supplied t o 3G Americas.
44
Source: “ WiMAX and Broadband Wireless Access Equipment Market Analysis, Trends and Forecast s,
2009- 2014, ” Maravedis, June 1, 2009.

HSPA t o LTE- Advanced, Rysavy Research / 3G Americas, Sept ember 2009 Page 39
One specific area in which WiMAX has a t echnical disadvant age is cell size. I n fact , 3G
syst ems have a significant link budget advant age over mobile WiMAX because of soft -
handoff diversit y gain and an FDD duplexing advant age over TDD.
45
Art hur D. Lit t le
report s t hat t he radii of t ypical HSPA cells will be t wo t o four t imes great er t han t ypical
mobile WiMAX cells for high- t hroughput operat ion.
46
One vendor est imat es t hat for t he
same power out put , frequency, and capacit y, mobile WiMAX requires 1. 7 t imes more cell
sit es t han HSPA.
47
Given t hat many real world deployment s of HSPA will occur at
frequencies such as 850 MHz, and LTE at 700 MHz, WiMAX deployment s at 2. 5 GHz will
be at a significant disadvant age.
Wit h respect t o spect ral efficiency, WiMAX is comparable t o HSPA+ , as discussed in t he
sect ion “ Spect ral Efficiency” t hat follows. As for dat a performance, HSPA+ in Release 8—
wit h a peak rat e of 42 Mbps—exceeds mobile WiMAX in 10 MHz in TDD 2: 1 DL: UL using
2X2 MI MO wit h a peak rat e of 40 Mbps.
48
The somet imes- quot ed peak rat e of 63. 4 Mbps
for mobile WiMAX in 10 MHz assumes no bandwidt h applied t o t he uplink.
Some have cit ed int ellect ual propert y right s as an area in which WiMAX has an
advant age. There is lit t le subst ant ial, publicly available informat ion, however, t o support
such claims. First , t he large HSPA vendors have invest ed heavily in t hese t echnologies—
hopefully giving t hem significant leverage wit h which t o negot iat e reasonable int ellect ual
propert y right s ( I PR) rat es wit h ot her vendors. Second, t he mobile WiMAX indust ry is in
it s infancy, and t here is considerable lack of clarit y when it comes t o how different
companies will assert and resolve I PR issues.
Finally, wireless- dat a business models must also be considered. Today’s cellular
net works can finance t he deployment of dat a capabilit ies t hrough a successful voice
business. Building new net works for broadband wireless mandat es subst ant ial capacit y
per subscriber. Consumers who download 1 gigabyt e of dat a each mont h represent a t en
t imes great er load on t he net work t han a 1, 000- minut e- a- mont h voice user. And if t he
fut ure is in mult imedia services such as movie downloads, it is import ant t o recognize
t hat downloading a single DVD- qualit y movie—even wit h advanced compression—
consumes approximat ely 2 gigabyt es. I t is not clear how easily t he available revenue per
subscriber will be able t o finance large- scale deployment of net work capacit y. Despit e
numerous at t empt s, no t errest rial wireless- dat a- only net work has ever succeeded as a
business.
49
Alt hough t here is discussion of providing voice services over WiMAX using
VoI P, mobile- voice users demand ubiquit ous coverage—including indoor coverage.
Mat ching t he cellular foot print wit h WiMAX will require nat ional roaming arrangement s,
complement ed by new dual- t echnology devices or significant operat or invest ment s.

45
Wit h a 2: 1 TDD syst em, t he reverse link only t ransmit s one t hird of t he t ime. To obt ain t he same
cell edge dat a rat es, t he mobile syst em must t ransmit at 4. 77 dB higher t ransmit power.
46
Source: "HSPA and mobile WiMAX for Mobile Broadband Wireless Access", 27 March 2007, Art hur D.
Lit t le Limit ed.
47
Source: Ericsson public whit e paper , “ HSPA, t he undisput ed choice for mobile broadband, May
2007” .
48
Source: Ericsson public whit e paper , “ HSPA, t he undisput ed choice for mobile broadband, May
2007” .
49
Source: Andy Seybold, January 18, 2006, comment ary: “ Will Dat a- Only Net works Ever Make
Money?” ht t p: / / www. out look4mobilit y. com/ comment ary2006/ j an1806. ht m

HSPA t o LTE- Advanced, Rysavy Research / 3G Americas, Sept ember 2009 Page 40
IEEE 802.20
I EEE 802. 20 is a mobile- broadband specificat ion developed by t he Mobile Broadband
Wireless Access Working Group of t he I EEE t hat was complet ed in 2008. Wit h vendors
focused heavily on LTE and WiMAX for next - generat ion wireless services, it is not clear
whet her t here is sufficient moment um in t his st andard t o make it a viable t echnology. At
t his t ime, no operat or has commit t ed t o t he possible st andard. Not e t hat 802. 20 is very
similar t o UMB. However, UMB has been cancelled, and I EEE 802. 20 has not gained any
moment um at t his point in t ime.
Wi-Fi and Municipal Wi-Fi Systems
I n t he local area, t he I EEE 802. 11 family of t echnologies has experienced rapid growt h,
mainly in privat e deployment s. The lat est 802. 11 st andard, 802. 11n offers users
t hroughput s in excess of 100 Mbps and improved range t hrough use of MI MO.
Complement ary st andards increase t he at t ract ion of t he t echnology. 802. 11e provides
qualit y- of- service enabling VoI P and mult imedia, 802. 11i enables robust securit y, and
802. 11r provides fast roaming, necessary for voice handover across access point s.
Leveraging t his success, operat ors—including cellular operat ors—are offering hot spot
service in public areas such as airport s, fast - food rest aurant s, and hot els. For t he most
part , hot spot s are complement ary wit h cellular- dat a net works, because t he hot spot can
provide broadband services in ext remely dense user areas and t he cellular net work can
provide broadband services across much larger areas. Various organizat ions are looking
at int egrat ing WLAN service wit h GSM- HSPA dat a services. The GSM Associat ion has
developed recommendat ions for SI M- based aut hent icat ion of hot spot s, and 3GPP has
mult iple init iat ives t hat address WLAN int egrat ion int o it s net works, including 3GPP
Syst em t o WLAN I nt erworking, UMA, I MS, and EPC/ SAE.
Many cit ies are now deploying met ro Wi- Fi syst ems t hat will provide Wi- Fi access in
downt own areas. These syst ems are based on a mesh t echnology, wherein access point s
forward packet s t o nodes t hat have backhaul connect ions. Alt hough some indust ry
observers are predict ing t hat t hese syst ems will have an adverse effect on 3G dat a
services, met ro Wi- Fi and 3G are more likely t o be complement ary in nat ure. Wi- Fi can
generally provide bet t er applicat ion performance over limit ed coverage areas, whereas
3G syst ems can provide access over much larger coverage areas.
Met ro syst ems t oday are st ill quit e immat ure and face t he following challenges:
 Many cit y proj ect s have been discont inued due t o t he difficult y of providing a
viable business model.
 Today’s mesh syst ems are all propriet ary. The I EEE is developing a mesh
net working st andard—I EEE 802. 11s—but t his may not be ready unt il 2010. Even
t hen, it is not clear t hat vendors will adopt t his st andard for out door syst ems.
 Coverage in most met ro syst ems is designed t o provide an out door signal. As
such, t he signal does not penet rat e many buildings in t he coverage area and
repeat ers are needed t o propagat e t he signal indoors. Many early net work
deployment s have experienced poorer coverage t han init ially expect ed, and t he
number of recommended access point s per square mile has increased st eadily.
 Operat ion is in unlicensed bands in t he 2. 4 GHz radio channel. Given only t hree
relat ively non- overlapping radio channels at 2. 4 GHz, int erference bet ween public
and privat e syst ems is inevit able.

HSPA t o LTE- Advanced, Rysavy Research / 3G Americas, Sept ember 2009 Page 41
 Alt hough mesh archit ect ure simplifies backhaul, t here are st ill considerable
expenses and net working considerat ions in backhauling a large number of
out door access point s.
Nevert heless, met ro net works have at t ract ed considerable int erest , and some number of
proj ect s are st ill proceeding. Technical issues will likely be resolved over t ime, and as
more devices support bot h 3G and Wi- Fi, users can look forward t o mult iple access
opt ions.
Comparison of Wireless Technologies
This sect ion of t he paper compares t he different wireless t echnologies looking at
t hroughput , lat ency, spect ral efficiency, and market posit ion. Finally, t he paper present s a
t able t hat summarizes t he compet it ive posit ion of t he different t echnologies across mult iple
dimensions.
Data Throughput
Dat a t hroughput is an import ant met ric for quant ifying net work t hroughput
performance. Unfort unat ely, t he ways in which various organizat ions quot e t hroughput
st at ist ics vary t remendously. This oft en result s in misleading claims. The int ent of t his
paper is t o realist ically represent t he capabilit ies of t hese t echnologies.
One met hod of represent ing a t echnology’s t hroughput is what people call “ peak
t hroughput ” or “ peak net work speed. ” This refers t o t he fast est possible t ransmission
speed over t he radio link, and it is generally based on t he highest order modulat ion
available and t he least amount of coding ( error correct ion) overhead. Peak net work
speed is also usually quot ed at layer 2 of t he radio link. Because of prot ocol overhead,
act ual applicat ion t hroughput may be 10 t o 20 percent lower ( or more) t han t his layer- 2
value. Even if t he radio net work can deliver t his speed, ot her aspect s of t he net work—
such as t he backhaul from base st at ion t o operat or- infrast ruct ure net work—can oft en
const rain t hroughput rat es t o levels below t he radio- link rat e.
Anot her met hod is t o disclose t hroughput s act ually measured in deployed net works wit h
applicat ions such as File Transfer Prot ocol ( FTP) under favorable condit ions, which
assume light net work loading ( as low as one act ive dat a user in t he cell sect or) and
favorable signal propagat ion. This number is useful because it demonst rat es t he high-
end, act ual capabilit y of t he t echnology. This paper refers t o t his rat e as t he “ peak user
rat e. ” Average rat es, however, are lower t han t his peak rat e and difficult t o predict ,
because t hey depend on a mult it ude of operat ional and net work fact ors. Except when
t he net work is congest ed, however, t he maj orit y of users should experience t hroughput
rat es higher t han one- half of t he peak- achievable rat e.
Some operat ors, primarily in t he US, also quot e t ypical t hroughput rat es. These rat es
are based on t hroughput t est s t he operat ors have done across t heir operat ing net works
and incorporat e a higher level of net work loading. Alt hough t he operat ors do not disclose
t he precise met hodology t hey use t o est ablish t hese figures, t he values provide a good
indicat ion of what users can t ypically expect .
Table 6 present s t he t echnologies in t erms of peak net work t hroughput rat es, peak user-
rat es ( under favorable condit ions) and t ypical rat es. I t omit s values t hat are not yet
known such as t hose associat ed wit h fut ure t echnologies.
The proj ect ed t ypical rat es for HSPA+ and LTE show a wide range. This is because t hese
t echnologies are designed t o exploit favorable radio condit ions t o achieve very high
t hroughput rat es. Under poor radio condit ions, however, t hroughput rat es are lower.

HSPA t o LTE- Advanced, Rysavy Research / 3G Americas, Sept ember 2009 Page 42
Tabl e 6: Thr oughput Per f or mance of Di f f er ent Wi r el ess Technol ogi es
( Bl ue I ndi cat es Theor et i cal Peak Rat es, Gr een Ty pi cal )

Dow nl i nk Upl i nk
Peak
Net w or k
Speed
Peak
and/ or
Ty pi cal
User Rat e
Peak
Net w or k
Speed
Peak
and/ or
Ty pi cal
User Rat e
EDGE ( t y pe 2 MS) 473. 6 kbps 473. 6 kbps
EDGE ( t y pe 1 MS)
( Pr act i cal Ter mi nal )
236. 8 kbps 200 kbps
peak
70 t o 135
kbps t ypical
236. 8 kbps 200 kbps
peak
70 t o 135
kbps t ypical
Evol ved EDGE
( t y pe 1 MS)
50

1184 kbps
51
1 Mbps peak
350 t o 700
kbps t ypical
expect ed
( Dual
Carrier)
473. 6
kbps
52

400 kbps
peak
150 t o 300
kbps t ypical
expect ed
Evol ved EDGE
( t y pe 2 MS)
53

1894. 4
54

kbps
947. 2
kbps
55



UMTS WCDMA Rel ease 99 2. 048 Mbps 768 kbps
UMTS WCDMA Rel ease 99
( Pr act i cal Ter mi nal )
384 kbps 350 kbps
peak
200 t o 300
kbps t ypical
384 kbps 350 kbps
peak
200 t o 300
kbps t ypical
HSDPA I ni t i al Devi ces
( 2006)
1. 8 Mbps > 1 Mbps
peak
384 kbps 350 kbps
peak
HSDPA 14. 4 Mbps 384 kbps
HSPA
56
I ni t i al
I mpl ement at i on
7. 2 Mbps > 5 Mbps 2 Mbps > 1. 5 Mbps

50
A t ype 1 Evolved EDGE MS can receive on up t o t en t imeslot s using t wo radio channels and can
t ransmit on up t o four t imeslot s in one radio channel using 32 QAM modulat ion ( wit h t urbo coding in
t he downlink) .
51
Type 1 mobile, 10 slot s downlink ( dual carrier) , DBS- 12( 118. 4 kbps/ slot ) .
52
Type 1 mobile, 4 slot s uplink, UBS- 12 ( 118. 4 kbps/ slot ) .
53
A t ype 2 Evolved EDGE MS can receive on up t o 16 t imeslot s using t wo radio channels and can
t ransmit on up t o eight t imeslot s in one radio channel using 32 QAM modulat ion ( wit h t urbo coding in
t he downlink) .
54
Type 2 mobile, 16 slot s downlink ( dual carrier) at DBS- 12 ( 118. 4 kbps/ slot ) .
55
Type 2 mobile, 8 slot s uplink, UBS- 12 ( 118. 4 kbps/ slot ) .

HSPA t o LTE- Advanced, Rysavy Research / 3G Americas, Sept ember 2009 Page 43

Dow nl i nk Upl i nk
Peak
Net w or k
Speed
Peak
and/ or
Ty pi cal
User Rat e
Peak
Net w or k
Speed
Peak
and/ or
Ty pi cal
User Rat e
peak
700 kbps t o
1. 7 Mbps
t ypical
57

peak
500 kbps t o
1. 2 Mbps
t ypical
HSPA Cur r ent
I mpl ement at i on
7. 2 Mbps 5. 76 Mbps
HSPA 14. 4 Mbps 5. 76 Mbps
HSPA+ ( DL 64 QAM, UL
16 QAM)
21. 6 Mbps 1. 5 Mbps t o
7 Mbps
13 Mbps
peak
58

11. 5 Mbps 1 Mbps t o
4 Mbps
HSPA+ ( 2X2 MI MO,
DL 16 QAM, UL 16 QAM)
28 Mbps

11. 5 Mbps
HSPA+ ( 2X2 MI MO,
DL 64 QAM, UL 16 QAM)
42 Mbps 11. 5 Mbps
HSPA+ ( 2X2 MI MO,
DL 64 QAM, UL 16 QAM,
Dual Car r i er )
84 Mbps 23 Mbps
LTE ( 2X2 MI MO) 173 Mbps 4 Mbps t o
24 Mbps
( in 2 x 20
MHz)
59

58 Mbps
LTE ( 4X4 MI MO) 326 Mbps 86 Mbps

CDMA2000 1XRTT
153 kbps 130 kbps
peak
153 kbps 130 kbps
peak
CDMA2000 1XRTT 307 kbps 307 kbps
CDMA2000 EV- DO Rel 0
2. 4 Mbps > 1 Mbps
peak
153 kbps 150 kbps
peak
CDMA2000 EV- DO Rev A 3. 1 Mbps > 1. 5 Mbps 1. 8 Mbps > 1 Mbps

56
High Speed Packet Access ( HSPA) consist s of syst ems support ing bot h High Speed Downlink Packet
Access ( HSDPA) and High Speed Uplink Packet Access ( HSUPA) .
57
Typical downlink and uplink t hroughput rat es based on AT&T press release, June 4, 2008
58
Source: Vodafone press release, “ Vodafone Trials HSPA+ Mobile Broadband at Speeds of Up To
16Mbps, ” January 15, 2009.
59
Downlink t hroughput will be about half in a 2 x 10 MHz deployment .

HSPA t o LTE- Advanced, Rysavy Research / 3G Americas, Sept ember 2009 Page 44

Dow nl i nk Upl i nk
Peak
Net w or k
Speed
Peak
and/ or
Ty pi cal
User Rat e
Peak
Net w or k
Speed
Peak
and/ or
Ty pi cal
User Rat e
peak
600 kbps t o
1. 4 Mbps
t ypical
60

peak
300 t o 500
kbps t ypical
CDMA2000 EV- DO Rev B
( 3 r adi o channel s MHz)
9. 3 Mbps 5. 4 Mbps
CDMA2000 EV- DO Rev B
Theor et i cal ( 15 r adi o
channel s)
73. 5 Mbps 27 Mbps


Wi MAX Rel ease 1. 0 ( 10
MHz TDD, DL/ UL= 3, 2x 2
MI MO)
46 Mbps 2 t o 4 Mbps
average
61

4 Mbps
Wi MAX Rel ease 1. 5 TBD TBD
I EEE 802. 16m TBD TBD


HSDPA Throughput in Representative Scenarios

I t is inst ruct ive t o look at act ual HSDPA t hroughput in commercial net works. Figure 13
shows t he t hroughput s measured in one net work wit h voice and dat a in one West ern
European count ry across t hree larger cit ies. The dat a shows t he percent age of samples
on t he X axis t hat fall below t he t hroughput shown on t he Y axis. For example, t he 75
percent ile is at 5 Mbps, meaning t hat 75% of samples are below 5 Mbps and 25% are
above. Significant ly, half of all t he measurement s showed 4 Mbps or higher t hroughput .

60
Typical downlink and uplink t hroughput rat es based on Sprint press release January 30, 2007.
61
Source: Sprint web page, June 2009.

HSPA t o LTE- Advanced, Rysavy Research / 3G Americas, Sept ember 2009 Page 45
Fi gur e 13: HSDPA Thr oughput Di st r i but i on i n Depl oy ed Net w or k s
62



I n anot her net work st udy, Figure 14 shows t he downlink t hroughput performance of a
7. 2 Mbps device ( peak dat a rat e capabilit y) . I t result s in a median t hroughput of 1. 9
Mbps when mobile, 1. 8 Mbps wit h poor coverage, and 3. 8 Mbps wit h good coverage.

62
Source: 3G Americas’ member company cont ribut ion.
0.0
1.0
2.0
3.0
4.0
5.0
6.0
1
0
0
%
9
5
%
9
0
%
8
5
%
8
0
%
7
5
%
7
0
%
6
5
%
6
0
%
5
5
%
5
0
%
4
5
%
4
0
%
3
5
%
3
0
%
2
5
%
2
0
%
1
5
%
1
0
%
5
%
0
%
T
h
r
o
u
g
h
p
u
t

[
M
b
p
s
]

HSPA t o LTE- Advanced, Rysavy Research / 3G Americas, Sept ember 2009 Page 46
Fi gur e 14: HSDPA Per f or mance of a 7.2 Mbps Devi ce i n a Commer ci al Net w or k
63



These rat es are consist ent wit h ot her vendor informat ion for t wo deployed HSPA
net works t hat support ed 7. 2 Mbps HSDPA. Test ers measured average FTP downlink
applicat ion t hroughput of 2. 1 Mbps in t he first net work, and 1. 9 Mbps in t he second
net work.
64

Release 99 and HSUPA Uplink Performance
HSUPA dramat ically increases uplink t hroughput s over 3GPP Release 99. Even Release
99 net works, however, have seen significant uplink increases. Many net works were
init ially deployed wit h a 64 kbps uplink rat e. Lat er, t his increased t o 128 kbps. Lat er
st ill, operat ors increased speeds t o 384 kbps peak rat es wit h peak user- achievable rat es
of 350 kbps.
The ant icipat ed 1 Mbps achievable uplink t hroughput wit h HSUPA can be seen in t he
measured t hroughput of a commercial net work as document ed in Figure 15. The X axis
shows t hroughput rat e, t he Y axis shows t he cumulat ive dist ribut ion funct ion, and t he
bars show t he number of samples obt ained for t hat t hroughput rat e on a relat ive basis.
The median bit rat e is 1. 0 Mbps.

63
Source: 3G Americas’ member company cont ribut ion.
64
Source: 3G Americas’ member company cont ribut ion.
Good Coverage
Median bitrate
3.8 Mbps
Median bitrate
1.8 Mbps
Bad Coverage
Median bitrate
1.8 Mbps
Bad Coverage
-106 dBm
Mobile
Median bitrate
1.9 Mbps
Mobile
Median bitrate
1.9 Mbps

HSPA t o LTE- Advanced, Rysavy Research / 3G Americas, Sept ember 2009 Page 47
Fi gur e 15: Upl i nk Thr oughput i n a Commer ci al Net w or k
65


These rat es are consist ent wit h ot her vendor informat ion for a deployed HSPA net work
t hat support ed 2. 0 Mbps HSUPA
66
uplink speed. Test ers measured average FTP downlink
applicat ion t hroughput of 1. 2 Mbps.
67

One operat or has not ed t hat in it s net works, peak rat es are oft en higher t han t he st at ed
t ypical rat es, because for a large percent age of cells and for a large percent age of t ime,
cells are only light ly loaded.
68

LTE Throughput
As part of t he LTE/ SAE/ EPC Trial I nit iat ive ( LSTI ) , vendors are t est ing LTE t echnology.
Figure 16 shows LTE t hroughput s in a 2X2 MI MO t rial net work reaching a maximum of
154 Mbps, a mean of 78 Mbps and a minimum of 16 Mbps. Unt il operat ors act ually
deploy complet e net works, t ypical rat es will not be available, but t he dat a suggest s t hat
users should be able t o obt ain t hroughput s an order of magnit ude higher t han t oday’s
3G net works.

65
Source: 3G Americas’ member company cont ribut ion.
66
2 x spreading fact or ( 2xSF2) code configurat ion.
67
Source: 3G Americas’ member company cont ribut ion.
68
Source: 3G Americas’ operat or member observat ion for 2009 condit ions.
0
7
0
1
4
0
2
1
0
2
8
0
3
5
0
4
2
0
4
9
0
5
6
0
6
3
0
7
0
0
7
7
0
8
4
0
9
1
0
9
8
0
1
0
5
0
1
1
2
0
1
1
9
0
1
2
6
0
1
3
3
0
1
4
0
0
0
10
20
30
40
50
60
70
80
90
100
Mobile
Median bitrate
1.0 Mbps
0
7
0
1
4
0
2
1
0
2
8
0
3
5
0
4
2
0
4
9
0
5
6
0
6
3
0
7
0
0
7
7
0
8
4
0
9
1
0
9
8
0
1
0
5
0
1
1
2
0
1
1
9
0
1
2
6
0
1
3
3
0
1
4
0
0
0
10
20
30
40
50
60
70
80
90
100
0
7
0
1
4
0
2
1
0
2
8
0
3
5
0
4
2
0
4
9
0
5
6
0
6
3
0
7
0
0
7
7
0
8
4
0
9
1
0
9
8
0
1
0
5
0
1
1
2
0
1
1
9
0
1
2
6
0
1
3
3
0
1
4
0
0
0
10
20
30
40
50
60
70
80
90
100
Mobile
Median bitrate
1.0 Mbps

HSPA t o LTE- Advanced, Rysavy Research / 3G Americas, Sept ember 2009 Page 48
Fi gur e 16: LTE Measur ed Thr oughput i n Test Net w or k
69



Figure 17 provides addit ional insight int o LTE downlink t hroughput , showing layer 1
t hroughput measured at 10 MHz bandwidt h using t he Ext ended Vehicular A 3 km/ hour
channel model. The figure shows t he increased performance obt ained wit h t he addit ion
of different orders of MI MO.



69
Source: 3G Americas’ member company cont ribut ion.
Base station located at x.
L1 Throughput
Max: 154 Mbps
Mean: 78 Mbps
Min: 16 Mbps
User Speed
Max: 45 km/h
Mean: 16 km/h
Min: 0 km/h
Suburban area with line-
of-sight: less than 40%
of the samples
Heights of surrounding
buildings: 15-25 m

12
23
37
54
74
97
123
154
100 meters

HSPA t o LTE- Advanced, Rysavy Research / 3G Americas, Sept ember 2009 Page 49
Fi gur e 17: LTE Thr oughput i n Var i ous Modes
70


For t ypical and average t hroughput s, it is reasonable t o expect an order of magnit ude
higher performance t han HSPA, which one can ant icipat e from radio channels t hat are
four t imes wider ( 20 MHz vs. 5 MHz) and at least a doubling of spect ral efficiency.
Act ual t hroughput rat es t hat users will experience will be lower t han t he peak rat es and
will depend on a variet y of fact ors including:
1. RF Condit ions and User Speed. Peak rat es depend on opt imal condit ions. Under
subopt imal condit ions, such as being at t he edge of t he cell or if t he user is
moving at high speed, t hroughput rat es will be lower.
2. Net work Loading. Like all wireless syst ems, t he t hroughput rat es will go down as
more users simult aneously use t he net work. This is largely a linear degradat ion.
3. Prot ocol Overhead. Peak rat es are generally st at ed for t he physical layer. Due t o
overhead at ot her layers, act ual dat a payload t hroughput rat es may be lower by
an approximat e 5% t o 20% amount . The precise amount depends on t he size of
packet s. Larger packet s ( e. g. , file downloads) result in a lower overhead rat io.
Figure 18 shows how t hroughput rat es can vary by number of act ive users and radio
condit ions. The higher curves are for bet t er radio condit ions.

70
Source: , “ I nit ial Field Performance Measurement s of LTE” , Jonas Karlsson, Mat hias
Riback, Ericsson Review No. 3 2008,
ht t p: / / www. ericsson. com/ ericsson/ corpinfo/ publicat ions/ review/ 2008_03/ files/ LTE. pdf.

HSPA t o LTE- Advanced, Rysavy Research / 3G Americas, Sept ember 2009 Page 50
Fi gur e 18: LTE Act ual Thr oughput Rat es Based on Condi t i ons
71


Verizon Wireless has st at ed t hat it expect s it s LTE net work t o deliver 8 t o 12 Mbps of
t hroughput .
72

Latency
Just as import ant as t hroughput is net work lat ency, defined as t he round- t rip t ime it
t akes dat a t o t raverse t he net work. Each successive dat a t echnology from GPRS forward
reduces lat ency, wit h HSDPA net works having lat ency as low as 70 milliseconds ( msec) .
HSUPA brings lat ency down even furt her, as will 3GPP LTE. Ongoing improvement s in
each t echnology mean all t hese values will go down as vendors and operat ors fine t une
t heir syst ems. Figure 19 shows t he lat ency of different 3GPP t echnologies.

71
Source: LTE/ SAE Trial I nit iat ive, “ Lat est Result s from t he LSTI , Feb 2009, ” www.lst iforum.org.
72
Source: ht t p: / / gigaom. com/ 2009/ 05/ 15/ verizons- lt e- plans- get - real/ .

HSPA t o LTE- Advanced, Rysavy Research / 3G Americas, Sept ember 2009 Page 51
Fi gur e 19: Lat ency of Di f f er ent Technol ogi es
73



The values shown in Figure 19 reflect measurement s of commercially deployed
t echnologies. Some vendors have report ed significant ly lower values in net works using
t heir equipment , such as 150 msec for EDGE, 70 msec for HSDPA, and 50 msec for
HSPA. Wit h furt her refinement s and t he use of 2 msec Transmission Time I nt erval ( TTI )
in t he HSPA uplink, 25 msec roundt rip is a realist ic goal. LTE will reduce lat ency even
furt her, t o as low as 10 msec in t he radio- access net work.
Spectral Efficiency
To bet t er underst and t he reasons for deploying t he different dat a t echnologies and t o
bet t er predict t he evolut ion of capabilit y, it is useful t o examine spect ral efficiency. The
evolut ion of dat a services will be charact erized by an increasing number of users wit h
ever- higher bandwidt h demands. As t he wireless- dat a market grows, deploying wireless
t echnologies wit h high spect ral efficiency will be of paramount import ance. Keeping all
ot her t hings equal such as frequency band, amount of spect rum, and cell sit e spacing,
an increase in spect ral efficiency t ranslat es t o a proport ional increase in t he number of
users support ed at t he same load per user—or, for t he same number of users, an
increase in t hroughput available t o each user. Delivering broadband services t o large
numbers of users can best be achieved wit h high spect ral- efficiency syst ems, especially

73
Source: 3G Americas' member companies. Measured bet ween subscriber unit and Gi int erface,
immediat ely ext ernal t o wireless net work. Does not include I nt er net lat ency. Not e t hat t here is some
variat ion in lat ency based on net work configurat ion and operat ing condit ions.
100
700
600
500
400
300
200
GPRS
Rel’97
EDGE
Rel’99
EDGE
Rel’4
WCDMA
Rel’99
Evolved
EDGE
M
i
l
l
i
s
e
c
o
n
d
s
LTE HSPA HSDPA

HSPA t o LTE- Advanced, Rysavy Research / 3G Americas, Sept ember 2009 Page 52
because t he only ot her alt ernat ives are using more spect rum or deploying more cell
sit es.
I ncreased spect ral efficiency, however, comes at a price. I t generally implies great er
complexit y for bot h user and base st at ion equipment . Complexit y can arise from t he
increased number of calculat ions performed t o process signals or from addit ional radio
component s. Hence, operat ors and vendors must balance market needs against net work
and equipment cost s. One core aspect of evolving wireless t echnology is managing t he
complexit y associat ed wit h achieving higher spect ral efficiency. The reason t echnologies
such as OFDMA are at t ract ive is t hat t hey allow higher spect ral efficiency wit h lower
overall complexit y; t hus t heir use in t echnologies such as LTE and WiMAX.
The roadmap for t he EDGE/ HSPA/ LTE family of t echnologies provides a wide port folio of
opt ions t o increase spect ral efficiency. The exact t iming for deploying t hese opt ions is
difficult t o predict , because much will depend on t he growt h of t he wireless dat a market
and what t ypes of applicat ions become popular.
When det ermining t he best area on which t o focus fut ure t echnology enhancement s, it is
int erest ing t o not e t hat HSDPA, 1xEV- DO, and I EEE 802. 16e- 2005 all have highly
opt imized links—t hat is, physical layers. I n fact , as shown in Figure 20, t he link layer
performance of t hese t echnologies is approaching t he t heoret ical limit s as defined by t he
Shannon bound. ( The Shannon bound is a t heoret ical limit t o t he informat ion t ransfer
rat e [ per unit bandwidt h] t hat can be support ed by any communicat ions link. The bound
is a funct ion of t he Signal t o Noise Rat io [ SNR] of t he communicat ions link. ) Figure 20
also shows t hat HSDPA, 1xEV- DO, and I EEE 802. 16e- 2005 are all wit hin 2 t o 3 decibels
( dB) of t he Shannon bound, indicat ing t hat t here is not much room for improvement
from a link layer perspect ive. Not e t hat differences do exist in t he design of t he MAC
layer ( layer 2) , and t his may result in lower t han expect ed performance in some cases
as described previously.

HSPA t o LTE- Advanced, Rysavy Research / 3G Americas, Sept ember 2009 Page 53
Fi gur e 20: Per f or mance Rel at i ve t o Theor et i cal Li mi t s f or HSDPA, EV- DO, and
I EEE 802.16e- 2005
74



The curves in Figure 20 apply t o an Addit ive Whit e Gaussian Noise Channel ( AWGN) . I f
t he channel is slowly varying and t he effect of frequency select ivit y can be overcome
t hrough an equalizer in eit her HSDPA or OFDM, t hen t he channel can be known almost
perfect ly and t he effect s of fading and non- AWGN int erference can be ignored—t hus
j ust ifying t he AWGN assumpt ion. For inst ance, at 3 km per hour, and fading at 2 GHz,
t he Doppler spread is about 5. 5 Hz. The coherence t ime of t he channel is t hus 1 sec/ 5. 5
or 180 msec. Frames are well wit hin t he coherence t ime of t he channel, because t hey
are t ypically 20 msec or less. As such, t he channel appears “ const ant ” over a frame and
t he Shannon bound applies. Much more of t he t raffic in a cellular syst em is at slow
speeds ( for example, 3 km/ hr) rat her t han at higher speeds. Thus, t he Shannon bound
is relevant for a realist ic deployment environment .
As t he speed of t he mobile st at ion increases and t he channel est imat ion becomes less
accurat e, addit ional margin is needed. This addit ional margin, however, would impact
t he different st andards fairly equally.
The Shannon bound only applies t o a single user; it does not at t empt t o indicat e
aggregat e channel t hroughput wit h mult iple users. I t does indicat e, however, t hat link
layer performance is reaching t heoret ical limit s. As such, t he focus of fut ure t echnology
enhancement s should be on improving syst em performance aspect s t hat maximize t he
experienced SNRs in t he syst em rat her t han on invest igat ing new air int erfaces t hat
at t empt t o improve t he link layer performance.

74
Source: A 3G Americas’ member company.
-15 -10 -5 0 5 10 15 20
0
1
2
3
4
5
6
Required SNR (dB)
A
c
h
i
e
v
a
b
l
e

E
f
f
i
c
i
e
n
c
y

(
b
p
s
/
H
z
)
Shannon bound
Shannon bound with 3dB margin
EV-DO
IEEE 802.16e-2005
HSDPA

HSPA t o LTE- Advanced, Rysavy Research / 3G Americas, Sept ember 2009 Page 54
Examples of t echnologies t hat improve SNR in t he syst em are t hose t hat minimize
int erference t hrough int elligent ant ennas or int erference coordinat ion bet ween sect ors
and cells. Not e t hat MI MO t echniques using spat ial mult iplexing t o pot ent ially increase
t he overall informat ion t ransfer rat e by a fact or proport ional t o t he number of t ransmit
or receive ant ennas do not violat e t he Shannon bound, because t he per- ant enna
t ransfer rat e ( t hat is, t he per- communicat ions link t ransfer rat e) is st ill limit ed by t he
Shannon bound.
Figure 21 compares t he spect ral efficiency of different wireless t echnologies based on a
consensus view of 3G Americas cont ribut ors t o t his paper. I t shows t he cont inuing
evolut ion of t he capabilit ies of all t he t echnologies discussed. The values shown are
conservat ive and int ended t o be reasonably represent at ive of real- world condit ions. Most
simulat ion result s produce values under idealized condit ions; as such, some of t he
values shown are lower ( for all t echnologies) t han t he values indicat ed in ot her papers
and publicat ions. For inst ance, 3GPP st udies indicat e higher HSDPA and LTE spect ral
efficiencies t han t hose shown below.
Fi gur e 21: Compar i son of Dow nl i nk Spect r al Ef f i ci ency
75






75
Joint analysis by 3G Americas’ members. 5+ 5 MHz for UMTS- HSPA/ LTE and CDMA2000, and 10 MHz
DL/ UL= 29: 18 TDD for WiMAX. Mix of mobile and st at ionary users.
0.1
0.7
0.6
0.5
0.4
0.3
0.2
UMTS/HSPA
S
p
e
c
t
r
a
l

E
f
f
i
c
i
e
n
c
y

(
b
p
s
/
H
z
/
s
e
c
t
o
r
)

0.8
0.9
UMTS R’99
HSDPA
EV-DO Rev 0
Rev B
Cross-Carrier
Scheduling
Rev A,
MRxD,
Equalizer
Rel 1.0
2X2 MIMO
CDMA2000 WiMAX
1.4
1.3
1.2
1.1
1.0
LTE
2X2 MIMO
HSPA+
2X2 MIMO
HSPA+
SIC, 64 QAM
HSDPA
MRxD,
Equalizer
1.5
2.1
2.0
1.9
1.8
1.7
1.6
2.2
2.3
2.5
2.4
Future
improvements
LTE
4X2 MIMO
Rel 1.5
2X2 MIMO
Rel 1.5
4X2 MIMO
Future
improvements
Future
improvements
LTE
Future
improvements
LTE
4X4 MIMO

HSPA t o LTE- Advanced, Rysavy Research / 3G Americas, Sept ember 2009 Page 55
The values shown in Figure 21 are not all t he possible combinat ions of available
feat ures. Rat her, t hey are represent at ive milest ones in ongoing improvement s in
spect ral efficiency. For inst ance, t here are t erminals t hat employ mobile- receive
diversit y but not equalizat ion.
The figure does not include EDGE, but EDGE it self is spect rally efficient at 0. 3
bps/ Hz/ sect or. Relat ive t o WCDMA Release 99, HSDPA increases capacit y by almost a
fact or of t hree. Type 3 receivers t hat include Minimum Mean Square Error ( MMSE)
equalizat ion and Mobile Receive Diversit y ( MRxD) will effect ively double HSDPA spect ral
efficiency. HSPA+ in Release 7 includes 2X2 MI MO, which furt her increases spect ral
efficiency by about 20 percent and exceeds WiMAX Release 1. 0 spect ral efficiency.
Met hods like successive int erference cancellat ion ( SI C) and 64 QAM allow gains in
spect ral efficiency as high as 1. 3 bps/ Hz/ sect or, which is close t o LTE performance in
5+ 5 MHz channel bandwidt h. Terminals wit h SI C can also be used wit h Release 7
syst ems. Dual- carrier HSPA will offer a furt her modest gain in spect ral efficiency from
cross- carrier scheduling wit h possible gains of about 10%.
76
Wit h Release 8, operat ors
can deploy eit her MI MO or dual- carrier operat ion. Wit h Release 9, dual- carrier operat ion
can be combined wit h MI MO.
Wit h respect t o act ual deployment , some enhancement s, such as 64 QAM, will be
simpler for some operat ors t o deploy t han ot her enhancement s such as 2X2 MI MO. The
former can be done as a soft ware upgrade, whereas t he lat t er requires addit ional
hardware at t he base st at ion. Thus, t he figure does not necessarily show t he act ual
progression of t echnologies t hat operat ors will deploy t o increase spect ral efficiency.
Beyond HSPA, 3GPP LTE will also result in furt her spect ral efficiency gains, init ially wit h
2X2 MI MO, and t hen opt ionally wit h SI C, 4X2 MI MO and 4X4 MI MO. The gain for 4X2
MI MO is shown at a modest 15% increase for LTE. This assumes a simplified swit ched-
beam approach defined in Release 8. Higher gains are possible wit h more advanced
adapt ive ant enna and beam- forming algorit hms, but are based on propriet ary
implement at ions and, t hus, t he act ual gains will depend on implement at ion. The same is
t rue for WiMAX. Downloadable codebooks in Release 9 LTE provide one avenue for such
addit ional gains.
LTE is even more spect rally efficient wit h wider channels, such as 10 and 20 MHz,
alt hough most of t he gain is realized at 10 MHz.
Similar gains t o t hose for HSPA and LTE are available for CDMA2000. Mobile WiMAX also
experiences gains in spect ral efficiency as various opt imizat ions, like MRxD and MI MO,
are applied. WiMAX Release 1. 0 includes 2X2 MI MO. Enhancement s t o WiMAX will come
wit h Release 1. 5, as well as ot her fut ure enhancement s.
The main reason t hat HSPA+ wit h MI MO is shown as more spect rally efficient t han
WiMAX Release 1. 0 wit h MI MO is because HSPA MI MO support s closed- loop operat ion
wit h precode weight ing and mult i- codeword MI MO, which enables t he use of SI C
receivers. Ot her reasons are t hat HSPA support s increment al- redundancy HARQ, while
WiMAX support s only Chase combining HARQ, and t hat WiMAX has larger cont rol
overhead in t he downlink t han HSPA, because t he uplink in WiMAX is fully scheduled.
OFDMA t echnology requires scheduling t o avoid t wo mobile devices t ransmit t ing on t he
same t ones simult aneously. An uplink MAP zone in t he downlink channel does t his
scheduling.

76
Source: 3G Americas’ member analysis. Vendor est imat es for spect ral- efficiency gains from dual-
carrier operat ion range from 5% t o 20%. Lower spect ral efficiency gains are due t o full- buffer t raffic
assumpt ions. I n more realist ic operat ing scenarios, gains will be significant ly higher.

HSPA t o LTE- Advanced, Rysavy Research / 3G Americas, Sept ember 2009 Page 56
LTE has higher spect ral efficiency t han WiMAX Wave 2 for a number of reasons
77
:
- Closed- loop operat ion wit h precoded weight ing.
- Mult i- codeword MI MO, which enables t he use of SI C receivers.
- Lower Channel Qualit y I ndicat or delay t hrough use of 1 msec frames inst ead of 5
msec frames.
- Great er cont rol channel efficiency.
- I ncrement al redundancy in error correct ion.
- Finer granularit y of modulat ion and coding schemes.
WiMAX Release 1. 5 addresses some of t hese it ems and will t hus have increased spect ral
efficiency. Expect ed feat ures include reduced MAC overhead, adapt ive modulat ion and
coding, and ot her physical- layer enhancement s.
One available improvement for LTE spect ral efficiency not shown in t he figure is
successive int erference cancellat ion. This will result in a gain of 5% in a low- mobilit y
environment and a gain of 10 t o 15% in environment s such as picocells in which t here is
cell isolat ion.
The following t able summarizes t he most import ant feat ures of LTE and WiMAX
t echnology t hat impact spect ral efficiency.
Tabl e 7: LTE and Wi MAX Feat ur es
Feat ur e LTE Wi MAX
Rel ease 1.0
Wi MAX
Rel ease 1.5
I mpact
Mul t i pl e
Access
OFDM in
downlink,
Discr et e Four ier
Transform ( DFT) -
spread OFDM in
uplink
OFDM in downlink
and uplink
OFDM in downlink
and uplink
DFT- spread OFDM reduces
t he peak- t o- average
power rat io and reduces
t er minal complexit y,
requir es one- t ap equalizer
in base st at ion receiver.
Upl i nk Pow er
Cont r ol
Fract ional pat h-
loss
compensat ion
Full pat h- loss
compensat ion
Full pat h- loss
compensat ion
Fract ional pat h- loss
compensat ion enables
flexible t radeoff bet ween
average and cell- edge
dat a rat es
Schedul i ng Channel
dependent in
t ime and
frequency
domains
Channel dependent
in t ime domain
Channel dependent
in t ime and
frequency domains
Access t o t he frequency
domain yields larger
scheduling gains
MI MO Scheme Mult i- codewor d
( horizont al) ,
closed loop wit h
pre- coding
Single codewor d
( vert ical)
Single codewor d
( vert ical) , wit h
rank- adapt ive
MI MO ( TDD) and
wit h closed- loop
pre- coding ( FDD)
Horizont al encoding
enables per- st r eam link
adapt at ion and successive
int erference cancellat ion
receivers.
Modul at i on
and Codi ng
Scheme
Gr anul ar i t y
Fine granular it y
( 1- 2 dB apart )
Coarse granularit y
( 2- 3 dB apart )
Coarse granularit y
( 2- 3 db apart )
Finer granular it y enables
bet t er link adapt at ion
precision.

77
I EEE I nt ernat ional Symposium on Personal, I ndoor and Mobile Radio Communicat ions: Anders
Furuskär et al “ The LTE Radio I nt erface – Key Charact erist ics and Performance” , 2008.

HSPA t o LTE- Advanced, Rysavy Research / 3G Americas, Sept ember 2009 Page 57
Feat ur e LTE Wi MAX
Rel ease 1.0
Wi MAX
Rel ease 1.5
I mpact
Hy br i d
Aut omat i c
Repeat
Request
( ARQ)
I ncrement al
redundancy
Chase combining Chase combining I ncrement al redundancy is
more efficient ( lower SNR
requir ed for given error
rat e)
Fr ame
Dur at i on
1 msec
subframes
5 msec subframes 5 msec subframes Short er subframes yield
lower user plane delay and
reduced channel qualit y
feedback delays
Over head /
Cont r ol
Channel
Ef f i ci ency
Relat ively low
overhead
Relat ively high
overhead
Relat ively high
overhead apar t
from reduct ion in
pilot s
Lower overhead improves
performance

Figure 22 compares t he uplink spect ral efficiency of t he different syst ems.
Fi gur e 22: Compar i son of Upl i nk Spect r al Ef f i ci ency
78






78
Joint analysis by 3G Americas’ members. 5+ 5 MHz for UMTS- HSPA/ LTE and CDMA2000, and 10 MHz
DL/ UL= 29: 18 TDD for WiMAX. Mix of mobile and st at ionary users.
0.1
0.7
0.6
0.5
0.4
0.3
0.2
S
p
e
c
t
r
a
l

E
f
f
i
c
i
e
n
c
y

(
b
p
s
/
H
z
/
s
e
c
t
o
r
)

0.8
0.9
UMTS R’99
to Rel 5
HSUPA Rel 6
LTE 1X2
Receive
Diversity
EV-DO
Rev 0
EV-DO Rev B,
Interference
Cancellation
EV-DO
Rev A
HSPA+
Interference
Cancellation,
16 QAM
LTE 1x4
Receive
Diversity
Future
Improvements
Future
Improvements
1.0
UMTS/HSPA CDMA2000 WiMAX LTE
Rel
1.0
Future
Improvements
Rel 1.5 1X2
Rx Div
Rel 1.5
1X4
Receive
Diversity

HSPA t o LTE- Advanced, Rysavy Research / 3G Americas, Sept ember 2009 Page 58
The implement at ion of HSUPA in HSPA significant ly increases uplink capacit y, as does
Rev. A and Rev. B of 1xEV- DO, compared t o Rel. 0. OFDM- based syst ems can exhibit
improved uplink capacit y relat ive t o CDMA t echnologies, but t his improvement depends
on fact ors such as t he scheduling efficiency and t he exact deployment scenario. Wit h
LTE, spect ral efficiency gains increase by use of receive diversit y. I nit ial syst ems will
employ 1X2 receive diversit y ( t wo ant ennas at t he base st at ion) and lat er 1X4 diversit y,
which should increase spect ral efficiency by 50%. I t is also possible t o employ Mult i- User
MI MO ( MU- MI MO) , which allows simult aneous t ransmission by mult iple users on t he
uplink on t he same physical resource t o increase spect ral efficiency and is, in fact , easier
t o implement t han t rue MI MO, because it does not require an addit ional t ransmit t er in
t he mobile device. Spect ral efficiency gains, however, wit h MU- MI MO are not as great as
wit h t he receive diversit y schemes.
Figure 23 compares voice spect ral efficiency. I t assumes a round- robin t ype of
scheduler, as opposed t o a proport ional- fair scheduler t hat is normally used for
asynchronous dat a.
Fi gur e 23: Compar i son of Voi ce Spect r al Ef f i ci ency
79



Figure 23 shows UMTS Release 99 wit h bot h AMR 12. 2 kbps and 7. 95 kbps vocoders.
The AMR 12. 2 kbps vocoder provides superior voice qualit y in good ( e. g. , st at ic, indoors)
channel condit ions. UMTS has dynamic adapt at ion bet ween vocoder rat es, enabling
enhanced voice qualit y compared t o EVRC at t he expense of capacit y in sit uat ions t hat
are not capacit y limit ed.

79
Source: Joint analysis by 3G Americas’ members. 10 + 10 MHz for UMTS- HSPA/ LTE and
CDMA2000, and 20 MHz DL/ UL= 29: 18 TDD for WiMAX. Mix of mobile and st at ionary users.
50
350
300
250
200
150
100
E
r
l
a
n
g
s
,


1
0
+
1
0

M
H
z
UMTS R’99
AMR 7.95 kbps
Interference
Cancellation
AMR 5.9 kbps
1xRTT
EVRC 8 kbps
Interference
Cancellation
EVRC-B 6 kbps
Rel 1.5
EVRC-B
6kbps
EV-DO Rev A
EVRC 8 kbps
Rel 7, VoIP
AMR 7.95 kbps
500
450
400
LTE AMR 5.9 kbps
Rel 7 VoIP
AMR 5.9 kbps
Future
Improvements
Future
Improvements
EVRC-B 6 kbps
Future
Improvements
Rel 1.0
EVRC
8 kbps
LTE VoIP
AMR 7.95 kbps
UMTS R’99
AMR 12.2 kbps
UMTS/HSPA CDMA2000 WiMAX LTE
Future
Improvements

HSPA t o LTE- Advanced, Rysavy Research / 3G Americas, Sept ember 2009 Page 59
Opport unit ies will arise t o improve voice capacit y using VoI P over HSPA channels.
Depending on t he specific enhancement s implement ed, voice capacit y could double over
exist ing circuit - swit ched syst ems. I t should be not ed, however, t hat t he gains are not
relat ed specifically t o t he use of VoI P; rat her, gains relat e t o advances in radio
t echniques applied t o t he dat a channels. Many of t hese same advances may also be
applied t o current circuit - swit ched modes. However, ot her benefit s of VoI P are driving
t he migrat ion t o packet voice. Among t hese benefit s are a consolidat ed I P core net work
for operat ors and sophist icat ed mult imedia applicat ions for users.
There are a number of planned improvement s for CDMA2000 1xRTT t hat will result in
increased voice capacit y. EV- DO t echnologies could possibly exhibit a slight ly higher
spect ral efficiency for VoI P t han HSPA t echnologies ( alt hough not for packet dat a in
general) , as t hey operat e purely in t he packet domain and do not have circuit - swit ched
cont rol overhead.
80
Unt il VoI P over EV- DO becomes available, HSPA will have t he
significant advant age, however, of being able t o support simult aneous circuit - swit ched
and packet - swit ched users on t he same radio channel. I f adj acent carriers are available,
seven CDMA2000 carriers can be deployed in 10 MHz of spect rum, providing an
addit ional gain of 12%.
Wit h respect t o codecs, in VoI P syst ems such as LTE and WiMAX, a variet y of codecs can
be used. The figures show performance assuming specific codecs at represent at ive bit
rat es. For codecs such as EVRC ( Enhanced Variable Rat e Codec) , t he bit rat e shown is
an average value.
WiMAX voice spect ral efficiency is shown at 180 Erlangs for Release 1. 0 and 210 Erlangs
for Release 1. 5. A spect ral efficiency gain of 50% is available by changing t he DL: UL
rat io from 29: 18 t o 23: 24, since now 18 dat a symbols per frame are allocat ed for t he UL
compared t o 12. A furt her gain of 15% is available t hrough t he use of persist ent
scheduling and changing t he DL: UL from 23: 24 t o 20: 27.
81
Changing t his rat io,
however, may not be pract ical if t he same carrier frequency must support bot h voice
and dat a. Alt ernat ively, voice and dat a may be placed on different carriers using
different TDD rat ios.
Cost, Volume and Market Comparison
So far, t his paper has compared wireless t echnologies on t he basis of t echnical capabilit y
and demonst rat ed t hat many of t he different opt ions have similar t echnical at t ribut es.
This is for t he simple reason t hat t hey employ many of t he same approaches.
There is a point of comparison, however, in which t he differences bet ween t he
t echnologies diverge t remendously; namely, t he difference in volume involved including
subscribers and t he amount of infrast ruct ure required. This difference should t ranslat e t o
dramat ically reduced cost s for t he highest volume solut ions, specifically GSM- HSPA.
Based on proj ect ions and numbers already present ed in t his paper, 3G subscribers on
UMTS net works will number in t he many hundreds of millions by t he end of t his decade,
whereas subscribers t o emerging wireless t echnologies, such as WiMAX, will number in
t he t ens of millions. See Figure 24 for det ails.

80
Transmit Power Cont rol ( TPC) bit s on t he uplink Dedicat ed Physical Cont rol Channel DPCCH in UMTS
R’99. See also I EEE Journal on Select ed Areas in Communicat ion, Vol 24, No. 1, Qi Bi, “ An Analysis of
VoI P Service Using 1 EV- DO Revision A Syst em” , January, 2006.
81
Source: I EEE Communicat ions Magazine, Mo- Han Fong and Robert Novak, Nort el Net works, Sean
McBeat h, Huawei Technologies, Roshni Srinivasan, I nt el Corporat ion, “ I mproved VoI P Capacit y in
Mobile WiMAX Syst ems Using Persist ent Resource Allocat ion, ” Oct ober, 2008.

HSPA t o LTE- Advanced, Rysavy Research / 3G Americas, Sept ember 2009 Page 60
Fi gur e 24: Rel at i ve Vol ume of Subscr i ber s Acr oss Wi r el ess Technol ogi es
82



I n t he chart above, t he small “ Ot her” cat egory represent s bot h WiMAX and LTE. I nforma
proj ect ions on HSPA, LTE and WiMAX in millions of subscribers are as follows:
83

2008 2009 2010 2011 2012 2013 2014
WiMAX 0. 5 2. 8 7. 5 16. 7 37. 1 82. 1
LTE 0. 0 0. 0 0. 5 3. 5 13. 1 44. 5 131. 5
HSPA 304 438 649 957 1400 2000 2700
By mid next decade, it is highly likely t hat LTE will mat ch WiMAX subscript ions and by
t he end of t he decade, will significant ly exceed t hem. Ovum st at es “ By 2014, LTE will
have 109 million connect ions worldwide. I n comparison, mobile WiMAX will have almost
55 million connect ions. This is in st ark cont rast t o 2013, when parit y bet ween t he t wo
t echnologies is expect ed. ”
84

Alt hough proponent s for t echnologies such as mobile WiMAX point t o lower cost s for
t heir alt ernat ives, t here doesn’t seem t o be any inherent cost advant age—even on an
equal- volume basis. And when fact oring in t he lower volumes, any real- world cost
advant age is debat able.

82
Source: I nforma Telecoms & Media, WCI S+ , June 2009.
83
Source: I nforma WiMAX and LTE proj ect ions, June 2009 supplied t o 3G Americas.
84
Source: Ovum, Telecom and Soft ware News, July 2, 2009.

1,000 
2,000 
3,000 
4,000 
5,000 
6,000 
7,000 
Dec 09 Dec 10 Dec 11 Dec 12 Dec 13 Dec 14
M
i
l
l
i
o
n
s
UMTS‐HSPA
GSM
CDMA
Other
6.3 Billion Total  
Subscriptions in 
2014
Subscriptions
4.6 
Billion
Total
2.7 B
2.7 B
3.7 B
3.9 B
4.0 B
3.8 B
3.4 B
2.0 B
1.4 B
957 M
649 M
438 M

HSPA t o LTE- Advanced, Rysavy Research / 3G Americas, Sept ember 2009 Page 61
From a deployment point of view, t he t ype of t echnology used ( for example, HSPA
versus WiMAX) only applies t o t he soft ware support ed by t he digit al cards at t he base
st at ion. This cost , however, is only a small fract ion of t he base st at ion cost wit h t he
balance covering ant ennas, power amplifiers, cables, racks, RF cards. As for t he rest of
t he net work including const ruct ion, backhaul, and core- net work component s, cost s are
similar regardless of Radio Access Net work ( RAN) t echnology. Spect rum cost s for each
t echnology can differ great ly depending on a count ry’s regulat ions and t he spect rum
band. As a general rule in most part s of t he world, spect rum sold at 3. 5 GHz will cost
much less t han spect rum sold at 850 MHz ( all ot her t hings being equal) .
As for UMTS- HSPA versus CDMA2000, higher deployment —by a fact or of five—could
t ranslat e t o significant cost savings. For example, research and development
amort izat ion result s in a four- t o- one difference in base st at ion cost s.
85
Similarly, j ust as
GSM handset s are considered much less expensive t han 1xRTT handset s, UMTS- HSPA
wholesale t erminal prices could be t he market leader in low- cost or mass- market 3G
t erminals. Development s such as single- chip UMTS complement ary met al oxide
semiconduct or ( CMOS) t ransceivers could be part icularly effect ive in making
UMTS/ HSDPA devices more affordable t o t he mass market .
86

Even LTE is on t he road t o a robust wireless ecosyst em and significant economies of
scale. I n June of 2008, t he Next Generat ion Mobile Net works ( NGMN) alliance confirmed
it s select ion of LTE. Dr. Pet er Meissner, Operat ing Officer of NGMN announced t hat
“ based on int ensive and det ailed t echnology evaluat ions, 3GPP LTE/ SAE is t he first
t echnology which broadly meet s it s recommendat ions and is approved by it s Board. ”
87

The NGMN is comprised of 18 mobile net work operat ors, 29 vendor sponsors and 3
Universit y research inst it ut es. I t s operat or members include: Allt el, AT&T, China Mobile,
France Telecom, Royal KPN, MSV Mobile Sat ellit e Vent ures, NTT DoComo, Reliance
Communicat ions, SK Telecom, Telecom I t alia, Telefonica, Telenor, TeliaSonera, Telst ra,
Telus, T- Mobile and Vodafone.
Competitive Summary
Based on t he informat ion present ed in t his paper,
Table 8 summarizes t he compet it ive posit ion of t he different t echnologies discussed.

Tabl e 8: Compet i t i ve Posi t i on of Maj or Wi r el ess Technol ogi es
Technol ogy EDGE/ HSPA/ LTE CDMA2000 Wi MAX
Subscr i ber s Over 3. 7 billion
t oday; 4 billion
expect ed by 2010
455 million
88

t oday; slower
growt h expect ed
t han GSM- HSPA
82 million
ant icipat ed by 2013
Mat ur i t y Ext remely mat ure Ext remely mat ure Emerging/ immat ure


85
Source: 3G Americas’ member analysis.
86
Source: Qualcomm press release Feb 13, 2007.
87
ht t p: / / www. umt s- forum. org/ cont ent / view/ 2479/ 172/ .
88
Source: CDG, June 2009 for Q4 2008.

HSPA t o LTE- Advanced, Rysavy Research / 3G Americas, Sept ember 2009 Page 62
Technol ogy EDGE/ HSPA/ LTE CDMA2000 Wi MAX
Adopt i on Cellular operat ors
globally
Cellular operat ors
globally.
Limit ed t o dat e
Cov er age/ Foot pr i nt Global Global wit h t he
general except ion
of West ern Europe
Very limit ed
Depl oy ment Fewer cell sit es
required at 700
and 850 MHz.
Fewer cell sit es
required at 700
and 850 MHz.
Many more cell
sit es required at 2. 5
GHz.
Dev i ces Broad select ion of
GSM/ EDGE/ UMTS/
HSPA devices
Broad select ion of
1xRTT/ EV- DO
devices
I nit ial devices
emphasize dat a
Radi o Technol ogy Highly opt imized
TDMA for EDGE,
highly opt imized
CDMA for HSPA,
highly opt imized
OFDMA for LTE
Highly opt imized
CDMA for
Rev 0/ A/ B,
highly opt imized
OFDMA for Rev C
Opt imized OFDMA
in Release 1. 0.
More opt imized in
Release 1. 5
Spect r al Ef f i ci ency Very high wit h
HSPA, mat ches
OFDMA approaches
in 5 MHz wit h
HSPA+
Very high wit h EV-
DO Rev A/ B
Very high, but not
higher t han HSPA+
for Release 1. 0, and
not higher t han LTE
for Release 1. 5
Thr oughput
Capabi l i t i es
Peak downlink
user- achievable
rat es of over 4
Mbps t oday, wit h
significant ly higher
rat es in t he fut ure
Peak downlink
user- achievable
rat es of over 1. 5
Mbps, wit h
significant ly higher
rat es in t he fut ure
2 t o 4 Mbps
average, 12 Mbps
peak
89

Voi ce Capabi l i t y Ext remely efficient
circuit - voice
available t oday;
smoot hest
migrat ion t o VoI P
of any t echnology
Ext remely efficient
circuit - voice
available t oday
EV- DO radio
channels wit h VoI P
cannot support
circuit - voice users
Relat ively inefficient
VoI P init ially; more
efficient in lat er
st ages, but lower
t han LTE.
Voice coverage will
be much more
limit ed t han cellular
Si mul t aneous Voi ce
and Dat a
Available wit h
GSM
90
and UMTS
t oday
Not available t oday
Available wit h VoI P
Pot ent ially
available, t hough
init ial services will
emphasize dat a


89
Sprint web page, June 2009.
90
Wit h t he applicat ion of Dual Transfer Mode.

HSPA t o LTE- Advanced, Rysavy Research / 3G Americas, Sept ember 2009 Page 63
Technol ogy EDGE/ HSPA/ LTE CDMA2000 Wi MAX
Ef f i ci ent Spect r um
Usage
Ent ire UMTS radio
channel available
for any mix of
voice and high-
speed dat a
Radio channel
t oday limit ed t o
eit her
voice/ medium
speed dat a or high-
speed dat a only
Current ly only
efficient for dat a-
cent ric net works

Conclusion
Through const ant innovat ion, t he EDGE/ HSPA/ LTE family of t echnologies has proven it self
as t he predominant wireless net work solut ion and offers operat ors and subscribers a t rue
mobile- broadband advant age. The cont inued use of GSM and EDGE t echnology t hrough
ongoing enhancement s allows operat ors t o leverage exist ing invest ment s. Wit h UMTS- HSPA,
t he t echnologies’ advant ages provide for broadband services t hat will deliver increased dat a
revenue and provide a pat h t o all- I P archit ect ures. Wit h LTE, now t he most widely chosen
t echnology plat form for t he fort hcoming decade and wit h deployment imminent , t he
advant ages offer a best - of- breed, long- t erm solut ion t hat mat ches or exceeds t he
performance of compet ing approaches. I n all cases, t he different radio- access t echnologies
can coexist using t he same core archit ect ure.
Today, HSPA offers t he highest peak dat a rat es of any widely available, wide- area wireless
t echnology. Wit h cont inued evolut ion, peak dat a rat es will cont inue t o increase, spect ral
efficiency will improve, and lat ency will decrease. The result is support for more users wit h
more support ed applicat ions. The scope of applicat ions will also increase as new services
t hrough st andardized net work int erfaces become available such as locat ion informat ion,
video, and call cont rol. Great er efficiencies and capabilit ies t ranslat e t o more compet it ive
offers, great er net work usage, and increased revenues.
Because of pract ical benefit s and deployment moment um, t he migrat ion pat h from EDGE t o
HSPA t hen t o LTE is proving inevit able. Benefit s include t he abilit y t o roam globally, huge
economies of scale, widespread accept ance by operat ors, complement ary services such as
messaging and mult imedia, and an ast onishing variet y of compet it ive handset s and ot her
devices. Current ly more t han 264 commercial UMTS- HSPA net works are already in
operat ion. UMTS- HSPA and/ or LTE offer an excellent migrat ion pat h for GSM operat ors, as
well as an effect ive t echnology solut ion for greenfield operat ors.
EDGE has proven t o be a remarkably effect ive and efficient t echnology for GSM net works. I t
achieves high spect ral efficiency and dat a performance t hat t oday support s a wide range of
applicat ions. Evolved EDGE will great ly enhance EDGE capabilit ies—doubling and,
pot ent ially, quadrupling t hroughput s, as well as halving lat ency and increasing spect ral
efficiency—making t he t echnology viable for many years t o come.
Whereas EDGE is efficient for narrowband dat a services, t he UMTS- HSPA radio link is
efficient for wideband services. Unlike some compet ing t echnologies, UMTS t oday offers
users simult aneous voice and dat a. I t also allows operat ors t o support voice and dat a across
t heir ent ire available spect rum.
HSPA has significant ly enhanced UMTS by providing a broadband dat a service wit h user-
achievable rat es t hat oft en exceed 1 Mbps on t he downlink in init ial deployment s and t hat
now exceed 4 Mbps in some commercial net works. Many net works are now being upgraded
t o include HSUPA providing users wit h uplink rat es in excess of 1 Mbps.

HSPA t o LTE- Advanced, Rysavy Research / 3G Americas, Sept ember 2009 Page 64
Not only expect ed cont inual improvement s in radio t echnology, but improvement s t o t he
core net work t hrough flat t er archit ect ures—part icularly EPC/ SAE—will reduce lat ency, speed
applicat ions, simplify deployment , enable all services in t he I P domain, and allow a common
core net work t o support bot h LTE and legacy GSM- HSPA syst ems.
Ot her innovat ions, such as MI MO and higher order modulat ion are now being deployed.
Evolved HSPA+ syst ems, wit h peak rat es of 42 Mbps, will largely mat ch t he t hroughput and
capacit y of OFDMA- based approaches in 5 MHz, 3GPP adopt ed OFDMA wit h 3GPP LTE, which
will provide a growt h plat form for t he next decade.
Wit h t he cont inued growt h in mobile comput ing, powerful new handheld- comput ing
plat forms, an increasing amount of mobile cont ent , mult imedia messaging, mobile
commerce, and locat ion services, wireless dat a has slowly, but inevit ably, become a huge
indust ry. EDGE/ HSPA/ LTE provides one of t he most robust port folios of mobile- broadband
t echnologies, and it is an opt imum framework for realizing t he pot ent ial of t his market .

HSPA t o LTE- Advanced, Rysavy Research / 3G Americas, Sept ember 2009 Page 65
Appendix: Technology Details
The EDGE/ HSPA/ LTE family of dat a t echnologies provides ever- increasing capabilit ies t hat
support ever more demanding applicat ions. EDGE, now available globally, already makes a
wealt h of applicat ions feasible including ent erprise applicat ions, messaging, e- mail, Web
browsing, consumer applicat ions, and even some mult imedia applicat ions. Wit h UMTS and
HSPA, users are enj oying videophones, high- fidelit y music, richer mult imedia applicat ions,
and efficient access t o t heir ent erprise applicat ions.
I t is import ant t o underst and t he needs ent erprises and consumers have for t hese services.
The obvious needs are broad coverage and high dat a t hroughput . Less obvious for users,
but as crit ical for effect ive applicat ion performance, are t he needs for low lat ency, QoS
cont rol, and spect ral efficiency. Spect ral efficiency, in part icular, is of paramount concern,
because it t ranslat es t o higher average t hroughput s ( and t hus more responsive
applicat ions) for more act ive users in a coverage area. The discussion below, which
examines each t echnology individually, det ails how t he progression from EDGE t o HSPA t o
LTE is one of increased t hroughput , enhanced securit y, reduced lat ency, improved QoS, and
increased spect ral efficiency.
I t is also helpful t o specifically not e t he t hroughput requirement s necessary for different
applicat ions:
 Microbrowsing ( for example, Wireless Applicat ion Prot ocol [ WAP] ) : 8 t o 128 kbps
 Mult imedia messaging: 8 t o 64 kbps
 Video t elephony: 64 t o 384 kbps
 General- purpose Web browsing: 32 kbps t o more t han 1 Mbps
 Ent erprise applicat ions including e- mail, dat abase access, and VPNs: 32 kbps t o
more t han 1 Mbps
 Video and audio st reaming: 32 kbps t o 2 Mbps
Not e t hat EDGE already sat isfies t he demands of many applicat ions. Wit h HSPA, applicat ions
operat e fast er and t he range of support ed applicat ions expands even furt her.
Under favorable condit ions, EDGE delivers peak user- achievable t hroughput rat es close t o
200 kbps and init ial deployment s of HSPA deliver peak user- achievable downlink t hroughput
rat es of well over 1 Mbps, easily meet ing t he demands of many applicat ions. Lat ency has
cont inued t o improve, t oo, wit h HSPA net works t oday having round- t rip t imes as low as 70
msec. The combinat ion of low lat ency and high t hroughput t ranslat es t o a broadband
experience for users in which applicat ions are ext remely responsive.
I ncreasingly, devices will be mult i- modal support ing mult iple t ypes of wireless t echnologies.
Users equipped wit h such mult imode devices may, t herefore, be grant ed quit e different
levels of connect ivit y ranging from a dense urban environment where t hey may obt ain t he
lat est wireless t echnology t o slower speeds in a rural net work deployment or when roaming
in a visit ed net work. I n t hese cases, users will benefit from knowing what service level t o
expect such as from indicat ions on t he device screen. These are current ly available at a
rudiment ary level ( e. g. , 2G vs. 3G) , but fut ure improvement s will enable display of
addit ional det ails ( e. g. , Evolved EDGE vs. EDGE, HSUPA) . I n t his sect ion, we consider
different t echnical approaches for wireless and t he parallel evolut ion of 3GPP t echnologies.
We t hen provide det ails on EDGE, UMTS- HSPA, HSPA+ , LTE, and support ing t echnologies
such as I MS.

HSPA t o LTE- Advanced, Rysavy Research / 3G Americas, Sept ember 2009 Page 66
EDGE/EGPRS
Today, most GSM net works support EDGE. I t is an enhancement applicable t o GPRS,
which is t he original packet dat a service for GSM net works, as well as t o GSM circuit -
swit ched services, t he lat t er not being considered furt her in t his document . GPRS
provides a packet - based I P connect ivit y solut ion support ing a wide range of ent erprise
and consumer applicat ions. GSM net works wit h EDGE operat e as wireless ext ensions t o
t he I nt ernet and give users I nt ernet access, as well as access t o t heir organizat ions from
anywhere. Wit h peak user- achievable
91
t hroughput rat es of up t o 200 kbps wit h EDGE
using four t imeslot devices, users have t he same effect ive access speed as a modem,
but wit h t he convenience of connect ing from anywhere.
To underst and t he evolut ion of dat a capabilit y, we briefly examine how t hese dat a
services operat e, beginning wit h t he archit ect ure of GSM and EDGE, as depict ed in
Figure 25.
Fi gur e 25: GSM/ GPRS/ EDGE Ar chi t ect ur e

EDGE is essent ially t he addit ion of a packet - dat a infrast ruct ure t o GSM. I n fact , t his
same dat a archit ect ure is preserved in UMTS and HSPA net works, and it is t echnically
referred t o as GPRS for t he core- dat a funct ion in all t hese net works. The t erm GPRS may
also be used t o refer t o t he init ial radio int erface, now supplant ed by EDGE. Funct ions of
t he dat a element s are as follows:
1. The base st at ion cont roller direct s/ receives packet dat a t o/ from t he SGSN, an
element t hat aut hent icat es and t racks t he locat ion of mobile st at ions.
2. The SGSN performs t he t ypes of funct ions for dat a t hat t he MSC performs for
voice. Each serving area has one SGSN, and it is oft en collocat ed wit h t he MSC.
3. The SGSN forwards/ receives user dat a t o/ from t he GGSN, which can be viewed
as a mobile I P rout er t o ext ernal I P net works. Typically, t here is one GGSN per
ext ernal net work ( for example, t he I nt ernet ) . The GGSN also manages I P
addresses, dynamically assigning t hem t o mobile st at ions for t heir dat a sessions.

91
“ Peak user- achievable” means users, under favorable condit ions of net work loading and signal
propagat ion, can achieve t his rat e as measured by applicat ions such as file t ransfer. Average rat es
depend on many fact ors and will be lower t han t hese rat es.
Public Switched
Telephone Network
External Data
Network (e.g., Internet)
Base
Station
Controller
Base
Transceiver
Station
Base
Transceiver
Station
Mobile
Switching
Center
Home
Location
Register
Serving
GPRS
Support
Node
Gateway
GPRS
Support
Node
IP
Traffic
Circuit-Switched
Traffic
Mobile
Station
Mobile
Station
Mobile
Station
GPRS/EDGE Data
Infrastructure

HSPA t o LTE- Advanced, Rysavy Research / 3G Americas, Sept ember 2009 Page 67
Anot her import ant element is t he HLR, which st ores users’ account informat ion for bot h
voice and dat a services. Of significance is t hat t his same dat a archit ect ure support s dat a
services in GSM and in UMTS- HSPA net works, t hereby simplifying operat or net work
upgrades.
I n t he radio link, GSM uses radio channels of 200 kilohert z ( kHz) widt h, divided in t ime
int o eight t imeslot s comprising 577 microseconds ( us) t hat repeat every 4. 6 msec, as
shown in Figure 26. The net work can have mult iple radio channels ( referred t o as
t ransceivers) operat ing in each cell sect or. The net work assigns different funct ions t o
each t imeslot such as t he Broadcast Cont rol Channel ( BCCH) , circuit - swit ched funct ions
like voice calls or dat a calls, t he opt ional Packet Broadcast Cont rol Channel ( PBCCH) ,
and packet dat a channels. The net work can dynamically adj ust capacit y bet ween voice
and dat a funct ions, and it can also reserve minimum resources for each service. This
enables more dat a t raffic when voice t raffic is low or, likewise, more voice t raffic when
dat a t raffic is low, t hereby maximizing overall use of t he net work. For example, t he
PBCCH, which expands t he capabilit ies of t he normal BCCH, may be set up on a t imeslot
of a TDMA frame when j ust ified by t he volume of dat a t raffic.
Fi gur e 26: Ex ampl e of GSM/ EDGE Ti mesl ot St r uct ur e
92


EDGE offers close coupling bet ween voice and dat a services. I n most net works, while in
a dat a session, users can accept an incoming voice call, which suspends t he dat a
session, and t hen resume t heir dat a session aut omat ically when t he voice session ends.
Users can also receive SMS messages and dat a not ificat ions
93
while on a voice call. Wit h
net works support ing DTM, users wit h DTM- capable devices can engage in simult aneous
voice/ dat a operat ion.
Wit h respect t o dat a performance, each dat a t imeslot can deliver peak user- achievable
dat a rat es of up t o about 50 kbps. The net work can aggregat e up t o four of t hese
t imeslot s on t he downlink wit h current devices.
I f mult iple dat a users are act ive in a sect or, t hey share t he available dat a channels. As
demand for dat a services increases, however, an operat or can accommodat e cust omers
by assigning an increasing number of channels for dat a service t hat is limit ed only by
t hat operat or’s t ot al available spect rum and radio planning.

92
Source: 3G Americas’ member company cont ribut ion.
93
Example: WAP not ificat ion message delivered via SMS.
BCCH TCH TCH TCH TCH PDTCH PDTCH PDTCH
0 1 2 3 4 5 6 7
577 S
per timeslot
4.615 ms per frame of 8 timeslots
Possible BCCH
carrier configuration
PBCCH TCH TCH PDTCH PDTCH PDTCH PDTCH PDTCH
0 1 2 3 4 5 6 7
Possible TCH carrier
configuration
BCCH: Broadcast Control Channel – carries synchronization, paging and other signalling information
TCH: Traffic Channel – carries voice traffic data; may alternate between frames for half-rate
PDTCH: Packet Data Traffic Channel – carries packet data traffic for GPRS and EDGE
PBCCH: Packet Broadcast Control Channel – additional signalling for GPRS/EDGE; used only if needed

HSPA t o LTE- Advanced, Rysavy Research / 3G Americas, Sept ember 2009 Page 68
EDGE is an official 3G cellular t echnology t hat can be deployed wit hin an operat or' s
exist ing 850, 900, 1800, and 1900 MHz spect rum bands. EDGE capabilit y is now largely
st andard in new GSM deployment s. A GPRS net work using t he EDGE radio int erface is
t echnically called an Enhanced GPRS ( EGPRS) net work, and a GSM net work wit h EDGE
capabilit y is referred t o as GERAN. EDGE has been an inherent part of GSM
specificat ions since Release 99. I t is fully backward- compat ible wit h older GSM
net works, meaning t hat GPRS devices work on EDGE net works and t hat GPRS and EDGE
t erminals can operat e simult aneously on t he same t raffic channels. I n addit ion, any
applicat ion developed for GPRS will work wit h EDGE.
Many operat ors t hat originally planned t o use only UMTS for next - generat ion dat a
services have deployed EDGE as a complement ary 3G t echnology.
I t is import ant t o not e t hat EDGE t echnology is cont inuing t o improve. For example,
Release 4 significant ly reduced EDGE lat ency ( net work round- t rip t ime) —from t he
t ypical 500 t o 600 msec t o about 300 msec. Operat ors also cont inue t o make
improvement s in how EDGE funct ions including net work opt imizat ions t hat boost
capacit y and reduce lat ency. The impact for users is t hat EDGE net works t oday are more
robust wit h applicat ions funct ioning more responsively. Release 7’s Evolved EDGE also
int roduces significant new feat ures.
Devices t hemselves are increasing in capabilit y. Dual Transfer Mode ( DTM) devices,
already available from vendors, allow simult aneous voice and dat a communicat ions. For
example, during a voice call, users will be able t o ret rieve e- mail, do mult imedia
messaging, browse t he Web, and do I nt ernet conferencing. This is part icularly useful
when connect ing phones t o lapt ops via cable or Bluet oot h and using t hem as modems.
DTM is a 3GPP- specified t echnology t hat enables new applicat ions like video sharing
while providing a consist ent service experience ( service cont inuit y) wit h UMTS. Typically,
a DTM end- t o- end solut ion requires only a soft ware upgrade t o t he GSM/ EDGE radio
net work. There are a number of net works and devices now support ing DTM.
Evolved EDGE
Recognizing t he value of t he huge inst alled base of GSM net works, 3GPP has worked t o
improve EDGE capabilit ies for Release 7. This work was part of t he GERAN Evolut ion
effort , which also includes voice enhancement s not discussed in t his paper.
Alt hough EDGE t oday already serves many applicat ions like wireless e- mail ext remely
well, it makes good sense t o cont inue t o evolve EDGE capabilit ies. From an economic
st andpoint , it is less cost ly t han upgrading t o UMTS, because most enhancement s are
designed t o be soft ware based, and it is highly asset - efficient , because it involves fewer
long- t erm capit al invest ment s t o upgrade an exist ing syst em. Wit h 85 percent of t he
world market using GSM, which is already equipped for simple roaming and billing, it is
easy t o offer global service t o subscribers. Evolved EDGE offers higher dat a rat es and
syst em capacit y, and reduced lat ency, and cable- modem speeds are realist ically
achievable.
I n addit ion, many regions do not have licensed spect rum for deployment of a new radio
t echnology such as UMTS- HSPA or LTE. Also, Evolved EDGE provides bet t er service
cont inuit y bet ween EDGE and HSPA or LTE, meaning t hat a user will not have a hugely
different experience when moving bet ween environment s, for example when a LTE user

HSPA t o LTE- Advanced, Rysavy Research / 3G Americas, Sept ember 2009 Page 69
moves t o a GSM/ Evolved EDGE net work t o est ablish a ( circuit - swit ched) voice call
94
or
when leaving LTE coverage.
Alt hough GSM and EDGE are already highly opt imized t echnologies, advances in radio
t echniques will enable furt her efficiencies. Some of t he obj ect ives of Evolved EDGE
included:
 A 100 percent increase in peak dat a rat es.
 A 50 percent increase in spect ral efficiency and capacit y in C/ I - limit ed scenarios.
 A sensit ivit y increase in t he downlink of 3 dB for voice and dat a.
 A reduct ion of lat ency for init ial access and round- t rip t ime, t hereby enabling
support for conversat ional services such as VoI P and PoC.
 To achieve compat ibilit y wit h exist ing frequency planning, t hus facilit at ing
deployment in exist ing net works.
 To coexist wit h legacy mobile st at ions by allowing bot h old and new st at ions t o
share t he same radio resources.
 To avoid impact s on infrast ruct ure by enabling improvement s t hrough a soft ware
upgrade.
 To be applicable t o DTM ( simult aneous voice and dat a) and t he A/ Gb mode
int erface. The A/ Gb mode int erface is part of t he 2G core net work, so t his goal is
required for full backward- compat ibilit y wit h legacy GPRS/ EDGE.
The met hods st andardized in Release 7 t o achieve or surpass t hese obj ect ives include:
 Downlink dual- carrier recept ion t o double t he number of t imeslot s t hat can be
received for a 100 percent increase in t hroughput .
 The addit ion of Quadrat ure Phase Shift Keying ( QPSK) , 16 QAM and 32 QAM, as
well as an increased symbol rat e ( 1. 2x) and a new set of modulat ion/ coding
schemes t hat will increase maximum t hroughput per t imeslot by up t o 100
percent ( EGPRS2- B) . Current ly, EDGE uses 8- PSK modulat ion.
 A reduct ion in overall lat ency. This is achieved by lowering t he TTI t o 10 msec
and by including t he acknowledgement informat ion in t he dat a packet . These
enhancement s will have a dramat ic effect on t hroughput for many applicat ions.
 Downlink diversit y recept ion of t he same radio channel t o increase t he robust ness
in int erference and t o improve t he receiver sensit ivit y. Simulat ions have
demonst rat ed sensit ivit y gains of 3 dB and a decrease in required C/ I of up t o 18
dB for a single co- channel int erferer. Significant increases in syst em capacit y can
be achieved, as explained below.
Dual - Car r i er Recei ver
A key part of t he evolut ion of EDGE is t he ut ilizat ion of more t han one radio frequency
carrier. This overcomes t he inherent limit at ion of t he narrow channel bandwidt h of GSM.
Using t wo radio- frequency carriers requires t wo receiver chains in t he downlink, as
shown in Figure 27. Using t wo carriers enables t he recept ion of t wice ( or more t han
t wice for some mult i- slot classes) as many radio blocks simult aneously.

94
Some init ial LTE net works will be dat a- only, wit h voice operat ion provided by GSM.

HSPA t o LTE- Advanced, Rysavy Research / 3G Americas, Sept ember 2009 Page 70
Fi gur e 27: Ev ol v ed EDGE Tw o- Car r i er Oper at i on
95



Alt ernat ively, t he original number of radio blocks can be divided bet ween t he t wo
carriers. This eliminat es t he need for t he net work t o have cont iguous t imeslot s on one
frequency.
Fi gur e 28: EDGE Mul t i - Car r i er Recei ve Logi c – Mobi l e Par t
96


Channel capacit y wit h dual- carrier recept ion improves great ly, not by increasing basic
efficiencies of t he air int erface, but because of st at ist ical improvement in t he abilit y t o
assign radio resources, which increases t runking efficiency.
As net work loading increases, it is st at ist ically unlikely t hat cont iguous t imeslot s will be
available. Wit h t oday’s EDGE devices, it is not possible t o change radio frequencies when

95
Source: 3G Americas’ member company cont ribut ion.
96
Source: 3G Americas’ member company cont ribut ion.
Rx1
Tx (1)
Neighbor Cell Measurements
Uplink Timeslot
Downlink Timeslot
Slot N
Slot N + 1
(Idle Frame)
Slot N + 2 Slot N + 3
Rx2

HSPA t o LTE- Advanced, Rysavy Research / 3G Americas, Sept ember 2009 Page 71
going from one t imeslot t o t he next . Wit h an Evolved EDGE dual receiver, however, t his
becomes possible, t hus enabling cont iguous t imeslot s across different radio channels.
The result is t hat t he syst em can allocat e a larger set of t ime slot s for dat a even if t hey
are not cont iguous, which ot herwise is not possible. Figure 29 shows why t his is
import ant . As t he net work becomes busy, t he probabilit y of being assigned 1 t imeslot
decreases. As t his probabilit y decreases ( X axis) , t he probabilit y of being able t o obt ain
5 t imeslot s on t he same radio carrier decreases dramat ically. Being able t o obt ain
t imeslot s across t wo carriers in Evolved EDGE, however, significant ly improves t he
likelihood of obt aining t he desired t imeslot s.
Fi gur e 29: Pr obabi l i t i es of Ti me Sl ot Assi gnment s
97



Mobi l e St at i on Recei v e Di ver si t y
Figure 30 illust rat es how mobile- st at ion receive diversit y increases syst em capacit y.
( BCCH refers t o t he Broadcast Cont rol Channel and TCH refers t o t he Traffic Channel. )
The BCCH carrier repeat s over 12 cells in a 4/ 12 frequency reuse pat t ern, which requires
2. 4 MHz for GSM. A fract ionally loaded syst em may repeat f12 t hrough f15 on each of
t he cells. This is a 1/ 1 frequency reuse pat t ern wit h higher syst em ut ilizat ion, but also
pot ent ially high co- channel int erference in loaded condit ions.

97
Source: 3G Americas’ member company cont ribut ion.

HSPA t o LTE- Advanced, Rysavy Research / 3G Americas, Sept ember 2009 Page 72
Fi gur e 30: Ex ampl e of 4/ 12 Fr equency Reuse w i t h 1/ 1 Ov er l ay
98


I n t oday’s EDGE syst ems, f12 t hrough f15 in t he 1/ 1 reuse layer can only be loaded t o
around 25 percent of capacit y. Thus, wit h four of t hese frequencies, it is possible t o
obt ain 100 percent of t he capacit y of t he frequencies in t he 4/ 12 reuse layer or t o
double t he capacit y by adding 800 KHz of spect rum.
Using Evolved EDGE and receive- diversit y- enabled mobile devices t hat have a high
t olerance t o co- channel int erference, however, it is possible t o increase t he load on t he
1/ 1 layer from 25 t o 50 percent and possibly t o as high as 75 percent . An increase t o 50
percent t ranslat es t o a doubling of capacit y on t he 1/ 1 layer wit hout requiring any new
spect rum and t o a 200 percent gain compared t o a 4/ 12 reuse layer.
Hi gher Or der Modul at i on and Hi gher Symbol Rat e Schemes
The addit ion of higher order modulat ion schemes enhances EDGE net work capacit y wit h
lit t le capit al invest ment by ext ending t he range of t he exist ing wireless t echnology. More
bit s- per- symbol means more dat a t ransmit t ed per unit t ime. This yields a fundament al
t echnological improvement in informat ion capacit y and fast er dat a rat es. Use of higher
order modulat ion exploit s localized opt imal coverage circumst ances, t hereby t aking
advant age of t he geographical locat ions associat ed wit h probabilit ies of high C/ I rat io
and enabling very high dat a t ransfer rat es whenever possible.
These enhancement s are only now being considered, because fact ors such as processing
power, variabilit y of int erference, and signal level made higher order modulat ions
impract ical for mobile wireless syst ems j ust a few years ago. Newer t echniques for
demodulat ion, however, such as advanced receivers and receive diversit y, help enable
t heir use.
Two different levels of support for higher order modulat ion are defined for bot h t he
uplink and t he downlink: EGPRS2- A and EGPRS2- B. I n t he uplink, EGPRS2- A level
includes GMSK, 8- PSK, and 16 QAM at t he legacy symbol rat e. This level of support
reuses Modulat ion and Coding Schemes ( MCSs) 1 t hrough 6 from EGPRS and adds five
new 16 QAM modulat ed schemes called uplink EGPRS2- A schemes ( UAS) .


98
Source: 3G Americas’ member company cont ribut ion.
0 1
2
3 4
5
6 7
8
9 10
11
Example of a 4/12 frequency reuse pattern used for BCCH
carriers with a 1/1 frequency reuse pattern for TCH carriers.
BCCH carriers on
f0 - f11 are associated
with TCH carrier
frequencies f12 – f15
f12
f13
f14
f15
f12
f13
f14
f15
f12
f13
f14
f15

HSPA t o LTE- Advanced, Rysavy Research / 3G Americas, Sept ember 2009 Page 73
Tabl e 9: Upl i nk Modul at i on and Codi ng Schemes
Modul at i on
and Codi ng
Scheme
Name
Upl i nk EGPRS2 Suppor t Level A
Modul at i on
Ty pe
Peak Thr oughput ( k bps) –
4 sl ot s
MCS- 1 GMSK 35. 2
MCS- 2 GMSK 44. 8
MCS- 3 GMSK 59. 2
MCS- 4 GMSK 70. 4
MCS- 5 8- PSK 89. 6
MCS- 6 8- PSK 118. 4
UAS- 7 16 QAM 179. 2
UAS- 8 16 QAM 204. 8
UAS- 9 16 QAM 236. 8
UAS- 10 16 QAM 268. 8
UAS- 11 16 QAM 307. 2
The second support level in t he uplink includes QPSK, 16 QAM, and 32 QAM modulat ion
as well as a higher ( 1. 2x) symbol rat e. MCSs 1 t hrough 4 from EGPRS are reused, and
eight new uplink EGPRS2- B schemes ( UBS) are added.
Tabl e 10: Upl i nk Modul at i on and Codi ng Schemes w i t h Hi gher Symbol Rat e
Modul at i on
and Codi ng
Scheme
Name
Upl i nk EGPRS2 Suppor t Level B
Modul at i on
Ty pe
Peak Thr oughput ( k bps)
– 4 sl ot s
MCS- 1 GMSK 35. 2
MCS- 2 GMSK 44. 8
MCS- 3 GMSK 59. 2
MCS- 4 GMSK 70. 4
UBS- 5 QPSK 89. 6
UBS- 6 QPSK 118. 4
UBS- 7 16 QAM 179. 2
UBS- 8 16 QAM 236. 8
UBS- 9 16 QAM 268. 8
UBS- 10 32 QAM 355. 2
UBS- 11 32 QAM 435. 2
UBS- 12 32 QAM 473. 6
The first downlink support level int roduces a modified set of 8- PSK coding schemes and
adds 16 QAM, and 32 QAM all at t he legacy symbol rat e. Turbo codes are used for all

HSPA t o LTE- Advanced, Rysavy Research / 3G Americas, Sept ember 2009 Page 74
new modulat ions. MCSs 1 t hrough 4 are reused, and eight new downlink EGPRS2- A
level schemes ( DAS) are added.

Tabl e 11: Dow nl i nk Modul at i on and Codi ng Schemes
Modul at i on
and Codi ng
Scheme
Name
Dow nl i nk HOM/ HSR Suppor t Level A
Modul at i on
Ty pe
Peak Thr oughput ( k bps) –
4 sl ot s
MCS- 1 GMSK 35. 2
MCS- 2 GMSK 44. 8
MCS- 3 GMSK 59. 2
MCS- 4 GMSK 70. 4
DAS- 5 8- PSK 89. 6
DAS- 6 8- PSK 108. 8
DAS- 7 8- PSK 131. 2
DAS- 8 16 QAM 179. 2
DAS- 9 16 QAM 217. 6
DAS- 10 32 QAM 262. 4
DAS- 11 32 QAM 326. 4
DAS- 12 32 QAM 393. 6

The second downlink support level includes QPSK, 16 QAM, and 32 QAM modulat ions at
a higher ( 1. 2x) symbol rat e. MCSs 1 t hrough 4 are reused, and eight new downlink
EGPRS2- B level schemes ( DBS) are defined.












HSPA t o LTE- Advanced, Rysavy Research / 3G Americas, Sept ember 2009 Page 75
Tabl e 12: Dow nl i nk Modul at i on and Codi ng Schemes w i t h Hi gher Symbol Rat e
99

Modul at i on
and Codi ng
Scheme
Name
Dow nl i nk HOM/ HSR Suppor t Level B
Modul at i on
Ty pe
Peak Thr oughput ( k bps) –
4 sl ot s
MCS- 1 GMSK 35. 2
MCS- 2 GMSK 44. 8
MCS- 3 GMSK 59. 2
MCS- 4 GMSK 70. 4
DBS- 5 QPSK 89. 6
DBS- 6 QPSK 118. 4
DBS- 7 16 QAM 179. 2
DBS- 8 16 QAM 236. 8
DBS- 9 16 QAM 268. 8
DBS- 10 32 QAM 355. 2
DBS- 11 32 QAM 435. 2
DBS- 12 32 QAM 473. 6

The combinat ion of Release 7 Evolved EDGE enhancement s shows a dramat ic pot ent ial
increase in t hroughput . For example, in t he downlink, a Type 2 mobile device ( one t hat
can support simult aneous t ransmission and recept ion) using DBS- 12 as t he MCS and a
dual- carrier receiver can achieve t he following performance:
Highest dat a rat e per t imeslot ( layer 2) = 118. 4 kbps
Timeslot s per carrier = 8
Carriers used in t he downlink = 2
Tot al downlink dat a rat e = 118. 4 kbps X 8 X 2 = 1894. 4 kbps
100

This t ranslat es t o a peak net work rat e close t o 2 Mbps and a user- achievable dat a rat e
of well over 1 Mbps!
Ev ol v ed EDGE I mpl ement at i on
Table 13 shows what is involved in implement ing t he different feat ures defined for
Evolved EDGE. For a number of feat ures, t here are no hardware changes required for
t he base t ransceiver st at ion ( BTS) . For all feat ures, Evolved EDGE is compat ible wit h
legacy frequency planning.

99
These dat a r at es require a wide- pulse shaping filt er t hat is not part of Release 7.
100
For t he near fut ure, t wo carriers will be a scenario more pract ically realized on a not ebook
comput er plat form t han handheld plat forms.

HSPA t o LTE- Advanced, Rysavy Research / 3G Americas, Sept ember 2009 Page 76
Tabl e 13: Ev ol v ed EDGE I mpl ement at i on
101



I n conclusion, it is int erest ing t o not e t he sophist icat ion and capabilit y t hat is achievable
wit h, and planned for, by GSM.
UMTS-HSPA Technology
UMTS has garnered t he overwhelming maj orit y of new 3G spect rum licenses wit h 283
commercial net works already in operat ion.
102
Compared t o emerging wireless
t echnologies, UMTS t echnology is mat ure and benefit s from research and development
t hat began in t he early 1990s. I t has been t horoughly t rialed, t est ed, and commercially
deployed. UMTS deployment is now accelerat ing wit h st able net work infrast ruct ures and
at t ract ive, reliable mobile devices t hat have rich capabilit ies. Wit h t he addit ion of HSPA
for high- speed packet dat a services, UMTS- HSPA is quickly emerging as t he dominant
global mobile- broadband net work.
UMTS employs a wideband CDMA radio- access t echnology. The primary benefit s of
UMTS include high spect ral efficiency for voice and dat a, simult aneous voice and dat a
capabilit y for users, high user densit ies t hat can be support ed wit h low infrast ruct ure
cost s, support for high- bandwidt h dat a applicat ions, and a clean migrat ion t o VoI P in t he
fut ure. Operat ors can also use t heir ent ire available spect rum for bot h voice and high-
speed dat a services.
Addit ionally, operat ors can use a common core net work t hat support s mult iple radio-
access net works including GSM, EDGE, WCDMA, HSPA, and evolut ions of t hese
t echnologies. This is called t he UMTS mult i- radio net work, and it gives operat ors
maximum flexibilit y in providing different services across t heir coverage areas ( see
Figure 31) .

101
Source: 3G Americas’ member company cont ribut ion.
102
Source: I nforma Telecoms & Media, “ World Cellular I nformat ion Service, ” June 2009.
 

HSPA t o LTE- Advanced, Rysavy Research / 3G Americas, Sept ember 2009 Page 77
Fi gur e 31: UMTS Mul t i - r adi o Net w or k

The UMTS radio- access net work consist s of base st at ions referred t o as Node B
( corresponding t o GSM base t ransceiver syst ems) t hat connect t o RNCs ( corresponding
t o GSM base st at ion cont rollers [ BSCs] ) . The RNCs connect t o t he core net work as do
t he BSCs. When bot h GSM and WCDMA access net works are available, t he net work can
hand over users bet ween t hese net works. This is import ant for managing capacit y, as
well as in areas in which t he operat or has cont inuous GSM coverage, but has only
deployed WCDMA in some locat ions.
Whereas GSM can effect ively operat e like a spread- spect rum syst em
103
, based on t ime
division in combinat ion wit h frequency hopping, WCDMA is a direct - sequence, spread-
spect rum syst em. WCDMA is spect rally more efficient t han GSM, but it is t he wideband
nat ure of WCDMA t hat provides it s great est advant age—t he abilit y t o t ranslat e t he
available spect rum int o high dat a rat es. This wideband t echnology approach result s in
t he flexibilit y t o manage mult iple t raffic t ypes including voice, narrowband dat a, and
wideband dat a.
WCDMA allocat es different codes for different channels, whet her for voice or dat a, and it
can adj ust t he amount of capacit y, or code space, of each channel every 10 msec wit h
WCDMA Release 99 and every 2 msec wit h HSPA. WCDMA creat es high- bandwidt h t raffic
channels by reducing t he amount of spreading ( using a short er code) wit h WCDMA
Release 99 and higher- order modulat ion schemes for HSPA. Packet dat a users can share
t he same codes as ot her users, or t he net work can assign dedicat ed channels t o users.
To furt her expand t he number of effect ively operat ing applicat ions, UMTS employs a
sophist icat ed QoS archit ect ure for dat a t hat provides four fundament al t raffic classes
including:
1. Conv er sat i onal . Real- t ime int eract ive dat a wit h cont rolled bandwidt h and
minimum delay such as VoI P or video conferencing.
2. St r eami ng. Cont inuous dat a wit h cont rolled bandwidt h and some delay such as
music or video.

103
Spread spect rum syst ems can eit her be direct sequence or frequency hopping.
UMTS
Core Network
(MSC, HLR,
SGSN, GGSN)
GSM/EDGE
WCDMA,
HSDPA
Other
e.g., WLAN
Radio-Access Networks External Networks
Packet-Switched
Networks
Circuit-Switched
Networks
Other Cellular
Operators

HSPA t o LTE- Advanced, Rysavy Research / 3G Americas, Sept ember 2009 Page 78
3. I nt er act i ve. Back- and- fort h dat a wit hout bandwidt h cont rol and some delay
such as Web browsing.
4. Back gr ound. Lower priorit y dat a t hat is non- real- t ime such as bat ch t ransfers.
This QoS archit ect ure, available t hrough all HSPA versions, involves negot iat ion and
priorit izat ion of t raffic in t he radio- access net work, t he core net work, and t he int erfaces
t o ext ernal net works such as t he I nt ernet . Consequent ly, applicat ions can negot iat e QoS
paramet ers on an end- t o- end basis bet ween a mobile t erminal and a fixed- end syst em
across t he I nt ernet or privat e int ranet s. This capabilit y is essent ial for expanding t he
scope of support ed applicat ions, part icularly mult imedia applicat ions including
packet ized video t elephony and VoI P.
UMTS Release 99 Data Capabilities
I nit ial UMTS net work deployment s were based on 3GPP Release 99 specificat ions, which
included voice and dat a capabilit ies. Since t hen, Release 5 has defined HSDPA and
Release 6 has defined HSUPA. Wit h HSPA- capable devices, t he net work uses HSPA
( HSDPA/ HSUPA) for dat a. Operat ors wit h Release 99 net works are upgrading t hem t o
HSPA capabilit y. I n advance of Release 6, t he uplink in HSDPA ( Release 5) net works
uses t he Release 99 approach.
I n UMTS Release 99, t he maximum t heoret ical downlink rat e is j ust over 2 Mbps.
Alt hough exact t hroughput depends on t he channel sizes t he operat or chooses t o make
available, t he capabilit ies of devices and t he number of users act ive in t he net work limit
t he peak t hroughput rat es a user can achieve t o about 350 kbps in commercial
net works. Peak downlink net work speeds are 384 kbps. Uplink peak- net work t hroughput
rat es are also 384 kbps in newer deployment s wit h user- achievable peak rat es of 350
kbps.
104
This sat isfies many communicat ions- orient ed applicat ions.
Channel t hroughput s are det ermined by t he amount of channel spreading. Wit h more
spreading, as in voice channels, t he dat a st ream has great er redundancy, and t he
operat or can employ more channels. I n comparison, a high- speed dat a channel has less
spreading and fewer available channels. Voice channels use downlink spreading fact ors
of 128 or 256, whereas a 384 kbps dat a channel uses a downlink spreading fact or of 8.
The commonly quot ed rat e of more t han 2 Mbps downlink t hroughput for UMTS can be
achieved by combining t hree dat a channels of 768 kbps, each wit h a spreading fact or of
4.
WCDMA has lower net work lat ency t han EDGE, wit h about 100 t o 200 msec measured in
act ual net works. Alt hough UMTS Release 99 offers at t ract ive dat a services, t hese
services become much more efficient and more powerful wit h HSPA.
HSDPA
HSPA refers t o net works t hat support bot h HSDPA and HSUPA. All new deployment s
t oday are HSPA, and many operat ors have upgraded t heir HSDPA net works t o HSPA. For
example, in 2008, AT&T upgraded most of it s net work t o HSPA. By t he end of 2008,
HSPA was deployed t hroughout t he Americas. This sect ion covers t echnical aspect s of
HSDPA, while t he next sect ion covers HSUPA.
HSDPA, specified in 3GPP Release 5, is a high- performance, packet - dat a service t hat
delivers peak t heoret ical rat es of 14 Mbps. Peak user- achievable t hroughput rat es in

104
I nit ial UMTS net works had peak uplink rat es of 64 kbps or 128 kbps, but many deployment s
emphasize 384 kbps.

HSPA t o LTE- Advanced, Rysavy Research / 3G Americas, Sept ember 2009 Page 79
init ial deployment s are well over 1 Mbps and as high as 4 Mbps in some net works. The
same radio carrier can simult aneously service UMTS voice and dat a users, as well as
HSDPA dat a users. HSDPA also has significant ly lower lat ency measured t oday on some
net works as low as 70 msec on t he dat a channel.
HSDPA achieves it s high speeds t hrough t echniques similar t o t hose t hat push EDGE
performance past GPRS including higher order modulat ion, variable coding, and soft
combining, as well as t hrough t he addit ion of powerful new t echniques such as fast
scheduling. The higher spect ral efficiency and higher dat a rat es not only enable new
classes of applicat ions, but also support a great er number of users accessing t he
net work.
HSDPA achieves it s performance gains from t he following radio feat ures:
 High- speed channels shared in bot h code and t ime domains
 Short TTI
 Fast scheduling and user diversit y
 Higher order modulat ion
 Fast link adapt at ion
 Fast HARQ
These feat ures funct ion as follows:
Hi gh- Speed Shar ed Channel s and Shor t Tr ansmi ssi on Ti me I nt er val : First ,
HSDPA uses high- speed dat a channels called High Speed Physical Downlink Shared
Channels ( HS- PDSCH) . Up t o 15 of t hese channels can operat e in t he 5 MHz WCDMA
radio channel. Each uses a fixed spreading fact or of 16. User t ransmissions are assigned
t o one or more of t hese channels for a short TTI of 2 msec. The net work can t hen
readj ust how users are assigned t o different HS- PDSCH every 2 msec. The result is t hat
resources are assigned in bot h t ime ( t he TTI int erval) and code domains ( t he HS- PDSCH
channels) . Figure 32 illust rat es different users obt aining different radio resources.

HSPA t o LTE- Advanced, Rysavy Research / 3G Americas, Sept ember 2009 Page 80
Fi gur e 32: Hi gh Speed–Dow nl i nk Shar ed Channel s ( Ex ampl e)


Fast Schedul i ng and User Di v er si t y : Fast scheduling exploit s t he short TTI by
assigning users channels t hat have t he best inst ant aneous channel condit ions, rat her
t han in a round- robin fashion. Because channel condit ions vary somewhat randomly
across users, most users can be serviced wit h opt imum radio condit ions and t hereby
obt ain opt imum dat a t hroughput . Figure 33 shows how a scheduler might choose
bet ween t wo users based on t heir varying radio condit ions t o emphasize t he user wit h
bet t er inst ant aneous signal qualit y. Wit h about 30 users act ive in a sect or, t he net work
achieves significant user diversit y and significant ly higher spect ral efficiency. The syst em
also makes sure t hat each user receives a minimum level of t hroughput . This approach
is somet imes called proport ional fair scheduling.
2 msec
Time
C
h
a
n
n
e
l
i
z
a
t
i
o
n

C
o
d
e
s
User 4 User 3 User 2 User 1

HSPA t o LTE- Advanced, Rysavy Research / 3G Americas, Sept ember 2009 Page 81
Fi gur e 33: User Di v er si t y


Hi gher Or der Modul at i on: HSDPA uses bot h t he modulat ion used in WCDMA—namely
QPSK—and, under good radio condit ions, an advanced modulat ion scheme—16 QAM.
The benefit of 16 QAM is t hat 4 bit s of dat a are t ransmit t ed in each radio symbol as
opposed t o 2 bit s wit h QPSK. Dat a t hroughput is increased wit h 16 QAM, while QPSK is
available under adverse condit ions. HSPA Evolut ion will add 64 QAM modulat ion t o
furt her increase t hroughput rat es. Not e t hat 64 QAM was available in Release 7, and t he
combinat ion of MI MO and 64 QAM became available t his year in Release 8.
Fast Li nk Adapt at i on: Depending on t he condit ion of t he radio channel, different levels
of forward- error correct ion ( channel coding) can also be employed. For example, a
t hree- quart er coding rat e means t hat t hree quart ers of t he bit s t ransmit t ed are user bit s
and one quart er are error- correct ing bit s. The process of select ing and quickly updat ing
t he opt imum modulat ion and coding rat e is referred t o as fast link adapt at ion. This is
done in close coordinat ion wit h fast scheduling, as described above.
Fast Hybr i d Aut omat i c Repeat Request : Anot her HSDPA t echnique is Fast Hybrid
Aut omat ic Repeat Request ( Fast Hybrid ARQ) . “ Fast ” refers t o t he medium- access
cont rol mechanisms implement ed in Node B ( along wit h scheduling and link adapt at ion) ,
as opposed t o t he BSC in GPRS/ EDGE, and “ hybrid” refers t o a process of combining
repeat ed dat a t ransmissions wit h prior t ransmissions t o increase t he likelihood of
successful decoding. Managing and responding t o real- t ime radio variat ions at t he base
st at ion, as opposed t o an int ernal net work node, reduces delays and furt her improves
overall dat a t hroughput .
Using t he approaches j ust described, HSDPA maximizes dat a t hroughput s and capacit y
and minimizes delays. For users, t his t ranslat es t o bet t er net work performance under
loaded condit ions, fast er applicat ion performance, a great er range of applicat ions t hat
funct ion well, and increased product ivit y.
Field result s validat e t he t heoret ical t hroughput result s. Wit h init ial 1. 8 Mbps peak- rat e
devices, vendors measured consist ent t hroughput rat es in act ual deployment s of more
t han 1 Mbps. These rat es rose t o more t han 2 Mbps for 3. 6 Mbps devices and are close
High data rate
Low data rate
Time
User 2
User 1
User 2
User 1 User 2 User 1 User 2 User 1
S
i
g
n
a
l

Q
u
a
l
i
t
y

HSPA t o LTE- Advanced, Rysavy Research / 3G Americas, Sept ember 2009 Page 82
t o 4 Mbps for 7. 2 Mbps devices, assuming ot her port ions of t he net work ( for example,
backhaul) can support t he high t hroughput rat es.
I n 2008, t ypical devices support ing peak dat a rat es of 3. 6 Mbps or 7. 2 Mbps became
available. Many operat or net works support 7. 2 Mbps peak operat ion, and some even
support t he maximum rat e of 14. 4 Mbps.
HSPA t echnology is not st anding st ill. Advanced radio t echnologies are becoming
available. Among t hese t echnologies are mobile- receive diversit y and equalizat ion ( for
example, MMSE) , which improve t he qualit y of t he received radio signal prior t o
demodulat ion and decoding. This improvement enables not only higher peak HSDPA
t hroughput speeds, but makes t hese speeds available over a great er percent age of t he
coverage area.
HSUPA
Whereas HSDPA opt imizes downlink performance, HSUPA—which uses t he Enhanced
Dedicat ed Channel ( E- DCH) —const it ut es a set of improvement s t hat opt imizes uplink
performance. Net works and devices support ing HSUPA became available in 2007. These
improvement s include higher t hroughput s, reduced lat ency, and increased spect ral
efficiency. HSUPA is st andardized in Release 6. I t result s in an approximat ely 85 percent
increase in overall cell t hroughput on t he uplink and more t han 50 percent gain in user
t hroughput . HSUPA also reduces packet delays, a significant benefit result ing in much
improved applicat ion performance on HSPA net works
Alt hough t he primary downlink t raffic channel support ing HSDPA serves is a shared
channel designed for t he support of services delivered t hrough t he packet - swit ched
domain, t he primary uplink t raffic channel defined for HSUPA is a dedicat ed channel t hat
could be used for services delivered t hrough eit her t he circuit - swit ched or t he packet -
swit ched domains. Nevert heless, by ext ension and for simplicit y, t he WCDMA- enhanced
uplink capabilit ies are oft en ident ified in t he lit erat ure as HSUPA.
Such an improved uplink benefit s users in a number of ways. For inst ance, some user
applicat ions t ransmit large amount s of dat a from t he mobile st at ion such as sending
video clips or large present at ion files. For fut ure applicat ions like VoI P, improvement s
will balance t he capacit y of t he uplink wit h t he capacit y of t he downlink.
HSUPA achieves it s performance gains t hrough t he following approaches:
 An enhanced dedicat ed physical channel
 A short TTI , as low as 2 msec, which allows fast er responses t o changing radio
condit ions and error condit ions
 Fast Node B- based scheduling, which allows t he base st at ion t o efficient ly
allocat e radio resources
 Fast Hybrid ARQ, which improves t he efficiency of error processing
The combinat ion of TTI , fast scheduling, and Fast Hybrid ARQ also serves t o reduce
lat ency, which can benefit many applicat ions as much as improved t hroughput . HSUPA
can operat e wit h or wit hout HSDPA in t he downlink, alt hough it is likely t hat most
net works will use t he t wo approaches t oget her. The improved uplink mechanisms also
t ranslat e t o bet t er coverage and, for rural deployment s, larger cell sizes.
HSUPA can achieve different t hroughput rat es based on various paramet ers including t he
number of codes used, t he spreading fact or of t he codes, t he TTI value, and t he
t ransport block size in byt es.

HSPA t o LTE- Advanced, Rysavy Research / 3G Americas, Sept ember 2009 Page 83
I nit ial devices enabled peak user rat es of close t o 2 Mbps as measured in act ual net work
deployment s. Fut ure devices will ult imat ely approach speeds close t o 5 Mbps, alt hough
only wit h t he addit ion of int erference cancellat ion met hods t hat boost SNR.
Beyond t hroughput enhancement s, HSUPA also significant ly reduces lat ency. I n
opt imized net works, lat ency will fall below 50 msec, relat ive t o current HSDPA net works
at 70 msec. And wit h a lat er int roduct ion of a 2 msec TTI , lat ency will be as low as 30
msec.
Evolution of HSPA (HSPA+)
OFDMA syst ems have at t ract ed considerable at t ent ion t hrough t echnologies such as
3GPP LTE and WiMAX. As already discussed in t his paper, however, CDMA approaches
can mat ch OFDMA approaches in reduced channel bandwidt hs. The goal in evolving
HSPA is t o exploit available radio t echnologies—largely enabled by increases in digit al
signal processing power—t o maximize CDMA- based radio performance. This not only
makes HSPA compet it ive, it significant ly ext ends t he life of sizeable operat or
infrast ruct ure invest ment s.
Wireless and net working t echnologist s have defined a series of enhancement s for HSPA,
some of which are specified in Release 7 and some of which are being finalized in
Release 8. These include advanced receivers, MI MO, Cont inuous Packet Connect ivit y,
Higher- Order Modulat ion and One Tunnel Archit ect ure.
Adv anced Recei v er s.
One import ant area is advanced receivers for which 3GPP has specified a number of
designs. These designs include Type 1, which uses mobile- receive diversit y; Type 2,
which uses channel equalizat ion; and Type 3, which includes a combinat ion of receive
diversit y and channel equalizat ion. Type 3i devices, which are not yet available, will
employ int erference cancellat ion. Not e t hat t he different t ypes of receivers are release-
independent . For example, Type 3i receivers will work and provide a capacit y gain in a
Release 5 net work.
The first approach is mobile- receive diversit y. This t echnique relies on t he opt imal
combinat ion of received signals from separat e receiving ant ennas. The ant enna spacing
yields signals t hat have somewhat independent fading charact erist ics. Hence, t he
combined signal can be more effect ively decoded, which result s in an almost doubling of
downlink capacit y when employed in conj unct ion wit h t echniques such as channel
equalizat ion. Receive diversit y is effect ive even for small devices such as PC Card
modems and smart phones.
Current receiver archit ect ures based on rake receivers are effect ive for speeds up t o a
few megabit s per second. But at higher speeds, t he combinat ion of reduced symbol
period and mult ipat h int erference result s in int er- symbol int erference and diminishes
rake receiver performance. This problem can be solved by advanced- receiver
archit ect ures wit h channel equalizers t hat yield addit ional capacit y gains over HSDPA
wit h receive diversit y. Alt ernat e advanced- receiver approaches include int erference
cancellat ion and generalized rake receivers ( G- Rake) . Different vendors are emphasizing
different approaches. The performance requirement s for advanced- receiver
archit ect ures, however, are specified in 3GPP Release 6. The combinat ion of mobile-
receive diversit y and channel equalizat ion ( Type 3) is especially at t ract ive, because it
result s in a large capacit y gain independent of t he radio channel.
What makes such enhancement s at t ract ive is t hat t he net works do not require any
changes ot her t han increased capacit y wit hin t he infrast ruct ure t o support t he higher
bandwidt h. Moreover, t he net work can support a combinat ion of devices including bot h

HSPA t o LTE- Advanced, Rysavy Research / 3G Americas, Sept ember 2009 Page 84
earlier devices t hat do not include t hese enhancement s and lat er devices t hat do. Device
vendors can select ively apply t hese enhancement s t o t heir higher performing devices.
MI MO
Anot her st andardized capabilit y is MI MO, a t echnique t hat employs mult iple t ransmit
ant ennas and mult iple receive ant ennas, oft en in combinat ion wit h mult iple radios and
mult iple parallel dat a st reams. The most common use of t he t erm “ MI MO” applies t o
spat ial mult iplexing. The t ransmit t er sends different dat a st reams over each ant enna.
Whereas mult ipat h is an impediment for ot her radio syst ems, MI MO—as illust rat ed in
Figure 34—act ually exploit s mult ipat h, relying on signals t o t ravel across different
uncorrelat ed communicat ions pat hs. This result s in mult iple dat a pat hs effect ively
operat ing somewhat in parallel and, t hrough appropriat e decoding, in a mult iplicat ive
gain in t hroughput .
Fi gur e 34: MI MO Usi ng Mul t i pl e Pat hs t o Boost Thr oughput and Capaci t y

Test s of MI MO have proven very promising in WLANs operat ing in relat ive isolat ion
where int erference is not a dominant fact or. Spat ial mult iplexing MI MO should also
benefit HSPA “ hot spot s” serving local areas such as airport s, campuses, and malls,
where t he t echnology will increase capacit y and peak dat a rat es. I n a fully loaded
net work wit h int erference from adj acent cells, however, overall capacit y gains will be
more modest —in t he range of 20 t o 33 percent over mobile- receive diversit y. Relat ive t o
a 1x1 ant enna syst em, however, 2X2 MI MO can deliver cell t hroughput gains of about
80 percent . 3GPP has st andardized spat ial mult iplexing MI MO in Release 7 using Double
Transmit Adapt ive Array ( D- TxAA) .
105

Alt hough MI MO can significant ly improve peak rat es, ot her t echniques such as Space
Division Mult iple Access ( SDMA) —also a form of MI MO—may be even more effect ive

105
For furt her det ails on t hese t echniques, refer t o t he 3G Americas’ whit e paper “ Mobile Broadband:
The Global Evolut ion of UMTS- HSPA. 3GPP Release 7 and Beyond. ”
Encoder Decoder

HSPA t o LTE- Advanced, Rysavy Research / 3G Americas, Sept ember 2009 Page 85
t han MI MO for improving capacit y in high spect ral efficiency syst ems employing a reuse
fact or of 1.
Cont i nuous Pack et Connect i vi t y
I n Release 7, CPC enhancement s reduce t he uplink int erference creat ed by t he
dedicat ed physical cont rol channels of packet dat a users when t hose channels have no
user dat a t o t ransmit . This, in t urn, increases t he number of simult aneously connect ed
HSUPA users. CPC allows bot h discont inuous uplink t ransmission and discont inuous
downlink recept ion, wherein t he modem can t urn off it s receiver aft er a cert ain period of
HSDPA inact ivit y. CPC is especially beneficial t o VoI P on t he uplink, which consumes t he
most power, because t he radio can t urn off bet ween VoI P packet s. See Figure 35.
Fi gur e 35: Cont i nuous Pack et Connect i vi t y


Hi gher Or der Modul at i on
Anot her way of increasing performance is t o use higher order modulat ion. HSPA uses 16
QAM on t he downlink and QPSK on t he uplink. But radio links can achieve higher
t hroughput s—adding 64 QAM on t he downlink and 16 QAM on t he uplink—precisely what
is added in HSPA+ . Higher order modulat ion requires a bet t er SNR, which is enabled
t hrough ot her enhancement s such as receive diversit y and equalizat ion.
HSPA+
Taking advant age of t hese various radio t echnologies, 3GPP has st andardized a number
of feat ures in Release 7 including higher order modulat ion and MI MO. Collect ively, t hese
capabilit ies are referred t o as HSPA+ . Release 8 will include furt her enhancement s.
The goals of HSPA+ are t o:
 Exploit t he full pot ent ial of a CDMA approach before moving t o an OFDM plat form
in 3GPP LTE.
 Achieve performance close t o LTE in 5 MHz of spect rum.
Data
Pilot
Data
Pilot
Without CPC
With CPC

HSPA t o LTE- Advanced, Rysavy Research / 3G Americas, Sept ember 2009 Page 86
 Provide smoot h int erworking bet ween HSPA+ and LTE, t hereby facilit at ing t he
operat ion of bot h t echnologies. As such, operat ors may choose t o leverage t he
EPC/ SAE planned for LTE.
 Allow operat ion in a packet - only mode for bot h voice and dat a.
 Be backward- compat ible wit h previous syst ems while incurring no performance
degradat ion wit h eit her earlier or newer devices.
 Facilit at e migrat ion from current HSPA infrast ruct ure t o HSPA+ infrast ruct ure.
Depending on t he feat ures implement ed, HSPA+ can exceed t he capabilit ies of I EEE
802. 16e- 2005 ( mobile WiMAX) in t he same amount of spect rum. This is mainly because
MI MO in HSPA support s closed- loop operat ion wit h precode weight ing, as well as
mult icode- word MI MO and enables t he use of SI C receivers. I t is also part ly because
HSPA support s I ncrement al Redundancy ( I R) and has lower overhead t han WiMAX.
Table 14 summarizes t he capabilit ies of HSPA and HSPA+ based on various met hods.
Tabl e 14: HSPA Thr oughput Evol ut i on
Technol ogy
Dow nl i nk
( Mbps) Peak
Dat a Rat e
Upl i nk ( Mbps)
Peak Dat a
Rat e
HSPA as def i ned i n Rel ease 6 14. 4 5. 76
Rel ease 7 HSPA+ DL 64 QAM,
UL 16 QAM
21. 1 11. 5
Rel ease 7 HSPA+ 2X2 MI MO,
DL 16 QAM, UL 16 QAM
28. 0 11. 5
Rel ease 8 HSPA+ 2X2 MI MO
DL 64 QAM, UL 16 QAM
42. 2 11. 5
Rel ease 9 HSPA+ 2X2 MI MO,
Dual Car r i er
84. 0 23. 0

Beyond t he peak rat e of 42 Mbps defined in Release 8, Release 9 may specify 2X2 MI MO
in combinat ion wit h dual- carrier operat ion, which would furt her boost peak net work
rat es t o 84 Mbps. Fut ure releases of HSPA+ could also use a quad- carrier approach for
even higher t hroughput s. Dual- and mult i- carrier operat ion are explained furt her below.
HSPA+ will also have improved lat ency performance of below 50 msec and improved
packet call set up t ime of below 500 msec.
HSPA+ wit h 28 Mbps capabilit y will be available for deployment by t he end of 2009, and
HSPA+ wit h 42 Mbps capabilit y on t he downlink and 11. 5 Mbps on t he uplink could be
ready for deployment by 2009 or 2010.
Given t he large amount of backhaul bandwidt h required t o support HSPA+ , as well as
addit ional MI MO radios at cell sit es, operat ors are likely t o init ially deploy HSPA+ in
limit ed “ hot spot ” coverage areas such as airport s, ent erprise campuses, and in- building
net works. Wit h advances in backhaul t ransport like met ropolit an Et hernet , however,
operat ors will be able t o expand coverage.
The prior discussion emphasizes t hroughput speeds, but HSPA+ will also more t han
double HSPA capacit y as well as reduce lat ency below 25 msec. Sleep- t o- dat a- t ransfer
t imes of less t han 200 msec will improve users’ “ always- connect ed” experience, and

HSPA t o LTE- Advanced, Rysavy Research / 3G Americas, Sept ember 2009 Page 87
reduced power consumpt ion wit h VoI P will result in t alk t imes t hat are more t han 50
percent higher.
From a deployment point of view, operat ors will be able t o int roduce HSPA+ capabilit ies
t hrough eit her a soft ware upgrade or hardware expansions t o exist ing cabinet s t o
increase capacit y. Cert ain upgrades will be simpler t han ot hers. For example, upgrading
t o 64- QAM support will be easier t o implement t han 2X2 MI MO for many net works. For
net works t hat have implement ed uplink diversit y in t he base st at ion, however, t hose
mult iple ant ennas will facilit at e MI MO deployment .
Dual - Car r i er HSPA
3GPP has defined a capabilit y in Release 8 for dual- carrier HSPA operat ion. This
approach coordinat es t he operat ion of HSPA on t wo adj acent 5 MHz carriers so t hat dat a
t ransmissions can achieve higher t hroughput rat es, as shown in Figure 36. The work
it em assumes t wo adj acent carriers, downlink operat ion and no MI MO. I n t his
configurat ion, it is possible t o achieve a doubling of t he 21 Mbps maximum rat e
available on each channel t o 42 Mbps.
Fi gur e 36: Dual - Car r i er Oper at i on w i t h One Upl i nk Car r i er
106


There are a number of benefit s t o t his approach:
 An increase in spect ral efficiency of about 20%, comparable t o what can be
obt ained wit h 2X2 MI MO.
 Significant ly higher peak t hroughput s available t o users, especially in light ly-
loaded net works.
 Same maximum- t hroughput rat e of 42 Mbps as using MI MO, but wit h a less
expensive infrast ruct ure upgrade.
By scheduling packet s across t wo carriers, t here is bet t er resource ut ilizat ion, result ing
in what is called t runking gain. Mult i- user diversit y also improves because t here are
more users t o select from.
Under considerat ion for Release 10 is t he use of four channels.
Figure 37 shows an analysis of dual- carrier performance using a cumulat ive dist ribut ion
funct ion. CDF indicat es t he probabilit y of achieving a part icular t hroughput rat e and t he
figure demonst rat es a consist ent doubling of t hroughput .

106
Source: "LTE for UMTS, OFDMA and SC- FDMA Based Radio Access, ” Harri Holma and Ant t i Toskala,
Wiley, 2009.
2 x 5 MHz 1 x 5 MHz
2 x 5 MHz 1 x 5 MHz
UE1
UE2
Uplink Downlink

HSPA t o LTE- Advanced, Rysavy Research / 3G Americas, Sept ember 2009 Page 88
Fi gur e 37: Dual - Car r i er Per f or mance
107


One- Tunnel Ar chi t ect ur e
Anot her way HSPA performance can be improved is t hrough a flat t er archit ect ure. I n
Release 7, t here is t he opt ion of a one- t unnel archit ect ure by which t he net work
est ablishes a direct t ransfer pat h for user dat a bet ween RNC and GGSN, while t he SGSN
st ill performs all cont rol funct ions. This brings several benefit s such as eliminat ing
hardware in t he SGSN and simplified engineering of t he net work.
There is also an int egrat ed RNC/ NodeB opt ion in which RNC funct ions are int egrat ed in
t he Node B. This is part icularly beneficial in femt ocell deployment s, as an RNC would
ot herwise need t o support t housands of femt ocells. The int egrat ed RNC/ NodeB for
HSPA+ has been agreed- upon as an opt ional archit ect ure alt ernat ive for packet -
swit ched- based services.
These new archit ect ures, as shown in Figure 38, are similar t o t he EPC/ SAE archit ect ure,
especially on t he packet - swit ched core net work side where t hey provide synergies wit h
t he int roduct ion of LTE.

107
Source: 3G Americas’ member company cont ribut ion.
0 5 10 15 20 25 30 35 40
0
10
20
30
40
50
60
70
80
90
100
C
D
F

[
%
]
Achievable bitrate [Mbps]


RAKE, single-carrier
RAKE, multi-carrier
GRAKE, single-carrier
GRAKE, multi-carrier
GRAKE2, single-carrier
GRAKE2, multi-carrier
Ped A, 10% load

HSPA t o LTE- Advanced, Rysavy Research / 3G Americas, Sept ember 2009 Page 89
Fi gur e 38: HSPA One- Tunnel Ar chi t ect ur e
108




HSPA, HSPA+ , and ot her advanced funct ions provide a compelling advant age for UMTS
over compet ing t echnologies: The abilit y t oday t o support voice and dat a services on t he
same carrier and across t he whole available radio spect rum; t o offer t hese services
simult aneously t o users; t o deliver dat a at ever- increasing broadband rat es; and t o do
so in a spect rally efficient manner.
HS- FACH
I n Release 7, a new capabilit y called High- Speed Access Forward Access Channel ( HS-
FACH) , illust rat ed in Figure 39, reduces set up t ime t o pract ically zero and provides a
more efficient way of carrying applicat ion signaling for always- on applicat ions. The
net work accomplishes t his by using t he same HSDPA power/ code resources for access
request s ( CELL_FACH st at e) as for dedicat ed packet t ransfer ( CELL_DCH) . This allows
dat a t ransmission t o st art during t he HS- FACH st at e wit h increased dat a rat es
immediat ely available t o t he user equipment . During t he HS- FACH st at e, t he net work
allocat es dedicat ed resources for t ransit ioning t he user equipment t o a dedicat ed
channel st at e.


108
Source: 3G Americas’ whit e paper, 2007, “ UMTS Evolut ion from 3GPP Release 7 t o Release 8. ”

User Plane
Cont r ol Pl ane
Node B
RNC
SGSN
GGSN
Tradi t i onal HSPA
Archi t ect ure
Node B
SGSN
GGSN
Possi bl e HSPA+ wi t h
One-Tunnel Ar chi t ect ure
Node B
SGSN
GGSN
HSPA wi t h One-Tunnel
Archi t ect ure
RNC

HSPA t o LTE- Advanced, Rysavy Research / 3G Americas, Sept ember 2009 Page 90
Fi gur e 39: Hi gh- Speed For w ar d Access Channel
109






Figure 40 summarizes t he capabilit ies and benefit s of t he feat ures being deployed in
HSPA+ .

109
Source: "LTE for UMTS, OFDMA and SC- FDMA Based Radio Access, ” Harri Holma and Ant t i Toskala,
Wiley, 2009.
RACH +
FACH
HSDPA +
HSUPA
HSDPA + HSUPA
R99-R6 solution R7/R8 solution
6-32 kbps >1 Mbps
Seamless transition Delay >0.5 s
>1 Mbps
Cell_FACH Cell_DCH Cell_FACH Cell_DCH
PCH
FACH
DCH
/HSPA
No data flow
during transition
>500 ms
Cell update and
C-RNTI allocation
takes >300 ms
RB recon-
figuration
RB recon-
figuration
PCH
eFACH
HSPA
Data flows on HS-
FACH also during
transition.
Immediate transmission w/o
cell update. No PCH
required.
R99-RC RRC States R7/8 RRC States
DCH – Dedicated Channel
FACH – Forward Access Channel
RACH – Reverse Access Channel
PCH – Paging Channel
HS-FACH – High Speed FACH
RB – Radio Bearer
RRC – Radio Resource Control

HSPA t o LTE- Advanced, Rysavy Research / 3G Americas, Sept ember 2009 Page 91
Fi gur e 40: Summar y of HSPA Funct i ons and Benef i t s
110


HSPA Voice Support
Voice support wit h WCDMA- dedicat ed channels in UMTS net works is spect rally very
efficient . Moreover, current net works support simult aneous voice and dat a operat ion.
There are, however, reasons t o consider alt ernat e approaches including reducing power
consumpt ion and being able t o support even more users. One approach is called circuit -
swit ched voice over HSPA. The ot her is VoI P.
CS Voi ce ov er HSPA
HSPA channels employ many opt imizat ions t o obt ain a high degree of dat a t hroughput ,
which is why it makes sense t o use t hem t o carry voice communicat ions. Doing so wit h
VoI P, however, requires not only support ing packet ized voice in t he radio channel, but
also wit hin t he infrast ruct ure net work. There is an elegant alt ernat ive: To packet ize t he
circuit - swit ched voice t raffic which is already in digit al form, use t he HSPA channels t o
carry t he CS voice, but t hen t o connect t he CS voice t raffic back int o t he exist ing CS
infrast ruct ure ( MSCs, et c. ) immediat ely beyond t he radio access net work. This requires
relat ively st raight forward changes in j ust t he radio net work and in devices. The following
figure shows t he infrast ruct ure changes required at t he Node B and wit hin t he RNC.

110
Source: 3G Americas’ member cont ribut ion.
Uplink DTX + downlink
DRX
L2 optimization (Flexible
RLC)
High speed FACH + High
speed RACH
Downlink 64QAM, MIMO
and Dual carrier
CS voice over HSPA
Uplink 16QAM
Lower UE power consumption
Higher voice capacity
Higher L2 throughput and less
processing requirements
Lower latency = better response
times
More efficient common channels =
savings in channel elements
Higher downlink peak data rates
and higher data capacity
Higher uplink peak data rates
Flat architecture
optimization
Less network elements

HSPA t o LTE- Advanced, Rysavy Research / 3G Americas, Sept ember 2009 Page 92
Fi gur e 41: I mpl ement at i on of HSPA CS Voi ce
111



Wit h t his approach, legacy mobile phones can cont inue using WCDMA- dedicat ed t raffic
channels for voice communicat ions, while new devices use HSPA channels. HSPA CS
voice can be deployed wit h Release 7 or lat er net works.
The many benefit s of t his approach, list ed below, make it highly likely t hat operat ors will
adopt it :
 Relat ively easy t o implement and deploy.
 Transparent t o exist ing CS infrast ruct ure.
 Support s bot h narrowband and wideband codecs.
 Significant ly improves bat t ery life wit h voice communicat ions.
 Enables fast er call connect ions.
 Provides a 50 t o 100% capacit y gain over current voice implement at ions.
 Act s as a st epping st one t o VoI P over HSPA/ LTE in t he fut ure.
VoI P
Once HSDPA and HSUPA are available, operat ors will have anot her opt ion of moving
voice t raffic over t o t hese high- speed dat a channels, which is using VoI P. This will
event ually increase voice capacit y, allow operat ors t o consolidat e t heir infrast ruct ure on
an I P plat form, and enable innovat ive new applicat ions t hat combine voice wit h dat a
funct ions in t he packet domain. VoI P is possible in Release 6, but it is enhancement s in

111
Source: 3G Americas’ whit e paper, 2007, “ UMTS Evolut ion from 3GPP Release 7 t o Release 8. ”
IuCS
IuPS
RNC
CS R99
AMR
adapt.
Transport
queues etc
HSPA
PS R99
NodeB
HSPA scheduler
Combined
to one
carrier
AMR adaptation
possible
CS mapped to R99 or HSPA bearer
depending on terminal capability
Scheduler prioritizes
voice packets

HSPA t o LTE- Advanced, Rysavy Research / 3G Americas, Sept ember 2009 Page 93
Release 7 t hat make it highly efficient and t hus at t ract ive t o net work operat ors. VoI P will
be implement ed in conj unct ion wit h I MS, discussed lat er in t his paper.
One at t ract ive aspect of deploying VoI P wit h HSPA is t hat operat ors can smoot hly
migrat e users from circuit - swit ched operat ion t o packet - swit ched operat ion over t ime.
Because t he UMTS radio channel support s bot h circuit - swit ched voice and packet -
swit ched dat a, some voice users can be on legacy circuit - swit ched voice and ot hers can
be on VoI P. Figure 42 shows a syst em’s voice capacit y wit h t he j oint operat ion of circuit -
swit ched and I P- based voice services.
Fi gur e 42: Abi l i t y f or UMTS t o Suppor t Ci r cui t and Pack et Voi ce User s
112



VoI P capacit y gains are quant ified in det ail in t he main part of in t his paper. They range
from 20 percent t o as high as 100 percent wit h t he implement at ion of int erference
cancellat ion and t he minimizat ion of I P overhead t hrough a scheme called Robust
Header Compression ( ROHC) .
Whereas packet voice is t he only way voice will be support ed in LTE, wit h HSPA+ , it may
not be used immediat ely for primary voice services. This is because UMTS already has a
highly efficient , circuit - swit ched voice service and already allows simult aneous
voice/ dat a operat ion. Moreover, packet voice requires a considerable amount of new
infrast ruct ure in t he core net work. As a result , packet voice will likely be used init ially as
part of ot her services ( for example, t hose based on I MS) , and only over t ime will it
t ransit ion t o primary voice service.
3GPP LTE
Alt hough HSPA and HSPA+ offer a highly efficient broadband- wireless service t hat will
enj oy success for t he remainder of t his decade and well int o t he next , 3GPP has
complet ed t he specificat ion for Long Term Evolut ion as part of Release 8. LTE will allow
operat ors t o achieve even higher peak t hroughput s in higher spect rum bandwidt h. Work
on LTE began in 2004 wit h an official work it em st art ed in 2006 and a complet ed
specificat ion early 2009. I nit ial deployment s will occur in 2010.

112
Source: 3G Americas’ member cont ribut ion.
0
0.2
0.4
0.6
0.8
1
1.2
1.4
0 2 4 6 8 10 12 14
VoIP
CS
CS + VoIP
Power reserved for PS traffic (W)
R
e
l
a
t
i
v
e

C
a
p
a
c
i
t
y
PS Evolution
0
0.2
0.4
0.6
0.8
1
1.2
1.4
0 2 4 6 8 10 12 14
VoIP
CS
CS + VoIP
Power reserved for PS traffic (W)
R
e
l
a
t
i
v
e

C
a
p
a
c
i
t
y
0
0.2
0.4
0.6
0.8
1
1.2
1.4
0 2 4 6 8 10 12 14
VoIP
CS
CS + VoIP
Power reserved for PS traffic (W)
R
e
l
a
t
i
v
e

C
a
p
a
c
i
t
y
PS Evolution

HSPA t o LTE- Advanced, Rysavy Research / 3G Americas, Sept ember 2009 Page 94
LTE uses OFDMA on t he downlink, which is well suit ed t o achieve high peak dat a rat es in
high- spect rum bandwidt h. WCDMA radio t echnology is basically as efficient as OFDM for
delivering peak dat a rat es of about 10 Mbps in 5 MHz of bandwidt h. Achieving peak
rat es in t he 100 Mbps range wit h wider radio channels, however, would result in highly
complex t erminals, and it is not pract ical wit h current t echnology. This is where OFDM
provides a pract ical implement at ion advant age. Scheduling approaches in t he frequency
domain can also minimize int erference, t hereby boost ing spect ral efficiency. The OFDMA
approach is also highly flexible in channelizat ion, and LTE will operat e in various radio
channel sizes ranging from 1. 4 t o 20 MHz.
On t he uplink, however, a pure OFDMA approach result s in high Peak t o Average Rat io
( PAR) of t he signal, which compromises power efficiency and, ult imat ely, bat t ery life.
Hence, LTE uses an approach called SC- FDMA, which is somewhat similar t o OFDMA, but
has a 2 t o 6 dB PAR advant age over t he OFDMA met hod used by ot her t echnologies such
as WiMAX.
LTE capabilit ies include:
 Downlink peak dat a rat es up t o 326 Mbps wit h 20 MHz bandwidt h.
 Uplink peak dat a rat es up t o 86.4 Mbps wit h 20 MHz bandwidt h.
 Operat ion in bot h TDD and FDD modes.
 Scalable bandwidt h up t o 20 MHz covering 1. 4, 3, 5, 10, 15, and 20 MHz in t he
st udy phase.
 I ncreased spect ral efficiency over Release 6 HSPA by a fact or of t wo t o four.
 Reduced lat ency, t o 10 msec round- t rip t imes bet ween user equipment and t he
base st at ion, and t o less t han 100 msec t ransit ion t imes from inact ive t o act ive.
 Self- opt imizing capabilit ies under operat or cont rol, and preferences t hat will
aut omat e net work planning and will result in lower operat or cost s.
LTE Thr oughput Rat es
The overall obj ect ive is t o provide an ext remely high- performance, radio- access
t echnology t hat offers full vehicular speed mobilit y and t hat can readily coexist wit h
HSPA and earlier net works. Because of scalable bandwidt h, operat ors will be able t o
easily migrat e t heir net works and users from HSPA t o LTE over t ime.
Table 15 shows LTE peak dat a rat es based on different downlink and uplink designs.
Tabl e 15: LTE Peak Thr oughput Rat es
LTE Conf i gur at i on
Dow nl i nk ( Mbps)
Peak Dat a Rat e
Upl i nk ( Mbps)
Peak Dat a Rat e
Using 2X2 MI MO in t he Downlink and
16 QAM in t he Uplink

172. 8 57. 6
Using 4X4 MI MO in t he Downlink and
64 QAM in t he Uplink

326. 4 86. 4



HSPA t o LTE- Advanced, Rysavy Research / 3G Americas, Sept ember 2009 Page 95
LTE is not only efficient for dat a but , because of a highly efficient uplink, is ext remely
efficient for VoI P t raffic. I n 10 MHz of spect rum, LTE VoI P capacit y will reach almost 500
users.
113

OFDMA and Schedul i ng
LTE implement s OFDM in t he downlink. The basic principle of OFDM is t o split a high- rat e
dat a st ream int o a number of parallel low- rat e dat a st reams, each a narrowband signal
carried by a subcarrier. The different narrowband st reams are generat ed in t he
frequency domain, and t hen combined t o form t he broadband st ream using a
mat hemat ical algorit hm called an I nverse Fast Fourier Transform ( I FFT) t hat is
implement ed in digit al- signal processors. I n LTE, t he subcarriers have 15 kHz spacing
from each ot her. LTE maint ains t his spacing regardless of t he overall channel bandwidt h,
which simplifies radio design, especially in support ing radio channels of different widt hs.
The number of subcarriers ranges from 72 in a 1. 4 MHz channel t o 1, 200 in a 20 MHz
channel.
The composit e signal is obt ained aft er t he I FFT is ext ended by repeat ing t he init ial part
of t he signal ( called t he Cyclic Prefix [ CP] ) . This ext ended signal represent s an OFDM
symbol. The CP is basically a guard t ime during which reflect ed signals will reach t he
receiver. I t result s in an almost complet e eliminat ion of mult ipat h induced I nt ersymbol
I nt erference ( I SI ) , which ot herwise makes ext remely high dat a- rat e t ransmissions
problemat ic. The syst em is called ort hogonal, because t he subcarriers are generat ed in
t he frequency domain ( making t hem inherent ly ort hogonal) , and t he I FFT conserves t hat
charact erist ic. OFDM syst ems may lose t heir ort hogonal nat ure as a result of t he Doppler
shift induced by t he speed of t he t ransmit t er or t he receiver. 3GPP specifically select ed
t he subcarrier spacing of 15 kHz t o avoid any performance degradat ion in high- speed
condit ions. WiMAX syst ems t hat use a lower subcarrier spacing ( ~ 11 kHz) will be more
impact ed in high- speed condit ions t han LTE.
Fi gur e 43: OFDM Sy mbol w i t h Cy cl i c Pr ef i x


The mult iple- access aspect of OFDMA comes from being able t o assign different users
different subcarriers over t ime. A minimum resource block t hat t he syst em can assign t o
a user t ransmission consist s of 12 subcarriers over 14 symbols in 1. 0 msec. Figure 44
shows how t he syst em can assign t hese resource blocks t o different users over bot h
t ime and frequency.

113
Source: 3GPP Mult i- member analysis.
Cyclic Prefix
(4.8 usec)
Data
(66.7 usec)

HSPA t o LTE- Advanced, Rysavy Research / 3G Americas, Sept ember 2009 Page 96
Fi gur e 44: LTE OFDMA Dow nl i nk Resour ce Assi gnment i n Ti me and Fr equency

By having cont rol over which subcarriers are assigned in which sect ors, LTE can easily
cont rol frequency reuse. By using all t he subcarriers in each sect or, t he syst em would
operat e at a frequency reuse of 1; but by using a different one t hird of t he subcarriers in
each sect or, t he syst em achieves a looser frequency reuse of 1/ 3. The looser frequency
reduces overall spect ral efficiency, but delivers high peak rat es t o users.
Beyond cont rolling frequency reuse, frequency domain scheduling, as shown in Figure 45
can use t hose resource blocks t hat are not faded, somet hing t hat is not possible in
CDMA- based syst ems. Since different frequencies may fade different ly for different
users, t he syst em can allocat e t hose frequencies for each user t hat result in t he great est
t hroughput . This result s in up t o a 40% gain in average cell t hroughput for low user
speed ( 3 km/ hour) , assuming a large number of users and no MI MO. The benefit
decreases at higher user speeds.
Time
F
r
e
q
u
e
n
c
y
User 1
User 2
User 3
User 4
Minimum resource block consists of
14 symbols and 12 subcarriers

HSPA t o LTE- Advanced, Rysavy Research / 3G Americas, Sept ember 2009 Page 97
Fi gur e 45: Fr equency - Domai n Schedul i ng i n LTE
114


LTE is specified for a variet y of MI MO configurat ions. On t he downlink, t hese include
2X2, 4X2 ( four ant ennas at t he base st at ion) , and 4X4. I nit ial deployment will likely be
2x2. 4X4 will be most likely used init ially in femt ocells. On t he uplink, t here are t wo
possible approaches: single- user MI MO ( SU- MI MO) and mult i- user MI MO ( MU- MI MO) .
SU- MI MO is more complex t o implement as it requires t wo parallel radio t ransmit chains
in t he mobile device, whereas MU- MI MO does not require any addit ional implement at ion
at t he device. The first LTE release t hus incorporat es MU- MI MO wit h SU- MI MO deferred
for t he second LTE release.
LTE is designed t o operat e in channel bandwidt hs from 1. 4 MHz t o 20 MHz. The great est
efficiency, however, occurs wit h higher bandwidt h. A 3G Americas’ member analysis
predict s 40% lower spect ral efficiency wit h 1. 4 MHz radio channels and 13% lower
efficiency wit h 3 MHz channels.
115
The syst em, however, achieves nearly all of it s
efficiency wit h 5 MHz channels or wider.
TDD Har moni zat i on
3GPP developed LTE TDD t o be fully harmonized wit h LTE FDD including alignment of
frame st ruct ures, ident ical symbol- level numerology, t he possibilit y of using similar
reference signal pat t erns, and similar synchronizat ion and cont rol channels. Also, t here
is only one TDD variant . Furt hermore, LTE TDD has been designed t o co- exist wit h TD-
SCDMA and TD- CDMA/ UTRA ( bot h low- chip rat e and high- chip rat e versions) . LTE TDD
achieves compat ibilit y and co- exist ence wit h TD- SCDMA by defining frame st ruct ures
where t he DL and UL t ime periods can be t ime aligned t o prevent BTS t o BTS and UE t o
UE int erference t o support operat ion in adj acent carriers wit hout t he need for large
guardbands bet ween t he t echnologies. This will simplify deployment of LTE TDD in

114
3G Americas’ member cont ribut ion.
115
3G Americas’ member company analysis 2009.
Frequency
Resource block
Transmit on those resource
blocks that are not faded
Carrier bandwidth

HSPA t o LTE- Advanced, Rysavy Research / 3G Americas, Sept ember 2009 Page 98
count ries such as China t hat are deploying TD- SCDMA. Figure 46 demonst rat es t he
synchronizat ion bet ween TC- SCDMA and LTE- TDD in adj acent channels.
Fi gur e 46: TDD Fr ame Co- Ex i st ence Bet w een TD- SCDMA and LTE TDD
116


For LTE FDD and TDD t o coexist , large guardbands will be needed t o prevent
int erference. The organizat ion Next Generat ion Mobile Net works has a proj ect for LTE
TDD and FDD convergence.
117

4G, IMT-Advanced and LTE-Advanced
As int roduced earlier in t his paper, t he t erm 4G will apply t o net works t hat comply wit h
t he requirement s of I MT- Advanced t hat are art iculat ed in Report I TU- R M. 2134. Some of
t he key requirement s or st at ement s include:
- Support for scalable bandwidt h up t o and including 40 MHz.
- Encouragement t o support wider bandwidt hs ( e. g. , 100 MHz) .
- Minimum downlink peak spect ral efficiency of 15 bps/ Hz ( assumes 4X4 MI MO) .
- Minimum uplink peak spect ral efficiency of 6. 75 bps/ Hz ( assumes 2X4 MI MO) .
Table 16 shows t he requirement s for cell- spect ral efficiency.

116
Source: A 3G Americas’ member company.
117
Source: ht t p: / / www.ngmn. org/ workprogramme. ht ml.

HSPA t o LTE- Advanced, Rysavy Research / 3G Americas, Sept ember 2009 Page 99
Tabl e 16: I MT- Advanced Requi r ement s f or Cel l - Spect r al Ef f i ci ency
Test Env i r onment
118
Dow nl i nk ( bps/ Hz) Upl i nk ( bps/ Hz)
I ndoor 3. 0 2. 25
Microcellular 2. 6 1. 8
Base Coverage Urban 2. 2 1. 4
High Speed 1. 1 0. 7

Table 17 shows t he requirement s for voice capacit y.
Tabl e 17: I MT- Advanced Requi r ement s f or Voi ce Capaci t y
Test Env i r onment
119
Mi ni mum VoI P Capaci t y
( Act i v e User s/ Sect or / MHz)
I ndoor 50
Microcellular 40
Base Coverage Urban 40
High Speed 30

3GPP is addressing t he I MT- Advanced requirement s t hrough a version of LTE called LTE-
Advanced, a proj ect t hat is a st udy it em in 2009 wit h specificat ions expect ed in t he
second half of 2010 as part of Release 10. LTE- Advanced will be bot h backwards- and
forwards- compat ible wit h LTE, meaning LTE devices will operat e in newer LTE- Advanced
net works, and LTE- Advanced devices will operat e in older LTE net works.
3GPP is st udying t he following capabilit ies for LTE- Advanced:
- Wider bandwidt h support for up t o 100 MHz via aggregat ion of 20 MHz blocks.
- Uplink MI MO ( t wo t ransmit ant ennas in t he device) .
- Downlink MI MO of up t o 8 by 8 as described below.
- Coordinat ed mult ipoint t ransmission ( CoMP) wit h t wo proposed approaches:
coordinat ed scheduling and/ or beamforming, and j oint processing/ t ransmission.
The int ent is t o closely coordinat e t ransmissions at different cell sit es, t hereby
achieving higher syst em capacit y and improving cell- edge dat a rat es.
120

Figure 47 shows t he carrier aggregat ion, wit h up t o 100 MHz of bandwidt h support ed.

118
Test environment s are described in I T Report I TU- R M. 2135.
119
Test environment s are described in I T Report I TU- R M. 2135.
120
For furt her det ails, refer t o sect ion 7. 7. 5 of t he 3G Americas’ whit e paper “ The Mobile Broadband
Evolut ion: 3G Release 8 and Beyond, HSPA+ , SAE/ LTE and LTE- Advanced. ”

HSPA t o LTE- Advanced, Rysavy Research / 3G Americas, Sept ember 2009 Page 100
Fi gur e 47: Rel ease 10 LTE- Advanced Car r i er Aggr egat i on
121



Figure 48 shows t he carrier aggregat ion operat ing at different prot ocol layers.
Fi gur e 48: Car r i er Aggr egat i on at Di f f er ent Pr ot ocol Lay er s
122


Beyond wider bandwidt hs, LTE- Advanced will ext end performance t hrough more
powerful mult i- ant enna capabilit ies. For t he downlink, t he t echnology will be able t o
t ransmit in up t o 8 layers using an 8X8 configurat ion for a peak spect ral efficiency of 30
bps/ Hz t hat exceeds t he I MT- Advanced requirement s, conceivably support ing a peak
rat e of 1 Gbps in j ust 40 MHz and even higher rat es in wider bandwidt hs. This would
require addit ional reference signals for channel est imat ion and for measurement s such
as channel qualit y t o enable adapt ive, mult i- ant enna t ransmission. LTE- Advanced will

121
Source: "LTE for UMTS, OFDMA and SC- FDMA Based Radio Access, ” Harri Holma and Ant t i Toskala,
Wiley, 2009.
122
Source: “ The Evolut ion of LTE t owards I MT- Advanced” , St efan Parkvall and David Ast ely, Ericsson
Research, ht t p: / / www. academypublisher. com/ j cm/ vol04/ no03/ j cm0403146154. pdf
Rel’8
100 MHz bandwidth
Rel’8 Rel’8 Rel’8 Rel’8
Release 10 LTE-Advanced UE resource pool
Release 8 UE uses a
single 20 MHz block
20 MHz

HSPA t o LTE- Advanced, Rysavy Research / 3G Americas, Sept ember 2009 Page 101
also include four- layer t ransmission in t he uplink result ing in spect ral efficiency
exceeding 15 bps/ Hz.
Table 18 shows ant icipat ed performance relat ive t o I MT- Advanced Requirement s.
Tabl e 18: I MT- Advanced Requi r ement s and Ant i ci pat ed LTE- Advanced
Capabi l i t y .
I t em
I MT- Advanced
Requi r ement
LTE- Advanced
Pr oj ect ed Capabi l i t y
Peak Dat a Rat e Downlink 1 Gbps
Peak Dat a Rat e Uplink 500 Mbps
Spect rum Allocat ion Up t o 40 MHz Up t o 100 MHz
Lat ency User Plane 10 msec 10 msec
Lat ency Cont rol Plane 100 msec 50 msec
Peak Spect ral Efficiency DL
123
15 bps/ Hz 30 bps/ Hz
Peak Spect ral Efficiency UL 6. 75 bps/ Hz 15 bps/ Hz
Average Spect ral Efficiency DL 2. 2 bps/ Hz 2. 6 bps/ Hz
Average Spect ral Efficiency UL 1. 4 bps/ Hz 2. 0 bps/ Hz
Cell- Edge Spect ral Efficiency DL 0. 06 bps/ Hz 0. 09 bps/ Hz
Cell- Edge Spect ral Efficiency UL 0. 03 bps/ Hz 0. 07 bps/ Hz

I n all cases, proj ect ions of LTE- Advanced performance exceed t hat of t he I MT- Advanced
requirement s.
Anot her capabilit y being planned for LTE- Advanced is relays as shown in Figure 49.
The idea is t o relay frames at an int ermediat e node, result ing in much bet t er in- building
penet rat ion and, wit h bet t er signal qualit y, user rat es will be much improved.


123
Spect ral efficiency values based on 4 ant ennas at t he base st at ion and 2 ant ennas at t he t erminal.

HSPA t o LTE- Advanced, Rysavy Research / 3G Americas, Sept ember 2009 Page 102
Fi gur e 49: LTE- Advanced Rel ay
124



As demonst rat ed in t his sect ion, LTE- Advanced will have t remendous capabilit y. Though
init ial deployment s of LTE will be based on Release 8, as new spect rum becomes
available next decade, especially if it includes wide radio channels, t hen LTE- Advanced
will be t he ideal t echnology for t hese new bands. Even in exist ing bands, operat ors are
likely t o event ually upgrade t heir LTE net works t o LTE- Advanced t o obt ain spect ral
efficiency gains and capabilit ies such as relaying.
UMTS TDD
Most WCDMA and HSDPA deployment s are based on FDD, in which t he operat or uses
different radio bands for t ransmit and receive. An alt ernat e approach is TDD, in which
bot h t ransmit and receive funct ions alt ernat e in t ime on t he same radio channel. 3GPP
specificat ions include a TDD version of UMTS, called UMTS TDD.
TDD does not provide any inherent advant age for voice funct ions, which need balanced
links—namely, t he same amount of capacit y in bot h t he uplink and t he downlink. Many
dat a applicat ions, however, are asymmet ric, oft en wit h t he downlink consuming more
bandwidt h t han t he uplink, especially for applicat ions like Web browsing or mult imedia
downloads. A TDD radio int erface can dynamically adj ust t he downlink- t o- uplink rat io
accordingly, hence balancing bot h forward- link and reverse- link capacit y. Not e t hat for
UMTS FDD, t he higher spect ral efficiency achievable in t he downlink versus t he uplink is
crit ical in addressing t he asymmet rical nat ure of most dat a t raffic.
The UMTS TDD specificat ion also includes t he capabilit y t o use j oint det ect ion in
receiver- signal processing, which offers improved performance.
One considerat ion, however, relat es t o available spect rum. Various count ries around t he
world including t hose in Europe, Asia, and t he Pacific region have licensed spect rum
available specifically for TDD syst ems. For t his spect rum, UMTS TDD, or in t he fut ure
LTE in TDD mode, is a good choice. I t is also a good choice in any spect rum t hat does
not provide a duplex gap bet ween forward and reverse links.
I n t he Unit ed St at es, t here is limit ed spect rum specifically allocat ed for TDD syst ems.
125

UMTS TDD is not a good choice in FDD bands; it would not be able t o operat e effect ively
in bot h bands, t hereby making t he overall syst em efficiency relat ively poor.

124
Source: 3G Americas’ member cont ribut ion.
Relay Link
Access
Link
Direct Link

HSPA t o LTE- Advanced, Rysavy Research / 3G Americas, Sept ember 2009 Page 103
As discussed in more det ail in t he “ WiMAX” sect ion, TDD syst ems require net work
synchronizat ion and careful coordinat ion bet ween operat ors or guardbands, which may
be problemat ic in cert ain bands.
There has been lit t le deployment of UMTS TDD. Fut ure TDD deployment s of 3GPP
t echnologies are likely t o be based on LTE.
TD-SCDMA
TD- SCDMA is one of t he official 3G wireless t echnologies being developed, most ly for
deployment in China. Specified t hrough 3GPP as a variant of t he UMTS TDD Syst em and
operat ing wit h a 1. 28 megachips per second ( Mcps) chip rat e against 3. 84 Mcps for
UMTS TDD, t he primary at t ribut e of TD- SCDMA is t hat it is designed t o support very high
subscriber densit ies. This makes it a possible alt ernat ive for wireless local loops. TD-
SCDMA uses t he same core net work as UMTS, and it is possible for t he same core
net work t o support bot h UMTS and TD- SCDMA radio- access net works.
TD- SCDMA t echnology is not as mat ure as UMTS and CDMA2000, wit h 2008 being t he
first year of limit ed deployment s in China in t ime for t he Olympic Games. Alt hough t here
are no planned deployment s in any count ry ot her t han China, TD- SCDMA could
t heoret ically be deployed anywhere unpaired spect rum is available—such as t he bands
licensed for UMTS TDD—assuming appropriat e resolut ion of regulat ory issues.
IMS
I MS is a service plat form t hat allows operat ors t o support I P mult imedia applicat ions.
Pot ent ial applicat ions include video sharing, PoC, VoI P, st reaming video, int eract ive
gaming, and so fort h. I MS by it self does not provide all t hese applicat ions. Rat her, it
provides a framework of applicat ion servers, subscriber dat abases, and gat eways t o
make t hem possible. The exact services will depend on cellular operat ors and t he
applicat ion developers t hat make t hese applicat ions available t o operat ors.
The core net working prot ocol used wit hin I MS is Session I nit iat ion Prot ocol ( SI P) , which
includes t he companion Session Descript ion Prot ocol ( SDP) used t o convey configurat ion
informat ion such as support ed voice codecs. Ot her prot ocols include Real Time Transport
Prot ocol ( RTP) and Real Time St reaming Prot ocol ( RTSP) for t ransport ing act ual
sessions. The QoS mechanisms in UMTS will be an import ant component of some I MS
applicat ions.
Alt hough originally specified by 3GPP, numerous ot her organizat ions around t he world
are support ing I MS. These include t he I nt ernet Engineering Taskforce ( I ETF) , which
specifies key prot ocols such as SI P, and t he Open Mobile Alliance, which specifies end-
t o- end, service- layer applicat ions. Ot her organizat ions support ing I MS include t he GSM
Associat ion ( GSMA) , t he ETSI , CableLabs, 3GPP2, The Parlay Group, t he I TU, t he
American Nat ional St andards I nst it ut e ( ANSI ) , t he Telecoms and I nt ernet Converged
Services and Prot ocols for Advanced Net works ( TI SPAN) , and t he Java Communit y
Process ( JCP) .
I MS is relat ively independent of t he radio- access net work and can, and likely will, be
used by ot her radio- access net works or wireline net works. Operat ors already have I MS
t rails in progress, and one init ial applicat ion under considerat ion—PoC—is being specified
by t he Open Mobile Alliance. Ot her applicat ions include pict ure and video sharing t hat
occur in parallel wit h voice communicat ions. Operat ors looking t o roll out VoI P over

125
The 1910- 1920 MHz band t arget ed unlicensed TDD syst ems, but has never been used.

HSPA t o LTE- Advanced, Rysavy Research / 3G Americas, Sept ember 2009 Page 104
net works could also use I MS. 3GPP init ially int roduced I MS in Release 5 and has
enhanced it in each subsequent specificat ion release.
As shown in Figure 50, I MS operat es j ust out side t he packet core.
Fi gur e 50: I P Mul t i medi a Subsy st em

The benefit s of using I MS include handling all communicat ion in t he packet domain,
t ight er int egrat ion wit h t he I nt ernet , and a lower cost infrast ruct ure t hat is based on I P
building blocks used for bot h voice and dat a services. This allows operat ors t o pot ent ially
deliver dat a and voice services at lower cost , t hus providing t hese services at lower
prices and furt her driving demand and usage.
I MS applicat ions can reside eit her in t he operat or’s net work or in t hird- part y net works
including t hose of ent erprises. By managing services and applicat ions cent rally—and
independent ly of t he access net work—I MS can enable net work convergence. This allows
operat ors t o offer common services across 3G, Wi- Fi, and wireline net works.
Broadcast/Multicast Services
An import ant capabilit y for 3G and evolved 3G syst ems is broadcast ing and mult icast ing,
wherein mult iple users receive t he same informat ion using t he same radio resource. This
creat es a much more efficient approach for delivering cont ent such as video
programming t o which mult iple users have subscript ions. I n a broadcast , every
subscriber unit in a service area receives t he informat ion, whereas in a mult icast , only
users wit h subscript ions receive t he informat ion. Service areas for bot h broadcast and
mult icast can span eit her t he ent ire net work or a specific geographical area. Because
mult iple users in a cell are t uned t o t he same cont ent , broadcast ing and mult icast ing
result in much great er spect rum efficiency for services such as mobile TV.
3GPP defined highly- efficient broadcast / mult icast capabilit ies for UMTS in Release 6 wit h
MBMS. Release 7 includes opt imizat ions t hrough a solut ion called mult icast / broadcast ,
single- frequency net work operat ion t hat involves simult aneous t ransmission of t he exact
waveform across mult iple cells. This enables t he receiver t o const ruct ively superpose
Call Session Control Function ( CSCF)
( SIP Proxy)
Home Subscriber
Server ( HSS)
SIP Application
Server
SIP
DI AMETER
I MS
UMTS/HSPA
Packet Core
Network
Media Resource
Function Control
Media Resource
Gateway Cont rol
Wi-Fi DSL
Multiple Possible Access Net wor ks

HSPA t o LTE- Advanced, Rysavy Research / 3G Americas, Sept ember 2009 Page 105
mult iple MBSFN cell t ransmissions. The result is highly efficient , WCDMA- based
broadcast t ransmission t echnology t hat mat ches t he benefit s of OFDMA- based broadcast
approaches.
LTE will also have a broadcast / mult icast capabilit y. OFDM is part icularly well- suit ed for
broadcast ing, because t he mobile syst em can combine t he signal from mult iple base
st at ions and because of t he narrowband nat ure of OFDM. Normally, t hese signals would
int erfere wit h each ot her. As such, t he LTE broadcast capabilit y is expect ed t o be quit e
efficient .
Fi gur e 51: OFDM Enabl es Ef f i ci ent Br oadcast i ng

An alt ernat e approach for mobile TV is t o use an ent irely separat e broadcast net work
wit h t echnologies such as Digit al Video Broadcast ing–Handheld ( DVB- H) or Media
Forward Link Only ( MediaFLO) , which various operat ors around t he world have opt ed t o
do. Alt hough t his requires a separat e radio in t he mobile device, t he net works are highly
opt imized for broadcast .
EPC/SAE
3GPP is defining EPC/ SAE in Release 8 as a framework for an evolut ion or migrat ion of
t he 3GPP syst em t o a higher- dat a- rat e, lower- lat ency, packet - opt imized syst em t hat
support s mult iple radio- access t echnologies. The focus of t his work is on t he packet -
swit ched domain wit h t he assumpt ion t hat t he syst em will support all services—including
voice—in t his domain.
Alt hough it will most likely be deployed in conj unct ion wit h LTE, EPC/ SAE could also be
deployed for use wit h HSPA+ where it could provide a st epping- st one t o LTE. EPC/ SAE
will be opt imized for all services t o be delivered via I P in a manner t hat is as efficient as
possible—t hrough minimizat ion of lat ency wit hin t he syst em, for example. I t will support
service cont inuit y across het erogeneous net works, which will be import ant for LTE
operat ors who must simult aneously support GSM- HSPA cust omers.
One import ant performance aspect of EPC/ SAE is a flat t er archit ect ure. For packet flow,
EPC/ SAE includes t wo net work element s, called Evolved Node B ( eNodeB) and t he
Access Gat eway ( AGW) . The eNodeB ( base st at ion) int egrat es t he funct ions t radit ionally

HSPA t o LTE- Advanced, Rysavy Research / 3G Americas, Sept ember 2009 Page 106
performed by t he radio- net work cont roller, which previously was a separat e node
cont rolling mult iple Node Bs. Meanwhile, t he AGW int egrat es t he funct ions t radit ionally
performed by t he SGSN and GGSN. The AGW has bot h cont rol funct ions, handled
t hrough t he Mobile Management Ent it y ( MME) , and user plane ( dat a communicat ions)
funct ions. The user plane funct ions consist of t wo element s: A serving gat eway t hat
addresses 3GPP mobilit y and t erminat es eNodeB connect ions, and a Packet Dat a
Net work ( PDN) gat eway t hat addresses service requirement s and also t erminat es access
by non- 3GPP net works. The MME, serving gat eway, and PDN gat eways can be collocat ed
in t he same physical node or dist ribut ed, based on vendor implement at ions and
deployment scenarios.
The EPC/ SAE archit ect ure is similar t o t he HSPA One- Tunnel Archit ect ure discussed in
t he “ HSPA+ ” sect ion t hat allows for easy int egrat ion of HSPA net works t o t he EPC/ SAE.
Anot her archit ect ural opt ion is t o reverse t he t opology, so t hat t he EPC/ SAE Access
Gat eway is locat ed close t o t he RAN in a dist ribut ed fashion t o reduce lat ency, while t he
MME is cent rally locat ed t o minimize complexit y and cost .
EPC/ SAE also allows int egrat ion of non- 3GPP net works such as WiMAX. EPC/ SAE will use
I MS as a component . I t will also manage QoS across t he whole syst em, which will be
essent ial for enabling a rich set of mult imedia- based services.
Figure 52 shows t he EPC/ SAE archit ect ure.
Fi gur e 52: EPC/ SAE Ar chi t ect ur e



MME
GERAN
UTRAN
Rel’7 Legacy GSM/ UMTS
SGSN
Evolved RAN,
e.g., LTE
Serving
Gateway
PDN
Gateway
Non 3GPP
IP Access
PCRF
IP
Services,
IMS
EPC/ SAE Access Gat eway
Cont rol
User Plane
One-Tunnel
Opt ion

HSPA t o LTE- Advanced, Rysavy Research / 3G Americas, Sept ember 2009 Page 107
Element s of t he SAE archit ect ure include:
 Support for legacy GERAN and UTRAN net works connect ed via SGSN.
 Support for new radio- access net works such as LTE.
 The Serving Gat eway t hat t erminat es t he int erface t oward t he 3GPP radio- access
net works.
 The PDN gat eway t hat cont rols I P dat a services, does rout ing, allocat es I P
addresses, enforces policy, and provides access for non- 3GPP access net works.
 The MME t hat support s user equipment cont ext and ident it y, as well as
aut hent icat ing and aut horizing users.
 The Policy Cont rol and Charging Rules Funct ion ( PCRF) t hat manages QoS
aspect s.
3GPP is planning t o support voice in EPS t hrough VoI P and I MS. However, t here is an
alt ernat ive voice approach being discussed in t he indust ry, namely t ransport ing circuit -
swit ched voice over LTE, called Voice over LTE Generic Access ( VOLGA) . This approach is
not current ly part of any 3GPP specificat ions.
The need for support ing a broader variet y of applicat ions requiring higher bandwidt h and
lower lat ency led 3GPP t o alleviat e t he exist ing ( UMTS Release 99) Qualit y of Service
( QoS) principles wit h t he int roduct ion for EPS of a QoS Class I dent ifier ( QCI ) . The QCI is
a scalar denot ing a set of t ransport charact erist ics ( bearer wit h/ wit hout guarant eed bit
rat e, priorit y, packet delay budget , packet error loss rat e) and used t o infer nodes
specific paramet ers t hat cont rol packet forwarding t reat ment ( e. g. , scheduling weight s,
admission t hresholds, queue management t hresholds, link- layer prot ocol configurat ion,
et c. ) . Each packet flow is mapped t o a single QCI value ( nine are defined in t he Release
8 version of t he specificat ions) according t o t he level of service required by t he
applicat ion. The usage of t he QCI avoids t he t ransmission of a full set of QoS- relat ed
paramet ers over t he net work int erfaces and reduces t he complexit y of QoS negot iat ion.
The QCI , t oget her wit h Allocat ion- Ret ent ion Priorit y ( ARP) and, if applicable, Guarant eed
Bit Rat e ( GBR) and Maximum Bit Rat e ( MBR) , det ermines t he QoS associat ed t o an EPS
bearer. A mapping bet ween EPS and pre- Release 8 QoS paramet ers has been defined t o
allow proper int erworking wit h legacy net works.
White Space
The FCC in t he US has ruled t hat unlicensed devices t hat have mechanisms t o not
int erfere wit h TV broadcast channels may use TV channels t hat are not in use.
126
The
rules provide for fixed devices and personal/ port able devices. The FCC has suggest ed
t wo usage t ypes: broadband services t o homes and businesses at a higher power level
t o fixed devices over larger geographical areas; and wireless port able devices at a low-
power level in indoor environment s.
To prevent int erference wit h TV t ransmissions, bot h device t ypes must employ geo-
locat ion capabilit y wit h 50- met er accuracy ( alt hough fixed devices can st ore t heir
posit ion during inst allat ion) , as well as having t he abilit y t o access a dat abase t hat list s
permit t ed channels for a specific locat ion. I n addit ion, all devices must be able t o sense
t he spect rum t o det ect bot h TV broadcast ing and wireless microphone signals. The rules
include t ransmit power limit s and emission limit s.

126
FCC- 08- 260: 2nd Repor t & Order.

HSPA t o LTE- Advanced, Rysavy Research / 3G Americas, Sept ember 2009 Page 108
The frequency- sensing and channel- change requirement s are not support ed by t oday’s
3GPP, 3GPP2 and WiMAX t echnologies. The I EEE, however, has developed a st andard,
I EEE 802. 22, based on I EEE 802. 16 concept s, t hat complies wit h t he FCC requirement s.
I EEE 802. 22 is aimed at fixed or nomadic services such as DSL replacement .
The indust ry is in t he very early st ages of det ermining t he viabilit y of using whit e- space
spect rum and, at t his t ime, t here are no product s or services available.


HSPA t o LTE- Advanced, Rysavy Research / 3G Americas, Sept ember 2009 Page 109
Acronyms
The following acronyms are used in t his paper. Acronyms are defined on first use.
1xEV- DO – One Carrier Evolved, Dat a Opt imized
1xEV- DV – One Carrier Evolved, Dat a Voice
1XRTT – One Carrier Radio Transmission Technology
2G – Second Generat ion
3G – Third Generat ion
3GPP – Third Generat ion Part nership Proj ect
3GPP2 – Third Generat ion Part nership Proj ect 2
4G – Fourt h Generat ion ( meet ing requirement s set fort h by t he I TU I MT- Advanced proj ect )
8- PSK – Oct agonal Phase Shift Keying
AAS – Adapt ive Ant enna Syst ems
ABR – Allocat ion Ret ent ion Priorit y
AGW – Access Gat eway
AMR – Adapt ive Mult i Rat e
ANSI – American Nat ional St andards I nst it ut e
ARP – Allocat ion Ret ent ion Priorit y
ARQ – Aut omat ic Repeat Request
ARPU – Average Revenue Per User
ATM – Asynchronous Transfer Mode
AWGN – Addit ive Whit e Gaussian Noise Channel
BCCH – Broadcast Cont rol Channel
bps – bit s per second
BRS – Broadband Radio Service
BSC – Base St at ion Cont roller
BTS – Base Transceiving St at ion
C/ I – Carrier t o I nt erference Rat io
CAPEX- Capit al Expendit ure
CDF – Cumulat ive Dist ribut ion Funct ion
CDMA – Code Division Mult iple Access
CMOS – Complement ary Met al Oxide Semiconduct or
CP – Cyclic Prefix
CPC – Cont inuous Packet Connect ivit y
CRM – Cust omer Relat ionship Management
DAS – Downlink EGPRS2- A Level Scheme
dB – Decibel
DBS – Downlink EGPRS2- B Level Scheme
DC- HSPA – Dual Carrier HSPA
DFT – Discret e Fourier Transform

HSPA t o LTE- Advanced, Rysavy Research / 3G Americas, Sept ember 2009 Page 110
DSL – Digit al Subscriber Line
DTM – Dual Transfer Mode
D- TxAA – Double Transmit Adapt ive Array
DVB- H – Digit al Video Broadcast ing Handheld
E–DCH – Enhanced Dedicat ed Channel
EBCMCS – Enhanced Broadcast Mult icast Services
EDGE – Enhanced Dat a Rat es for GSM Evolut ion
EGPRS – Enhanced General Packet Radio Service
eNodeB – Evolved Node B
EPS – Evolved Packet Syst em
ERP – Ent erprise Resource Planning
ETRI – Elect ronic and Telecommunicat ions Research I nst it ut e
ETSI – European Telecommunicat ions I nst it ut e
E- UTRAN – Enhanced UMTS Terrest rial Radio Access Net work
EV- DO – One Carrier Evolved, Dat a Opt imized
EV- DV – One Carrier Evolved, Dat a Voice
EVRC – Enhanced Variable Rat e Codec
FCC – Federal Communicat ions Commission
FDD – Frequency Division Duplex
Flash OFDM – Fast Low- Lat ency Access wit h Seamless Handoff OFDM
FLO – Forward Link Only
FMC – Fixed Mobile Convergence
FP7 – Sevent h Framework Programme
FTP – File Transfer Prot ocol
G- Rake – Generalized Rake Receiver
Gbps – Gigabit s Per Second
GBR – Guarant eed Bit Rat e
GERAN – GSM EDGE Radio Access Net work
GGSN – Gat eway GPRS Support Node
GHz — Gigahert z
GMSK – Gaussian Minimum Shift Keying
GPRS – General Packet Radio Service
GSM – Global Syst em for Mobile communicat ions
GSMA – GSM Associat ion
HARQ – Hybrid Aut omat ic Repeat Request
HD – High Definit ion
HLR – Home Locat ion Regist er
HSDPA – High Speed Downlink Packet Access
HS- FACH – High Speed Forward Access Channel
HS- PDSCH - High Speed Physical Downlink Shared Channels

HSPA t o LTE- Advanced, Rysavy Research / 3G Americas, Sept ember 2009 Page 111
HSPA – High Speed Packet Access ( HSDPA wit h HSUPA)
HSPA+ – HSPA Evolut ion
HSUPA – High Speed Uplink Packet Access
Hz – Hert z
I CT – I nformat ion and Communicat ion Technologies
I EEE – I nst it ut e of Elect rical and Elect ronic Engineers
I ETF – I nt ernet Engineering Taskforce
I FFT – I nverse Fast Fourier Transform
I M – I nst ant Messaging
I MS – I P Mult imedia Subsyst em
I MT – I nt ernat ional Mobile Telecommunicat ions
I PR - I nt ellect ual Propert y Right s
I P – I nt ernet Prot ocol
I PTV – I nt ernet Prot ocol Television
I R – I ncrement al Redundancy
I SI – I nt ersymbol I nt erference
I SP – I nt ernet Service Provider
I TU – I nt ernat ional Telecommunicat ions Union
JCP – Java Communit y Process
kbps – Kilobit s Per Second
kHz — Kilohert z
km – Kilomet er
LSTI – LTE/ SAE Trial I nit iat ive
MAC – Medium Access Cont rol
MBMS - Mult imedia Broadcast / Mult icast Service
Mbps – Megabit s Per Second
MBR – Maximum Bit Rat e
Mcps – Megachips Per Second
MCS – Modulat ion and Coding Scheme
MediaFLO – Media Forward Link Only
MHz – Megahert z
MI MO – Mult iple I nput Mult iple Out put
mI TF – Japan Mobile I T Forum
MMDS – Mult ichannel Mult ipoint Dist ribut ion Service
MME – Mobile Management Ent it y
MMSE – Minimum Mean Square Error
MRxD – Mobile Receive Diversit y
MS – Mobile St at ion
MSA – Mobile Service Archit ect ure
MSC – Mobile Swit ching Cent er

HSPA t o LTE- Advanced, Rysavy Research / 3G Americas, Sept ember 2009 Page 112
MU- MI MO – Mult i- User MI MO
msec – millisecond
NGMC – Next Generat ion Mobile Commit t ee
OFDM – Ort hogonal Frequency Division Mult iplexing
OFDMA – Ort hogonal Frequency Division Mult iple Access
PAR – Peak t o Average Rat io
PBCCH – Packet Broadcast Cont rol Channel
PCRF – Policy Cont rol and Charging Rules Funct ion
PCS – Personal Communicat ions Service
PHY – Physical Layer
PDN – Packet Dat a Net work
PoC – Push- t o- t alk over Cellular
QAM – Quadrat ure Amplit ude Modulat ion
QCI – Qualit y of Service Class I dent ifier
QoS – Qualit y of Service
QPSK – Quadrat ure Phase Shift Keying
RAB – Radio Access Bearer
RAN – Radio Access Net work
RCS – Rich Communicat ions Suit e
REST – Represent at ional St at e Transfer
RF – Radio Frequency
RNC – Radio Net work Cont roller
ROHC – Robust Header Compression
RTP – Real Time Transport Prot ocol
RTSP – Real Time St reaming Prot ocol
SC- FDMA – Single Carrier Frequency Division Mult iple Access
SAE – Syst em Archit ect ure Evolut ion
SDMA – Space Division Mult iple Access
SDP – Session Descript ion Prot ocol
SGSN – Serving GPRS Support Node
SI C – Successive I nt erference Cancellat ion
SI P – Session I nit iat ion Prot ocol
SMS – Short Message Service
SNR – Signal t o Noise Rat io
SU- MI MO – Single User MI MO
TCH – Traffic Channel
TDD – Time Division Duplex
TDMA – Time Division Mult iple Access
TD- SCDMA – Time Division Synchronous Code Division Mult iple Access
TD- CDMA – Time Division Code Division Mult iple Access

HSPA t o LTE- Advanced, Rysavy Research / 3G Americas, Sept ember 2009 Page 113
TI A/ EI A – Telecommunicat ions I ndust ry Associat ion/ Elect ronics I ndust ry Associat ion
TI SPAN – Telecoms and I nt ernet converged Services and Prot ocols for Advanced Net works
TTI – Transmission Time I nt erval
UAS – Uplink EGPRS2- A Level Scheme
UBS – Uplink EGPRS2- B Level Scheme
UMA – Unlicensed Mobile Access
UMB – Ult ra Mobile Broadband
UMTS – Universal Mobile Telecommunicat ions Syst em
us – Microseconds
UTRAN – UMTS Terrest rial Radio Access Net work
VDSL – Very High Speed DSL
VoI P – Voice over I nt ernet Prot ocol
VOLGA – Voice over LTE Generic Access
VPN – Virt ual Privat e Net work
WAP – Wireless Applicat ion Prot ocol
WCDMA – Wideband CDMA
Wi- Fi – Wireless Fidelit y
WiMAX – Worldwide I nt eroperabilit y for Microwave Access
WLAN – Wireless Local Area Net work
WMAN – Wireless Met ropolit an Area Net work
WRC- 07 – World Radiocommunicat ion Conference 2007
Additional Information
3G Americas maint ains complet e and current list s of market informat ion including EDGE,
UMTS, and HSDPA deployment s worldwide, available for free download on it s Web sit e:
ht t p: / / www. 3gamericas. org.
I f t here are any quest ions regarding t he download of t his informat ion, please call + 1 425
372 8922 or e- mail Krissy Gochnour, Public Relat ions Administ rat or, at
info@3gamericas. org. ”
References
3G Americas: “ Mobile Broadband: The Global Evolut ion of UMTS- HSPA – 3GPP Release 7
and Beyond, ” July 2006.
3G Americas: “ UMTS Evolut ion from 3GPP Release 7 t o Release 8, HSPA and SAE/ LTE” , July
2007.
3G Americas: “ UMTS Evolut ion from 3GPP Release 7 t o Release 8, HSPA and SAE/ LTE” , June
2008 Updat e.
3G Americas: “ Global UMTS and HSPA Operat or St at us” , July 11, 2008.
3GPP: LTE Performance Summary, Downlink, Uplink, VoI P. Mult i- vendor assessment , 2007.
ABI Research: press release on st udy “ Mobile Business Applicat ions and Services, ” August
1, 2007.

HSPA t o LTE- Advanced, Rysavy Research / 3G Americas, Sept ember 2009 Page 114
Alcat el Lucent : “ CDMA2000 Pat h t o LTE” , Sam Samra, Senior Direct or- Technology
Programs, CDMA Development Group, ATI S 3GPP LTE Conference, Dallas, TX, January 26,
2009.
Alcat el Lucent : LTE TDD Harmonizat ion, May 2009, submission t o 3G Americas.
Alcat el Lucent : LTE Migrat ion Scenarios, May 2009, submission t o 3G Americas.
Alcat el Lucent : ” Technology Comparison” , June 2009, submission t o 3G Americas.
Alcat el Lucent : “ TV Whit e Space, ” May 2009, submission t o 3G Americas.
Art hur D Lit t le: “ HSPA and Mobile WiMAX for Mobile Broadband Wireless Access – An
I ndependent Report Prepared for t he GSM Associat ion, March 27, 2007.
AT&T: “ I deas on LTE- Advanced” , 3GPP LTE- Advanced Workshop, April 2008.
AT&T: Tom Keat hley, “ HSPA: Keys t o a Successful Broadband Access St rat egy” , 2008.
Berg I nsight : Smart phone Operat ing Syst ems,
ht t p: / / www. berginsight . com/ Report PDF/ Product Sheet / BI - SOS- PS. pdf, July 2007.
CDMA Developer Group: Deployment s, ht t p: / / www. cdg. org, June 5, 2009.
Cingular Wireless: uplink t hroughput dat a, June 2006, submission t o 3G Americas.
Cisco: “ Approaching t he Zet t abyt e Era, ” June 16, 2008
Cisco: “ Visual Net working I ndex: Global Mobile Dat a Traffic, Forecast Updat e, ” January 29,
2009.
Dow Jones NY: “ 3G Ret ains it s Buzz but Pot ent ial Remains Unclear, ” January 27, 2006.
GSA: The Global mobile Suppliers Associat ion, ht t p: / / www. gsacom. com, July 2007.
Ericsson: “ The 3G Long- Term Evolut ion – Radio I nt erface Concept s and Performance
Evaluat ion, ” 2006.
Ericsson: “ 3GPP I mproved UE Receiver Requirement s, ” June 2006, submission t o 3G
Americas.
Ericsson whit e paper: “ Basic Concept s of HSPA” , February 2007.
Ericsson: “ Cellular Evolut ion, ” May 2006, submission t o 3G Americas.
Ericsson: “ The Evolut ion of LTE t owards I MT- Advanced” , St efan Parkvall and David Ast ely,
ht t p: / / www. academypublisher. com/ j cm/ vol04/ no03/ j cm0403146154. pdf
Ericsson: “ HSPA and WiMAX Performance, ” July 2007, submission t o 3G Americas.
Ericsson: “ HSPA Spect rum Efficiency Evolut ion” , June 2008, submission t o 3G Americas.
Ericsson: Johan Bergman et al, “ HSPA Evolut ion – Boost ing t he performance of mobile
broadband access” , Ericsson Review No. 1, 2008.
Ericsson whit e paper: “ HSPA, t he Undisput ed Choice for Mobile Broadband, ” May 2007.
Ericsson whit e paper: “ HSDPA Performance and Evolut ion” , No 3, 2006.
Ericsson: “ I nit ial Field Performance Measurement s of LTE, ” Jonas Karlsson, Mat hias Riback,
Ericsson Review No. 3, 2008,
ht t p: / / www. ericsson. com/ ericsson/ corpinfo/ publicat ions/ review/ 2008_03/ files/ LTE. pdf
Ericsson: Per Beming et al, “ LTE- SAE archit ect ure and performance” , Ericsson Review No. 3,
2007.

HSPA t o LTE- Advanced, Rysavy Research / 3G Americas, Sept ember 2009 Page 115
Ericsson: Erik Dahlman et al, “ Key feat ures of t he LTE radio int erface” , Ericsson Review No.
2, 2008.
Ericsson: “ Mobile Broadband Backhaul: Addressing t he Challenge, ” Raj esh Chundury,
Ericsson Review No. 3, 2008.
Ericsson whit e paper: “ Technical Overview and Performance of HSPA and Mobile WiMAX” ,
June 2007.
Ericsson: “ Advanced Receivers for WCDMA Terminal Plat forms and Base St at ions, ” Ericsson
Review No. 2, 2006.
Ericsson: “ The Evolut ion of LTE t owards I MT- Advanced, ” St efan Parkvall and David Ast ely,
2009.
Ericsson: “ GSM/ EDGE Cont inued Evolut ion, ” Ericsson Review No. 1, 2006.
Ericsson: “ HSDPA performance - HSDPA and R99 on Same Carrier, ” June 2006, submission
t o 3G Americas.
Ericsson: HSPA voice migrat ion, June 2006, submission t o 3G Americas.
Ericsson: “ WCDMA vs. CDMA Business View, ” 2006, submission t o 3G Americas.
Ericsson whit e paper: “ WiMAX – Copper in t he Air, ” April 2006.
Fierce Wireless: Press release, ht t p: / / www. fiercewireless. com/ press- releases/ led- asia-
pacific- suppliers- cellular- modem- indust ry- will- exceed- 200- million- unit s- 2013, 2008.
Senza Fili Consult ing: Press release of June 19, 2007 describing t he report "WiMAX:
Ambit ions and Realit y. A det ailed market assessment and forecast at t he global, regional
and count ry level ( 2006- 2012) ".
Forward Concept s, Mobile I nt ernet Device and Chip Market Opport unit ies, June 2008.
Gart ner Group: “ Forecast : Mobile Terminals, Worldwide, 2000- 2009 ( 4Q05 Updat e) , ”
January 12, 2006.
Global Telecom I nsider report , "LTE’s Five- Year Global Forecast : Poised t o Grow Fast er t han
3G, " 2009
GSA: HSPA Devices Survey, July 21, 2008
GSM Associat ion: “ Mobile phones on t he cat walk” by Paul Rasmussen. Wireless Business
Review, Spring 2006.
I EEE: Communicat ions Magazine, Mo- Han Fong and Robert Novak, Nort el Net works, Sean
McBeat h, Huawei Technologies, Roshni Srinivasan, I nt el Corporat ion, “ I mproved VoI P
Capacit y in Mobile WiMAX Syst ems Using Persist ent Resource Allocat ion, ” Oct ober, 2008.
I EEE I nt ernat ional Symposium on Personal, I ndoor and Mobile Radio Communicat ions:
Anders Furuskär et al “ The LTE Radio I nt erface – Key Charact erist ics and Performance” ,
2008.
I EEE: Journal on Select ed Areas in Communicat ion, Vol 24, No. 1, Qi Bi, “ An Analysis of VoI P
Service Using 1 EV- DO Revision A Syst em, ” January 2006.
I nforma Telecoms & Media, press release, “ Mobile dat a revenue t o exceed US$200 billion in
2007, ” July 23, 2008
I nforma Telecoms & Media: WiMAX proj ect ion, June 2009, supplied t o 3G Americas.
I nforma Telecoms & Media: World Cellular I nformat ion Service, May 2009.

HSPA t o LTE- Advanced, Rysavy Research / 3G Americas, Sept ember 2009 Page 116
iSuppli Corp. : “ Mobile- Phone Premium Cont ent Market t o Reach $40 billion by 2010, ”
www. cellular- news. com/ st ory/ 16425. php, March 8, 2006.
I nt ernat ional Telecommunicat ions Union, “ Report I TU- R M. 2134, Requirement s relat ed t o
t echnical performance for I MT- Advanced radio int erface( s) , ” 2009.
Juniper Research, LTE Report , July 2009.
Lehman Global Equit y Research: Paul Wuh, “ Global 3G Development s: 3G subs accelerat e;
more dat a revenue in ’09” May 23, 2008.
Art hur D. Lit t le Limit ed: "HSPA and mobile WiMAX for Mobile Broadband Wireless Access",
27 March 2007.
"LTE for UMTS, OFDMA and SC- FDMA Based Radio Access, ” Harri Holma and Ant t i Toskala,
Wiley, 2009.
Maravedis: “ WiMAX and Broadband Wireless Access Equipment Market Analysis, Trends and
Forecast s, 2009- 2014” , June 1, 2009.
LSTI Forum: "LTE/ SAE Trial I nit iat ive Lat est Result s from t he LSTI ” , Feb 2009.
Nokia: “ 3GPP vs. 3GPP2 Cellular VoI P Driver Comparison, ” June 2006, submission t o 3G
Americas.
Nokia: “ HSDPA Performance Measurement s wit h Commercial QPSK ( CAT 12) and 16QAM
( CAT 6) , ” June 2006, submission t o 3G Americas.
Nokia: “ Fut ure Voice Traffic: Primary vs. Secondary?” June 2006 submission t o 3G
Americas.
Nokia: “ VoI P over HSPA wit h 3GPP Release 7, ” by Harri Holma, et al, 2006.
Nokia: “ Overview on HSPA+ , ” June 2006, submission t o 3G Americas.
Nokia: “ HSUPA Simulat ion Result s, ” May 2006, submission t o 3G Americas.
Nokia: “ SAE Evolved Archit ect ure, ” June 2006, submission t o 3G Americas.
Nokia: “ WCDMA CS vs. HSPA VoI P Capacit y Difference, ” June 2006, submission t o 3G
Americas.
Nokia Siemens Net works: “ HSPA/ LTE Performance, ” May 2009, submission t o 3G Americas.
Nokia Siemens Net works, “ Performance Evaluat ion on Dual- Cell HSDPA Operat ion, ” Danielle
Morais de Andrade, Axel Klein, Harri Holma, 2009.
Nort el whit e paper: “ GSM t o LTE Evolut ion, ” June 2007.
Ovum Comment , Adam Leach, Devices principal analyst , “ Smart phones: t he silver lining of
t he declining handset market , ” June 2009.
Ovum, Telecom and Soft ware News, July 2, 2009.
Port io Research: “ Mobile Dat a Services Market s 2008” , June 11, 2008.
Pyramid Research: “ Europe t o See Huge Growt h in Mobile Broadband Services despit e
Recession, ” 2009.
Qualcomm: press release,
ht t p: / / www. qualcomm. com/ press/ releases/ 2008/ 080207_Qualcomm_t o_Ship. ht ml.
Research in Mot ion: “ Evolut ion of High Speed Wireless Dat a St andards in 3GPP” , May 15,
2007.

HSPA t o LTE- Advanced, Rysavy Research / 3G Americas, Sept ember 2009 Page 117
Rysavy Research art icle: “ Reach Me if You Can, ” ht t p: / / www. rysavy. com/ papers. ht ml, May
2007.
Rysavy Research: “ Hard Numbers and Expert s' I nsight s on Migrat ion t o 4G Wireless
Technology, ” published by Dat acomm Research, February 2005.
Rysavy Research, “ Mobile Broadband Spect rum Demand, ” December 2008.
Andy Seybold: “ Will Dat a- Only Net works Ever Make Money?” January 18, 2006
comment ary, ht t p: / / www. out look4mobilit y. com/ comment ary2006/ j an1806. ht m.
Chet an Sharma: “ Managing Growt h and Profit s in t he Yot t abyt e Era” , July 2009.
Chet an Sharma: “ US Wireless Dat a Market Updat e - Q1 2009. ”
SNL Kagan: press release, “ SNL Kagan Expect s Wireless Dat a Revenue t o I ncrease at a
16% Annual Rat e over t he Next Decade” , ht t p: / / www1. snl. com/ press/ 20080731. asp, July
31, 2008
Sprint : Press release, January 30, 2007.
Sprint Next el: Ali Tabassi, Fierce Wireless Webcast , “ WiMAX: Mobilizing t he I nt ernet ” , March
5, 2008.
Telst ra: present at ion, “ HSPA as an Open Eco- Syst em Today – Telst ra Next G Net work” ,
2008.
Unit ed St at es Census Bureau, ht t p: / / www. census. gov/ ipc/ www/ idb/ worldpopinfo. ht ml,
2009.
Verizon Wireless: Verizon Broadband Access Web page, July 29, 2005.
Value Part ners: “ Get t ing t he Most Out of t he Digit al Divide – Allocat ing UHF Spect rum t o
Maximise t he Benefit s for European Societ y” , March 2008,
ht t p: / / www. spect rumst rat egy. com/ Pages/ GB/ perspect ives/ Spect rum- Get t ing- t he- most - out -
of- t he- digit a- dividend- 2008. pdf.
Vodafone: press release, “ Vodafone Trials HSPA+ Mobile Broadband at Speeds of Up To
16Mbps, ” January 15, 2009.
Wikipedia: ht t p: / / en. wikipedia. org/ wiki/ World_populat ion, July 2008.
Wireless I nt elligence and AT Kearney: ht t p: / / www. cellular-
news. com/ st ory/ 31730. php?source= newslet t er, 2008.
Wireless Week: “ One in Three Handset s Will Be a Smart phone by 2013” ,
ht t p: / / www. wirelessweek. com/ art icle. aspx?id= 158452, March 2008.


This whit e paper was writ t en for 3G Amer icas by Rysavy Research ( ht t p: / / www.rysavy. com) and ut ilized a
composit e of st at ist ical infor mat ion from mult iple resources.



HSPA t o LTE- Advanced, Rysavy Research / 3G Americas, Sept ember 2009 Page 118
The contents of this paper reflect the research, analysis and conclusions of Rysavy Research and
may not necessarily represent the comprehensive opinions and individual view points of each
particular 3G Americas Board member company.

Rysavy Research provides this document and the information contained herein to you for
informational purposes only. Rysavy Research provides this information solely on the basis that
you will take responsibility for making your own assessments of the information.

Although Rysavy Research has exercised reasonable care in providing this information to you,
Rysavy Research does not warrant that the information is error-free. Rysavy Research disclaims
and in no event shall be liable for any losses or damages of any kind, whether direct, indirect,
incidental, consequential, or punitive arising out of or in any way related to the use of the
information.

















Copyright ©2009 Rysavy Research, LLC. All rights reserved.
http://www.rysavy.com



TABLE OF CONTENTS
INTRODUCTION...................................................................................................... 4  BROADBAND DEVELOPMENTS ................................................................................ 6  Wireless versus Wireline ........................................................................................ 7  Bandwidth Management ........................................................................................ 8  WIRELESS DATA MARKET ..................................................................................... 10  Trends .............................................................................................................. 10  EDGE/HSPA/HSPA+ Deployment ........................................................................... 13  Statistics ........................................................................................................... 13  WIRELESS TECHNOLOGY EVOLUTION AND MIGRATION ....................................... 14  1G to 4G ........................................................................................................... 14  3GPP Evolutionary Approach ................................................................................ 16  Spectrum .......................................................................................................... 19  Core-Network Evolution ....................................................................................... 22  Service Evolution ................................................................................................ 23  Device Innovation ............................................................................................... 24  Network Interfaces for Applications ....................................................................... 25  Mobile Application Architectures ............................................................................ 26  Broadband-Wireless Deployment Considerations ..................................................... 27  Feature and Network Roadmap ............................................................................. 28  Deployment Scenarios ......................................................................................... 32  COMPETING TECHNOLOGIES ................................................................................ 33  CDMA2000 ........................................................................................................ 33  WiMAX .............................................................................................................. 35  IEEE 802.20....................................................................................................... 39  Wi-Fi and Municipal Wi-Fi Systems ........................................................................ 39  COMPARISON OF WIRELESS TECHNOLOGIES ....................................................... 40  Data Throughput ................................................................................................ 40  HSDPA Throughput in Representative Scenarios ...................................................... 43  Release 99 and HSUPA Uplink Performance ............................................................ 45  LTE Throughput .................................................................................................. 46  Latency ............................................................................................................. 49  Spectral Efficiency .............................................................................................. 50  Cost, Volume and Market Comparison ................................................................... 58  Competitive Summary ......................................................................................... 60 

HSPA to LTE-Advanced, Rysavy Research / 3G Americas, September 2009

Page 2

CONCLUSION........................................................................................................ 62  APPENDIX: TECHNOLOGY DETAILS ...................................................................... 64  EDGE/EGPRS ..................................................................................................... 65  Evolved EDGE .................................................................................................... 67  UMTS-HSPA Technology ...................................................................................... 75  UMTS Release 99 Data Capabilities........................................................................ 77  HSDPA .............................................................................................................. 77  HSUPA .............................................................................................................. 81  Evolution of HSPA (HSPA+) .................................................................................. 82  HSPA Voice Support ............................................................................................ 90  3GPP LTE .......................................................................................................... 92  4G, IMT-Advanced and LTE-Advanced.................................................................... 97  UMTS TDD ........................................................................................................ 101  TD-SCDMA ....................................................................................................... 102  IMS ................................................................................................................. 102  Broadcast/Multicast Services ............................................................................... 103  EPC/SAE........................................................................................................... 104  White Space ..................................................................................................... 106  ACRONYMS ......................................................................................................... 108  ADDITIONAL INFORMATION .............................................................................. 112  REFERENCES ...................................................................................................... 112 

HSPA to LTE-Advanced, Rysavy Research / 3G Americas, September 2009

Page 3

Introduction
We are at the dawn of a new decade that will bring to mass market the mobile broadband innovations introduced over the last several years. 3G technology has shown us the power and potential of always-on, everyplace network connectivity and has ignited a massive wave of industry innovation that spans devices, applications, Internet integration, and new business models. Already used by hundreds of millions of people, mobile broadband connectivity is on the verge of becoming ubiquitous. It will do so on a powerful foundation of networking technologies, including GSM with EDGE, HSPA, and LTE. LTE in a forthcoming release will be one of the first technologies to meet the requirements of International Mobile Telephone (IMT) Advanced, a project of the International Telecommunications Union (ITU) that this year defined official “4G” requirements. Through constant innovation, Universal Mobile Telecommunications System (UMTS) with High Speed Packet Access (HSPA) technology has established itself as the global, mobilebroadband solution. Building on the phenomenal success of Global System for Mobile Communications (GSM), the GSM-HSPA ecosystem has become the most successful communications technology family ever. Through a process of constant improvement, the GSM family of technologies has not only matched or exceeded the capabilities of all competing approaches, but has significantly extended the life of each of its member technologies. UMTS-HSPA, in particular, has many key technical and business advantages over other mobile wireless technologies. Operators worldwide are now deploying both High Speed Downlink Packet Access (HSDPA) and High Speed Uplink Packet Access (HSUPA), the combination of the two technologies called simply HSPA. HSPA is the most capable cellulardata technology ever developed and deployed. HSPA, already widely available, follows the successful deployment of UMTS networks around the world and is now a standard feature. HSPA is strongly positioned to be the dominant mobile-data technology for the next five to ten years. To leverage operator investments in HSPA, the 3GPP (Third Generation Partnership Project) standards body has developed a series of enhancements to create “HSPA Evolution,” also referred to as “HSPA+.” HSPA Evolution represents a logical development of the Wideband Code Division Multiple Access (WCDMA) approach, and it is the stepping stone to an entirely new Third Generation Partnership Project (3GPP) radio platform called 3GPP Long Term Evolution (LTE). LTE, which uses Orthogonal Frequency Division Multiple Access (OFDMA), will be ready for deployment in 2010. Simultaneously, 3GPP — recognizing the significant worldwide investments in GSM networks—has defined enhancements that will significantly increase EDGE data capabilities through an effort called Evolved EDGE. Combined with these improvements in radio-access technology, 3GPP has also spearheaded the development of major core-network architecture enhancements such as the IP Multimedia Subsystem (IMS) and the Evolved Packet Core (EPC), previously called System Architecture Evolution or SAE, as well as developments in Fixed Mobile Convergence (FMC). These developments will facilitate new types of services, the integration of legacy and new networks, the convergence between fixed and wireless systems, and the transition from circuit-switched approaches for voice traffic to a fully packet-switched model. The result is a balanced portfolio of complementary technologies that covers both radioaccess and core networks, provides operators maximum flexibility in how they enhance their networks over time, and supports both voice and data services. This paper discusses the evolution of EDGE, HSPA enhancements, 3GPP LTE, the capabilities of these technologies, and their position relative to other primary competing technologies. It HSPA to LTE-Advanced, Rysavy Research / 3G Americas, September 2009 Page 4

explains how these technologies fit into the ITU roadmap that leads to IMT-Advanced. The following are some of the important observations and conclusions of this paper:

The wireless technology roadmap now extends to IMT-Advanced with LTE-Advanced being one of the first technologies defined to meet IMT-Advanced requirements. LTEAdvanced will be capable of peak throughput rates that exceed 1 gigabit per second (Gbps). Persistent innovation created EDGE, which was a significant advance over GPRS; HSPA and HSPA+, which are bringing UMTS to its full potential; and is now delivering LTE, the most powerful, wide-area wireless technology ever developed. GSM-HSPA1 has an overwhelming global position in terms of subscribers, deployment, and services. Its success will continue to marginalize other wide-area wireless technologies. In current deployments, HSPA users regularly experience throughput rates well in excess of 1 megabit per second (Mbps) under favorable conditions, on both downlinks and uplinks, with 4 Mbps downlink speed commonly being measured. Planned enhancements such as dual-carrier operation will double peak userachievable throughput rates. HSPA Evolution provides a strategic performance roadmap advantage for incumbent GSM-HSPA operators. Features such as dual-carrier operation, MIMO, and higherorder modulation offer operators multiple options for upgrading their networks, with many of these features (e.g., dual-carrier, higher-order modulation) being available as network software upgrades. HSPA+ with 2x2 MIMO, successive interference cancellation, and 64 Quadrature Amplitude Modulation (QAM) is more spectrally efficient than competing technologies including Worldwide Interoperability for Microwave Access (WiMAX) Release 1.0. The LTE Radio Access Network technical specification was approved in 2008 for 3GPP Release 8, which was fully ratified in March, 2009. Initial deployments will occur in 2010 and will expand rapidly thereafter. The 3GPP OFDMA approach used in LTE matches or exceeds the capabilities of any other OFDMA system. Peak theoretical downlink rates are 326 Mbps in a 20 MHz channel bandwidth. LTE assumes a full Internet Protocol (IP) network architecture, and it is designed to support voice in the packet domain. LTE has become the technology platform of choice as GSM-UMTS and CDMA/EV-DO operators are making strategic, long-term decisions on their next-generation platforms. In June of 2008, after extensive evaluation, LTE was the first and thus far only technology recognized by the Next Generation Mobile Network alliance to meet its broad requirements. GSM-HSPA will comprise the overwhelming majority of subscribers over the next five to ten years, even as new wireless technologies are adopted. The deployment of LTE and its coexistence with UMTS-HSPA will be analogous to the deployment of UMTSHSPA and its coexistence with GSM. 3GPP has made significant progress on how to enhance LTE to meet the requirements of IMT-Advanced in a project called LTE-Advanced. LTE-Advanced is expected to be the first true “4G” system available.

1 This paper’s use of the term “GSM-HSPA” includes GSM, EDGE, UMTS, HSPA and HSPA+. “UMTS-HSPA” refers to UMTS technology deployed in conjunction with HSPA capability.

HSPA to LTE-Advanced, Rysavy Research / 3G Americas, September 2009

Page 5

 

HSPA-LTE has significant economic advantages over other wireless technologies. WiMAX has developed an ecosystem supported by many companies, but it will still only represent a very small percentage of wireless subscribers over the next five years. EDGE technology has proven extremely successful and is widely deployed on GSM networks globally. Advanced capabilities with Evolved EDGE can double and eventually quadruple current EDGE throughput rates, halve latency and increase spectral efficiency. With a UMTS multi-radio network, a common core network can efficiently support GSM, WCDMA, and HSPA access networks and offer high efficiency for both high and low data rates, as well as for both high- and low-traffic density configurations. In the future, EPC/SAE will provide a new core network that supports both LTE and interoperability with legacy GSM-UMTS radio-access networks. Innovations such as EPC/SAE and UMTS one-tunnel architecture will “flatten” the network, simplifying deployment and reducing latency. Circuit-switched, voice-over HSPA, then moving to voice over Internet Protocol (VoIP) over HSPA will add to voice capacity and reduce infrastructure costs. In the meantime, UMTS-HSPA enjoys high circuit-switched voice spectral efficiency, and it can combine voice and data on the same radio channel.

This paper begins with an overview of the market, looking at trends, EDGE and UMTS-HSPA deployments, and market statistics. It then examines the evolution of wireless technology, particularly 3GPP technologies, including spectrum considerations, core-network evolution, broadband-wireless deployment considerations, and a feature and network roadmap. Next, the paper discusses other wireless technologies, including Code Division Multiple Access 2000 (CDMA2000) and WiMAX. Finally, it compares the different wireless technologies technically, based on features such as performance and spectral efficiency. The appendix explains in detail the capabilities and workings of the different technologies including EDGE, Evolved EDGE, WCDMA2, HSPA, HSPA Evolution (HSPA+), LTE, LTEAdvanced, IMS, and SAE.

Broadband Developments
As wireless technology represents an increasing portion of the global communications infrastructure, it is important to understand overall broadband trends and the role between wireless and wireline technologies, as well as Internet trends. Sometimes wireless and wireline technologies compete with each other, but in most instances, they are complementary. For the most part, backhaul transport and core infrastructure for wireless networks are based on wireline approaches, whether optical or copper. This applies as readily to Wi-Fi networks as it does to cellular networks. Trends show explosive bandwidth growth of the Internet at large and for mobile broadband networks in particular. Cisco projects global IP traffic as nearly doubling every two years

Although many use the terms “UMTS” and “WCDMA” interchangeably, in this paper we use “WCDMA” when referring to the radio interface technology used within UMTS and “UMTS” to refer to the complete system. HSPA is an enhancement to WCDMA.

2

HSPA to LTE-Advanced, Rysavy Research / 3G Americas, September 2009

Page 6

5 With declining voice revenue. even though wireline technology. Developed countries continue to show tremendous uptake of mobile broadband services. has inherent capacity advantages. is managing bandwidth. conclusively demonstrates the desire for mobile-oriented communications. 1 exabyte is 1015 bytes. 1 terabyte is 1012 bytes. 5 Source: Cisco Visual Networking Index: Global Mobile Data Traffic. wireline networks have always had greater capacity. Forecast Update. and supplying attractive devices. however. 3 4 Source: Cisco. competing with digital subscriber line (DSL) for home use. cellular operators face a tremendous opportunity to develop a mobile broadband business. It also means nurturing an application ecosystem. Mobile broadband combines compelling high-speed data services with mobility. which will require a number of different approaches. Wireless versus Wireline Wireless technology is playing a profound role in networking and communications. however. but increasing data revenue. means more than just providing high speed networks. the opportunities are limitless when considering the many diverse markets mobile broadband can successfully address. The overwhelming global success of mobile telephony.through 20123. Rysavy Research / 3G Americas. reaching 2 exabytes4 per month by 2013. 2009 HSPA to LTE-Advanced. These are all areas in which the industry has done well. in developing countries. for whom 3G can be a cost-effective option. and historically have delivered faster throughput rates. providing complementary services. January 29. An emerging challenge. such as fiber links. 2008 One gigabyte is 109 bytes. Thus. Relative to wireless networks. Successful execution. 1 zettabyte is 1018 bytes.” June 16. September 2009 Page 7 . and mobile broadband traffic growing at a CAGR of 131 percent between 2008 and 2013. and now the growing adoption of mobile data. “Approaching the Zettabyte Era. there is no doubt that 3G technology will cater to both enterprises and their high-end mobile workers and consumers. Figure 1 shows advances in typical user throughput rates with a consistent 10x advantage of wireline technologies over wireless technologies. Additionally.

Thus. “Is it possible to match these rates using wireless approaches?” The answer is yes from a purely technical perspective. may limit its appeal and usage. it simply will not be possible to deliver the hundreds of gigabytes per month that users will eventually be consuming over their broadband connections with wide-area wireless networks. Otherwise. very highspeed DSL (VDSL) or fiber—especially for services such as high-definition IP Television (IPTV)—the question becomes.S. relative to wireline technology. wherein one subscriber could essentially consume the entire capacity of a WiMAX or HSPA cell sector. but it is no from a practical point of view. This is especially true when wireless is compared to optical fiber. September 2009 Page 8 . Rysavy Research / 3G Americas. These include: HSPA to LTE-Advanced.Figure 1: Wireline and Wireless Advances 100 Mbps FTTH 100 Mbps ADSL2+ 25 Mbps 10 Mbps ADSL 3 to 5 Mbps 1 Mbps ADSL 1 Mbps ISDN 128 kbps LTE 10 Mbps HSPA+ 5 Mbps HSDPA 1 Mbps UMTS 350 kbps EDGE 100 kbps GPRS 40 kbps 100 kbps 10 kbps 2000 2005 2010 The question is whether some of the limitations of wireless technology. With wireline operators looking to provide 50 to 100 Mbps to either people’s homes or businesses via next-generation cable-modem services. Consider today’s high definition (HD) television content that demands 6 to 9 Mbps of continuous connectivity. Bandwidth Management Although it is true that most 3G systems are now offering throughputs of about 1 Mbps—which is comparable to what many users experience with a basic DSL or cablemodem service—the overall capacity of wireless systems is generally lower than it is with wireline systems.) are now deploying to people’s homes. which some operators in developed countries (such as the U. and by using relatively small cell sizes. It is only possible to achieve these rates by using large amounts of spectrum. generally more than is available for current 3G systems. operators are both deploying and considering multiple approaches for managing bandwidth.

provides the technical means to deliver on proven business models.    More spectrum. Quality of service. movie downloads to femtocell connections. Newer technologies are spectrally efficient. By prioritizing traffic. Table 1: Strengths and Weakness of Broadband Approaches Strength Mobile broadband (EDGE.. which anticipates continual performance and capacity improvements. More cell sites. The 3GPP roadmap. LTE) Constant connectivity Broadband capability across extremely wide areas Good access solution for areas lacking wireline infrastructure Capacity enhancement via FMC Excellent voice communications Wireline broadband (e. September 2009 Page 9 . large downloads can occur with lower priority. LTE and LTE-Advanced will continue to provide a competitive platform for tomorrow’s new business opportunities. meaning greater throughput in the same amount of spectrum. Operators can offer lower rates or perhaps fewer restrictions on large data transfers that occur at off-peak hours such as overnight. Rysavy Research / 3G Americas.g. HSPA. DSL. HSPA. Femto cells can significantly offload the macro network. DOCSIS. Increased spectral efficiency. Pricing plans can encourage users to move high-bandwidth activities (e. especially in developing economies lacking infrastructure Weakness Lower capacity than wireline approaches Inability to serve highbandwidth applications such as IP TV 3GPP technologies clearly address proven market needs. more    It will take a creative blend of all of the above as well as other measures to make the mobile broadband market successful and to enable it to exist as a complementary solution to wired broadband. Wi-Fi networks offer another means of offloading heavy traffic. Spectrum correlates directly to capacity. Smaller cell sizes result in more capacity per subscriber. and more spectrum is becoming available globally for mobile broadband. HSPA+. Wi-Fi. Table 1 summarizes the strengths and weaknesses of wireless versus wireline broadband approaches. As the applications for mobile broadband continue to expand.. HSPA to LTE-Advanced. thus not affecting other active users.g. Off-peak hours. FTTH) High-capacity broadband at very high data rates Evolution to extremely high throughput rates Expensive to deploy new networks. Femtocells. hence their overwhelming success.

March 2009 9 Source: 3G Americas press release of June 5. of which 1. Source: Rysavy Research. Clearly. ”We are at a unique and pivotal time in history.html Source: Subscriber Data . and then provides market data that demonstrates the rapid growth of wireless data. wireless data worldwide now comprises a significant percentage of revenue per user (ARPU). In the United States. more than 3.Wireless Data Market By May 2009. consumer awareness and comfort with emerging wireless technology. US Wireless Data Market Update . September 2009 Page 10 . and is projected to hit 30% by the end of 2009. wireless data is now more than 26 percent of ARPU. World Cellular Information Service.”9 This growth continues. “This level of wireless technology growth exceeds that of almost all other lifestyle-changing innovations.gov/ipc/www/idb/worldpopinfo.8In 2007. Trends As stated in a Rysavy Research report for CTIA on mobile broadband spectrum demand. Source: US Census Bureau. 2007.7 By the end of 2013.6 billion will use 3GPP technologies.census. and industry innovation are converging to create mass-market acceptance of mobile broadband. 6 7 8 Source: Informa Telecoms & Media.”11 The market factors contributing to the surging growth in this market are shown in the following figure.10 This section examines trends and deployment. in which technology capability. the global 3G wireless market is expected to include more than 2 billion subscribers. 3G Americas President Chris Pearson stated. Chetan Sharma. http://www. May 2009. 10 11 HSPA to LTE-Advanced.Q1 2009. Although voice still constitutes most cellular traffic.Informa Telecoms and Media.” December 2008. Rysavy Research / 3G Americas.8 billion population. GSM-HSPA has established global dominance. “Mobile Broadband Spectrum Demand.7 billion subscribers were using GSM-HSPA6—approaching an astonishing 50 percent of the world’s total 6. representing 80% market share.

12 Based on one leading UMTS-HSPA infrastructure vendor’s statistics. This adoption will offer tremendous opportunities and the associated risks to operators as they choose the most commercially viable evolutionary path for migrating their customers. Operators that are the most aggressive with mobile broadband services are experiencing data growth rates even higher than these average values.com/pr/library/2009/06/22iphone. Figure 3 compares the rapid growth in wireless data traffic compared to voice traffic across multiple operators. it is essential that operators deploy spectrally efficient data technologies that meet customer requirements for performance—especially because data applications can demand significantly more network resources than traditional voice services.Figure 2: Market Factors Contributing to Growth of Mobile Broadband As data constitutes a rising percentage of total cellular traffic. there are now more than 50. Rysavy Research / 3G Americas.html . in HSPA coverage areas worldwide. By mid 2009. As a consequence. For example. The EDGE/HSPA/LTE evolutionary paths provide data capabilities that address market needs and deliver ever-higher data throughputs. Operators have a huge investment in spectrum and in their networks. lower latency. Page 11 HSPA to LTE-Advanced.apple. Because of its growing size—and its unassailable potential—application and content developers are making the wireless market a high priority. September 2009 . data services must leverage these investments. this rich network and device environment is spawning the availability of a wide range of wireless applications and content. Traffic has continued to increase since. and increased spectral efficiency. the volume of data traffic significantly exceeded voice traffic.000 applications for the Apple iPhone. It is only a matter of time before today’s more than 4 billion cellular customers start taking full advantage of data capabilities. 12 Source: http://www.

data demands are expected to grow significantly. Source: “Managing Growth and Profits in the Yottabyte Era”. September 2009 Page 12 . HSPA to LTE-Advanced. July 2009. One Terabyte is 1000 gigabytes. Rysavy Research / 3G Americas.Figure 3: UMTS-HSPA Voice and Data Traffic13 19 17 15 13 11 9 7 5 3 1 35 05 75 45 5 85 55 25 95 65 Relative Network Load ~ 18x Packet data Voice ~ 2x Jan Mar May Jul Sep Nov Jan Mar May July Sep Nov Jan Mar May 07 07 07 07 07 07 08 08 08 08 08 08 09 09 09 Over time. Chetan Sharma. Figure 4 shows a projection by Chetan Sharma of mobile data growth in the US through 2014. Figure 4: Mobile Data Growth in the United States14 13 14 Based on leading UMTS-HSPA infrastructure vendor statistics.

Rysavy Research / 3G Americas. Statistics A variety of statistics show the rapid growth in wireless data. the first time the US market has crossed the $10B milestone. laptops.Q1 2009. HSPA makes such efficient use of spectrum for data that it results in a much lower overall cost per megabyte of data delivered. Source: Berg Insight.16 Almost all UMTS operators are deploying HSPA for two reasons: first. 11.20 15 16 17 18 19 20 Source: “World Cellular Information Service. Not only does it provide a platform for continual improvements in capabilities. Starhub (Singapore). representing more than 345 networks in approximately 170 countries.18 Berg Insight reported that in the European Union. http://www. Operators have begun deploying evolved HSPA features and HSPA+ launches include: Telstra (Australia). Source: GSMA.15 Because of the very low incremental cost of including EDGE capability in GSM network deployments.” Informa Telecoms & Media. routers. media players and cameras.com/News. but it does so over huge coverage areas and on a global basis. data cards. Chetan Sharma reported that in Q1 2009.aspx?m_m=6&s_m=1. HSPA to LTE-Advanced. As the technology matures.The key for operators is enhancing their networks to support the demands of consumer and business applications as they grow. Two hundred and sixty four operators in 114 countries offer HSDPA.berginsight. Ibid. the US wireless data market grew 32% over Q1 of 2008 to reach $10B in mobile data revenues. He also states that 62% of US subscribers were using some form of data service. Nearly all WCDMA handsets are also GSM handsets. This is where the GSM family of wireless-data technologies is the undisputed leader. Source: Pyramid Research. EDGE/HSPA/HSPA+ Deployment Three quarters of GSM networks today support EDGE.17 Devices include handsets. there are more than 1375 commercial HSPA devices available worldwide from 135 suppliers. virtually all new GSM infrastructure deployments are also EDGE-capable and nearly all new mid.19 Pyramid Research projects the number of European mobile broadband users to reach 117 million in 2014. Already. UMTS has established itself globally. June 2009. “Europe to See Huge Growth in Mobile Broadband Services despite Recession. CSL Limited (Hong Kong). US Wireless Data Market Update .” 2009. upgrading to HSPA+ will likely represent a minimal investment for operators in order to significantly boost network performance. Meanwhile. modems. up from 24 million in 2008.6% of broadband links at the end of 2008 were based on HSPA in both mobile and home environments. along with offering complementary capabilities such as IP-based multimedia. so WCDMA users can access the wide base of GSM networks and services. and 77 of these have HSUPA deployed. There are more than 336 million UMTS-HSPA customers worldwide spanning 283 commercial networks. Source: Chetan Sharma. the incremental cost of HSPA is relatively low and second.to high-level GSM devices include EDGE radio technology. September 2009 Page 13 . Mobilkom (Austria).

1G to 4G There is some confusion in the industry as to what technology falls into which cellular generation. Pyramid Research expects LTE networks to grow more quickly than prior 3G networks. 3G requirements were specified by the ITU as part of the International Mobile Telephone 2000 (IMT-2000) project. 2G technologies became available in the 1990s. for which digital networks had to provide 144 kbps of throughput at mobile speeds. CDMA2000 1xRTT and GSM are the primary 2G technologies. June 2009. 384 kbps at pedestrian speeds. 22 23 HSPA to LTE-Advanced. July 2009. also meets this requirement. followed by today’s enhanced 3G capabilities such as HSPA.21 According to Juniper Research.org). EDGE. the evolution and migration of wireless-data technologies from EDGE to LTE.22 According to 3G Americas (www. Informa WCIS projected in June 2009 the following sales growth rate for WCDMA handsets: 23 2009: 307 million 2010: 416 million 2011: 564 million 2012: 736 million 2013: 927 million It is clear that both EDGE and UMTS/HSDPA are dominant wireless technologies. although CDMA2000 1xRTT is sometimes called a 3G technology because it meets the 144 kbps mobile throughput requirement. Source: Juniper Research. 1G refers to analog cellular technologies and became available in the 1980s. LTE Report. first with EDGE. From a device perspective. however. And powerful data capabilities and global presence mean these technologies will likely continue to capture most of the available wireless-data market. Meanwhile. and 2 Mbps in indoor 21 Source: Global Telecom Insider report. "LTE’s Five-Year Global Forecast: Poised to Grow Faster than 3G. HSPA+ and now. 2G denotes initial digital systems. Rysavy Research / 3G Americas. September 2009 Page 14 . as well as the evolution of underlying wireless approaches.3gamericas. there are already in excess of 30 network operator commitments to LTE. LTE. LTE should see relatively rapid adoption as it becomes deployed starting in 2010. which is the basis of LTE. underlying approaches have evolved from Time Division Multiple Access (TDMA) to CDMA. and then UMTS. Progress in 3GPP has occurred in multiple phases.Though most mobile broadband growth today is based on HSPA (with some EV-DO). Source: “World Cellular Information Service." 2009. introducing services such as short messaging and lower speed data.” Informa Telecoms & Media. reaching 100 million subscribers in just four years from initial 2010 deployments. and now from CDMA to OFDMA. Wireless Technology Evolution and Migration This section discusses 1G to 4G designations. which itself is evolving to LTE-Advanced. there are more than 100 operators that have committed or expressed intentions to commit to LTE.

5G and LTE 3. but this is only accurate to the extent that such designation refers to the general approach or platform that will be enhanced to meet the 4G requirements. The ITU recommends operation in up to 100 MHz radio channels and peak spectral efficiency of 15 bps/Hz. Digital Technology. It will require new technologies such as LTE-Advanced (with work already underway) and IEEE 802. No technology meets these requirements yet.environments. Analog technology.16m. 384 kbps pedestrian. No official requirements. calling these technologies 3G clearly does not give them full credit as they are a generation beyond current technologies in capability. Deployed in the 1990s. No technology meets requirements today. 2 Mbps indoors Primary technologies include CDMA2000 EV-DO and UMTS-HSPA. UMTS-HSPA and CDMA2000 EV-DO are the primary 3G technologies. Table 2: 1G to 4G Generation 1G 2G Requirements No official requirements.5 Gbps. The following table summarizes the generations. Requirements include operation in up to 40 MHz radio channels and extremely high spectral efficiency. Rysavy Research / 3G Americas. IEEE 802. Primary technologies include CDMA2000 1xRTT and GSM. Unfortunately. WiMAX now an official 3G technology. First digital systems. New services such as SMS and low-rate data. But calling them 4G is not correct. 4G ITU’s IMT-Advanced requirements include ability to operate in up to 40 MHz radio channels and with very high spectral efficiency. 1 Gbps was frequently cited as a 4G goal. none is even close.16m and LTE Advanced being designed to meet requirements. The ITU has recently issued requirements for IMT-Advanced. Comments Deployed in the 1980s. Some have tried to label current versions of WiMAX and LTE as “4G”. Previous to the publication of the requirements. resulting in a theoretical throughput rate of 1. September 2009 Page 15 . although these are not official designations. which constitutes the official definition of 4G. although recently WiMAX was also designated as an official 3G technology.9G. Some people have even called technologies such as HSPA 3. the generational labels do not properly capture the scope of available technologies and have resulted in some amount of market confusion. With WiMAX and HSPA significantly outperforming 3G requirements. 3G ITU’s IMT-2000 required 144 kbps mobile. HSPA to LTE-Advanced.

” which will be available for deployment in the 2009 to 2010 timeframe. CDMA was chosen as the basis of 3G technologies including WCDMA for the frequency division duplex (FDD) mode of UMTS and Time Division CDMA (TD-CDMA) for the time division duplex (TDD) mode of UMTS. HSPA to LTE-Advanced. In the same way that 3G coexists with Second Generation (2G) systems in integrated networks. Meanwhile. By the end of the decade. To address ITU’s IMT-Advanced requirements. because of sheer market momentum. LTE is of crucial importance to operators since it provides the efficiencies and capabilities being demanded by the quickly growing mobile broadband market. September 2009 Page 16 . such as HSPA and HSPA+. this section provides a summary 24 3GPP also refers to LTE as Enhanced UMTS Terrestrial Radio Access Network (E-UTRAN). See the appendix section “4G. EPC/SAE provides a new core architecture that enables both flatter architectures and integration of LTE with both legacy GSM-HSPA networks. Standards bodies have already defined “Evolved EDGE. Innovations such as dual-carrier HSPA. explained in detail in the appendix section “Evolution of HSPA (HSPA+). there are nevertheless opportunities for additional optimizations and enhancements. especially in 5 MHz spectrum allocations.. 3GPP’s evolutionary plan is to recognize the strengths and weaknesses of every technology and to exploit the unique capabilities of each one accordingly. however. The TDD mode. cost per Mbyte) is almost directly proportional to the spectral efficiency of the technologies.” coordinate the operation of HSPA on two adjacent 5 MHz carriers for higher throughput rates.g. Evolved EDGE more than doubles throughput over current EDGE systems. halves latency. and increases spectral efficiency. Multimode devices will function across LTE/3G or even LTE/3G/2G. as well as other wireless technologies. will be important in enabling deployments where paired spectrum is unavailable. In combination with MIMO. depending on market circumstances. Rysavy Research / 3G Americas. Given some of the advantages of an OFDM approach. LTE incorporates best-of-breed radio techniques to achieve performance levels beyond what will be practical with CDMA approaches. The evolved data systems for UMTS. Beyond radio technology. The combination of EPC and LTE is referred to as the Evolved Packet System (EPS). IMT-Advanced and LTE-Advanced” for a detailed explanation. the majority of worldwide subscribers will still be using GSM/EDGE technologies.3GPP Evolutionary Approach Rather than emphasizing any one wireless approach. LTE has the highest spectral efficiency of any specified technology. LTE systems will coexist with both 3G systems and 2G systems. 3GPP is developing LTE-Advanced. is mature and broadly deployed. Already extremely efficient. particularly in larger channel bandwidths. introduce enhancements and simplifications that help CDMA-based systems match the capabilities of competing systems. Although later sections quantify performance and the appendix of the white paper presents functional details of the different technologies. based on a TDMA approach. Many deployments will be based on FDD in paired spectrum. The cost for operators to deliver data (e. a technology that will have peak rates of more than 1 Gbps. dual-carrier HSPA will achieve peak network speeds of 84 Mbps. LTE is available in both FDD and TDD modes. 3GPP has specified OFDMA as the basis of its Long Term Evolution24 effort. GSM. making it an essential technology as the market matures.

Table 3 summarizes the key 3GPP technologies and their characteristics.5 Mbps to 7 Mbps 1 Mbps to 4 Mbps LTE OFDMA 4 Mbps to 24 Mbps (in 2 x 20 MHz) LTEAdvanced OFDMA HSPA to LTE-Advanced. All communications handled in IP domain. 200 to 300 kbps HSPA CDMA 1 Mbps to 4 Mbps 500 kbps to 2 Mbps HSPA+ CDMA 1. Evolution of HSPA in various stages to increase throughput and capacity and to lower latency. halve latency and increase spectral efficiency. Table 3: Characteristics of 3GPP Technologies Technology Name GSM Type Characteristics Typical Downlink Speed Typical Uplink Speed TDMA Most widely deployed cellular technology in the world. Rysavy Research / 3G Americas. Advanced version of EDGE that can double and eventually quadruple throughput rates. New radio interface that can use wide radio channels and deliver extremely high throughput rates. An enhancement to original UMTS data service. Current deployments implement HSPA for data service. 70 kbps to 135 kbps 70 kbps to 135 kbps EDGE TDMA Evolved EDGE TDMA 175 kbps to 350 kbps expected (Single Carrier) 350 kbps to 700 kbps expected (Dual Carrier) 200 to 300 kbps 150 kbps to 300 kbps expected UMTS CDMA 3G technology providing voice and data capabilities. An enhancement to original GSM data service called GPRS. Provides voice and data service via GPRS/EDGE.intended to provide a frame of reference for the subsequent discussion. September 2009 Page 17 . Data service for GSM networks. Advanced version of LTE designed to meet IMTAdvanced requirements. Data service for UMTS networks.

Rysavy Research / 3G Americas.1 Mbps UL: 1. Radio channel bandwidths indicated. Majority of deployments today are based on Release 99.16m Notes: Throughput rates are peak theoretical network rates. HSPA to LTE-Advanced.  25 After Release 99. It is important to realize that the 3GPP releases address multiple technologies. September 2009 Page 18 . Figure 5: Evolution of TDMA. For example.89 Mbps UL: 947 kbps 2011 2012 2013 HSPA DL: 14.0 DL: 46 Mbps UL: 4 Mbps 10 MHz 3:1 TDD Rel 1.4 Mbps UL: 5. Dates refer to expected initial commercial network deployment except 2008. and OFDMA Systems 2008 EDGE DL: 474 kbps UL: 474 kbps 2009 2010 Evolved EDGE DL: 1.User achievable rates and greater details on typical rates are covered in Table 6 in the section “Data Throughput” later in this paper. Provides support for GSM/EDGE/GPRS/WCDMA radio-access networks. The development of GSM and UMTS-HSPA happens in stages referred to as 3GPP releases.5 Mbps In 5 MHz Rel 9 HSPA+ DL: 84 Mbps UL: 23 Mbps In 10 MHz LTE DL: 326 Mbps UL: 86 Mbps In 20 MHz LTE (Rel 9) LTE Advanced DL: > 1 Gbps EV-DO Rev A DL: 3. CDMA.9 Mbps In 5 MHz Fixed WiMAX Release 1. First steps toward using IP transport in the core network. First deployable version of UMTS. and equipment vendors produce hardware that supports particular versions of each specification. but also significantly enhanced GSM data functionality with Evolved EDGE.25 MHz EV-DO Rev B DL: 14. Enhancements to GSM data (EDGE).76 Mbps In 5 MHz Rel 7 HSPA+ DL: 28 Mbps UL: 11. release versions went to a numerical designation instead of designation by year. Figure 5 shows the evolution of the different wireless technologies and their peak network performance capabilities. Multimedia messaging support.7 Mbps UL: 4. Release 7 optimized VoIP for HSPA. Release 4: Completed. A summary of the different 3GPP releases is as follows: 25  Release 99: Completed.8 Mbps In 1. which shows available technologies that year.5 Mbps In 5 MHz Rel 8 HSPA+ DL: 42 Mbps UL: 11.5 IEEE 802.

Performance specifications for advanced receivers. which includes higher order modulation and MIMO. September 2009 Page 19 . Results include performance enhancements. Release 5: Completed. Release 9: Under development. Initial VoIP capability. As other frequency bands become available for deployment. For this reason. Full ability to use IP-based transport instead of just Asynchronous Transfer Mode (ATM) in the core network. Also includes optimization of MBMS capabilities through the multicast/broadcast. An increasing number of operators are also deploying UMTS at 900 MHz. Provides fine-tuning and incremental improvements of features from previous releases. the Advanced Wireless Services (AWS) band at 1710-1755 MHz with/2110-2155 MHz in the US. Specifies OFDMA-based 3GPP LTE. HSDPA. the detailed discussion of the technologies in this paper emphasizes features as opposed to 3GPP releases. standards bodies are adapting UMTS for these bands as well. as well as reductions in call set-up delay for PoC. HSPA to LTE-Advanced. HSUPA. UMTS-TDD equipment is already available for 450 MHz. Release 10: Under development. Rysavy Research / 3G Americas. and the forthcoming 2. most users of the technology are more interested in particular features and capabilities such as whether a device supports HSDPA. IMS enhancements. Will include HSPA and LTE enhancements including HSPA multi-carrier operation. increased capacity. the Federal Communications Commission auctioned the 700 MHz band in the United States in January 2008. a traditional GSM band. Radio enhancements to HSPA include 64 QAM in the downlink DL and 16 QAM in the uplink. single-frequency network (MBSFN) function. Release 6: Completed. Likely 2011. Defines EPC. only the frequency-dependent portions of the radios have to change. Comprises further HSPA Evolution features such as simultaneous use of MIMO and 64 QAM. Provides enhanced GSM data functionality with Evolved EDGE. Spectrum Another important aspect of UMTS-HSPA deployment is the expanding number of available radio bands. The availability of this band. as shown in Figure 6. Enhanced multimedia support through Multimedia Broadcast/Multicast Services (MBMS).7/2. Continuous Packet Connectivity (CPC) enables efficient “always-on” service and enhanced uplink UL VoIP capacity. WLAN integration option.1GHz band. The 1710-1770 uplink was matched with 2110-2170 downlink to allow for additional global harmonization of the 1. Will specify LTE-Advanced that meets the requirements set by ITU’s IMT-Advanced project. improved spectral efficiency. and the corresponding support from infrastructure and mobile-equipment vendors.6 GHz frequency band in Europe are providing operators with wider deployment options. The fundamental system design and networking protocols remain the same for each band. Meanwhile. and better resistance to interference. This includes 450 and 700 MHz. First phase of IMS. Specifies HSPA Evolution (HSPA+). Includes work item for dual-carrier HSPA (DC-HSPA) wherein two WCDMA radio channels can be combined for a doubling of throughput performance. Likely 2010. Release 8: Completed. Release 7: Completed.      Whereas operators and vendors actively involved in the development of wireless technology are heavily focused on 3GPP release versions.

supporting more subscribers and making many new types of applications feasible. the first devices operating on this band were compliant with the release 5 of the standard. according to release anterior to the introduction of that particular frequency band.1 GHz 1900 MHz 1800 MHz 1.As the total amount of available spectrum increases and as technologies simultaneously become spectrally more efficient. September 2009 Page 20 .7/2. Figure 6: FDD Bands for 3GPP Technologies 26 Operating band Band 1 Band 2 Band 3 Band 4 Band 5 Band 6 Band 7 Band 8 Band 9 Band 10 Band 11 Band 12 Band 13 Band 14 Band name 2. although band 5 (US Cellular Band) was introduced in Release 6. Figure 7 shows TDD bands defined for 3GPP Technologies.1452.9 2110-2170 1475. 26 Source: A 3G Americas’ member company. the 3GPP standard also specifies ways to implement devices and infrastructure operating on any frequency band.9 1710-1770 1427.9 728-746 746-756 758-768 It should be noted that although the support of a new frequency band may be introduced in a particular release. Rysavy Research / 3G Americas. The following figure shows the FDD bands defined for 3GPP technologies.1 GHz 850 MHz 800 MHz 2.6 GHz 900 MHz 1700 MHz Ext 1.9-1784.1500.9-1879.9 . total capacity rises rapidly.9 .9 698-716 777-787 788-798 Downlink [MHz] 2110-2170 1930-1990 1805-1880 2110-2155 869-894 875-885 2620-2690 925-960 1844. For example. public safety/private Total spectrum 2x60 MHz 2x60 MHz 2x75 MHz 2x45 MHz 2x25 MHz 2x10 MHz 2x70 MHz 2x35 MHz 2x35 MHz 2x60 MHz 2x25 MHz 2x18 MHz 2x10 MHz 2x10 MHz Uplink [MHz] 1920-1980 1850-1910 1710-1785 1710-1755 824-849 830-840 2500-2570 880-915 1749.7/2. HSPA to LTE-Advanced.1MHz 1500 MHz Lower 700 MHz Upper 700 MHz Upper 700 MHz.

Utilized efficiently without causing interference to existing spectrum holders. Emerging technologies such as LTE benefit from wider radio channels. Given the expanding size and economic significance of the mobilecomputing industry. the amount of spectrum required by the next generation of wireless technology (that is. this spectrum would fall below 5 GHz. but offer greater capacity. and it may be well into the next decade before any such new spectrum becomes available. and 3G bands. however.e. Ideally. it is important that such spectrum be: 1. 10 MHz. Personal Communications Service (PCS). after 3GPP LTE in projects such as International Mobile Telecommunications (IMT) Advanced) could be substantial given the desire to operate radio channels as wide as 100 MHz. an essential attribute because typical broadband usage contributes to a much higher load than a 27 Source: A 3G Americas’ member company. HSPA to LTE-Advanced. With the projected increase in the use of mobile-broadband technologies. 4. Made available in as wide radio channels as possible (i. Harmonized on a regional or global basis. As regulators make more spectrum available. 2. 20 MHz and more). whereas in Europe there are greater restrictions—although efforts are underway that are resulting in greater flexibility including the use of 3G technologies in current 2G bands.. 3. operators in the United States can use either 2G or 3G technologies in cellular. decisions made on new spectrum—especially with respect to global harmonization—will have profound consequences.Figure 7: TDD Bands for 3GPP Technologies27 Operating band Band 33 Band 34 Band 35 Band 36 Band 37 Band 38 Band 39 Band 40 Total spectrum 20 MHz 15 MHz 60 MHz 60 MHz 20 MHz 50 MHz 40 MHz 100 MHz Frequencies [MHz] 1900-1920 2010-2025 1850-1910 1930-1990 1910-1930 2570-2620 1880-1920 2300-2400 Different countries have regulated spectrum more loosely than others. These wider radio channels are not only spectrally more efficient. For example. Unencumbered by spectrum caps and other legacy voice-centric spectrum policies. Rysavy Research / 3G Americas. September 2009 Page 21 . This search for new spectrum is a long-term undertaking.

Core-Network Evolution 3GPP is defining a series of enhancements to the core network to improve network performance and the range of services provided. in particular. Rysavy Research / 3G Americas. initial US deployments will be at 700 MHz. For instance. Figure 8 shows increasing LTE spectral efficiency obtained with wider radio channels.6 GHz. an option called one-tunnel architecture allows operators to configure their networks so that user data bypasses a serving node and travels directly via a gateway 28 Source: 3G Americas’ member company analysis. One way to improve core-network performance is by using flatter architectures. HSPA to LTE-Advanced. watching a YouTube video consumes 100 times as many bits per second on the downlink as a voice call. however. For instance. with some many varying spectrum bands. The more hierarchical a network. the more easily it can be managed centrally. and to enable a shift to all-IP architectures. is reduced performance. especially for data communications. the tradeoff. September 2009 Page 22 .voice user.4 3 5 MHz 10 20 Of some concern in this regard is that spectrum for LTE is becoming available in different frequency bands in different countries. in Japan at 1500 MHz and in Europe at 2. In Release 7. to reduce latency (delays). To improve data performance and. Figure 8: LTE Spectral Efficiency as Function of Radio Channel Size28 100 90 % Efficiency  Relative to 20 MHz 80 70 60 50 40 30 20 10 0 1. with 20 MHz showing the most efficient configuration. it will most likely necessitate that roaming operation be based on GSM or HSPA on common regional or global bands. 3GPP has defined a number of enhancements in Release 7 and Release 8 that reduce the number of processing nodes and result in a flatter architecture. because packets must traverse and be processed by multiple nodes in the network. Thus.

UMA/GAN is a 3GPP technology. including Generic Access Network (GAN). the Third Generation HSPA to LTE-Advanced. For Release 8. femtocells leverage a subscriber's existing wirelinebroadband connection (for example. because it supports not only FMC. femtocells. like Wi-Fi. Users will also benefit from single voice mailboxes and single phone numbers. as well as the ability to control how and with whom they communicate. There are various approaches for FMC. including T-Mobile in the United States.   This paper provides further details in the sections on HSPA Evolution (HSPA+) and EPC/SAE. FMC allows the consolidation of core services across multiple-access networks. or 3G. Though FMC is still in its early stages of deployment by operators. and the appendix provides greater detail on each of these items. GSM-HSPA devices can connect via Wi-Fi or cellular connections for both voice and data. For operators. previously called System Architecture Evolution. It allows access to core services and applications via multiple-access networks. The key features and capabilities of EPC/SAE include:   Reduced latency and higher data performance through a flatter architecture. September 2009 Page 23 . Optimization for all services provided via IP. Although defined by 3GPP. Instead of operating on unlicensed bands. AT&T has committed to an IMS approach and has already deployed an IMS-based video sharing service. For instance. mobile-communications device a user has can now operate using the femtocells. an operator could offer complete VoIP-based voice service that supports access via DSL. Wi-Fi. The key advantage of the femtocell approach is that any single-mode. For users.node. Key service advances include Fixed Mobile Convergence (FMC). it promises to provide significant benefits to both users and operators. Support for both LTE radio-access networks and interworking with GSM-HSPA radio-access networks. femtocells use the operator’s licensed bands at very low power levels. IMS is another key technology for convergence. FMC refers to the integration of fixed services (such as telephony provided by wireline or Wi-Fi) with mobile cellular-based services. FMC will simplify how they communicate making it possible for them to use one device (for example. IMS is more powerful than GAN. This section provides an overview of these topics. FMC also offloads the macro network from data-intensive applications such as movie downloads. IMS. In the United States. The ability to integrate non-3GPP networks such as WiMAX. and IMS. With GAN. Service Evolution Not only do 3GPP technologies provide continual improvements in capacity and data performance. but also a much broader range of potential applications. There is also an option to integrate the functionality of the radio-network controller directly into the base station. a cell phone) at work and at home where it might connect via a Wi-Fi network or a femto cell. they also evolve capabilities that expand the services available to subscribers. and broadcasting technologies. formerly known as Unlicensed Mobile Access (UMA). When mobile. DSL). users connect via a cellular network. and it has been deployed by a number of operators. Rysavy Research / 3G Americas. These are tiny base stations that cost little more than a WiFi access point and. called the Evolved Packet Core. 3GPP has defined an entirely new core network. An alternative to using Wi-Fi for the “fixed” portion of FMC is femtocells.

and leading notebook vendors are now offering computers with integrated 3G (e.” with some form of data capability. IMS allows the creative blending of different types of communications and information. WiMAX. and it allows people to communicate in entirely new ways by dynamically using multiple services. and notebook computers and smartphones are now prevalent. a user could launch a voice call. Rysavy Research / 3G Americas. e-mail clients. video. including voice.. Modems are available in multiple formats including USB devices. According to a recent report by Forward Concepts. Computer manufacturers are also delivering new form factors such as netbooks. 29 Source: Forward Concepts. the global MID-only market is expected to grow from 305.000 shipments in 2008 to 40 million in 2012. All these capabilities consume data. HSPA) capabilities. Another important new service is support for mobile TV through what is called multicast or broadcast functions. IM.g. mobile Internet devices (MID) and smartbooks. IM clients. still cameras. during an interactive chat session. Enriched calls that include multimedia content (e. CableLabs and WiMAX have adopted IMS. A new initiative called Rich Communications Suite (RCS). a user could suddenly establish a simultaneous video connection or start transferring files. all mobile phones are becoming “smart. Or during a voice call. use cases. HSPA and LTE networks. IMS is how VoIP will (or could) be deployed in CDMA 2000 EV-DO. HSPA to LTE-Advanced.. builds upon IMS technology to provide a consistent feature set. June 2008. Mobile Internet Device and Chip Market Opportunities. graphics viewers. In fact. downloadable executable content capabilities. as well as implementation guidelines. September 2009 Page 24 . MP3 players. For example. Device Innovation Computing itself is becoming more mobile. RCS uses existing standards and specifications from 3GPP. and reference implementations. It provides application developers the means to create applications that have never before been possible. and documents. location. supported by many operators and vendors. The movement to open networks also allows a greater number of companies to develop products that use wireless networks in both vertical-market and horizontal-market scenarios.g. Core features include:    An enhanced phone book (device and/or network based) that includes service capabilities and presence-enhanced contact information. Enhanced messaging (supporting text. instant messaging and multimedia) with chat and messaging history. OMA and GSMA. a user could decide to speak to a customer-service representative. IMS will be a key platform for all-IP architectures for both HSPA and LTE. movie cameras. 3GPP has defined multicast/broadcast capabilities for both HSPA and LTE.Partnership Project 2 (3GPP2). video sharing) during voice calls. presence information. and ever more powerful browsers. While browsing the Web. Push-to-Talk over Cellular (PoC). PC Cards and Express cards.29 Cellular telephones are becoming more powerful and feature large color touch displays.

Rysavy Research / 3G Americas. HSPA to LTE-Advanced. Two widely deployed capabilities today include location queries and short message service. Parlay X Web Services include support for location and SMS. Devices principal analyst. location of friends for social networking. 30 Source: Ovum Comment. and worker dispatch. Network Interfaces for Applications Another important development related to service evolution is operators making interfaces available to external applications for information and control.30 The success of the iPhone demonstrates the potential of this market.g. The Open Mobile Alliance (OMA) now manages the Parlay X specifications. a set of functions specified through a joint project of the Parlay Group.. With SMS.” Smartphones. external applications can send user requested content such as flight updates. Until now.” June 2009. “Smartphones: the silver lining of the declining handset market. This can significantly enhance many applications including navigation.g.Meanwhile. as well as many other functions with which developers will be able to build innovative applications. Adam Leach. are now available at much lower price points and thus affordable to a much larger market segment. the European Telecommunications Standards Institute (ETSI) and 3GPP. there are now interfaces that span multiple functions using a consistent set of programming methods. applications operating on computers outside of the network) can query the location of a user. smartphones. Ovum predicts that smartphones will constitute 29% of phones by 2014. represent the convergence of the personal digital assistant.. subject to privacy restrictions. the interfaces for such functions have either been proprietary. September 2009 Page 25 . However. mobile devices or external applications (e. originally targeted for the high end of the market. One set is the Parlay X Web Services. all in a device that is only slightly larger than the average cellular telephone. With location. supplying location of nearby destinations (e. which emphasize a rich computing environment on a phone. or specific to that function. Many users would prefer to carry one device that “does it all. stores). restaurants. a fully capable mobile computer. and a phone.

parlay. The advantage of this approach is that developers can build applications that are compatible with multiple operator networks.Table 4 summarizes the available Parlay X specifications.org/en/specifications/pxws. HSPA to LTE-Advanced. 31 See http://www.31 Operators are beginning to selectively deploy these functions. September 2009 Page 26 . Rysavy Research / 3G Americas.asp for actual specifications.

will be developed specifically for mobile devices. a GSM Association project to also define network interfaces. Mobile Application Architectures Many applications used over wireless connections will be the same as those used over the Internet with desktop/laptop PCs. members. these are mainstream interfaces that will open wireless networks to thousands of Internet programmers who will be able to build applications that leverage the latent information and capabilities of wireless networks.Table 4: Parlay X Specifications Part 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 Title Common Third Party Call Call Notification Short Messaging Multimedia Messaging Payment Account Management Terminal Status Terminal Location Call Handling Audio Call Multimedia Conference Address List Management Presence Message Broadcast Geocoding Application-driven QoS Devices Capabilities and Configuration Multimedia Streaming Control Multimedia Multicast Session Management Functions Definitions common across Parlay X specifications Creates and manages calls Management of calls initiated by a subscriber Send and receive of SMS including delivery receipts Send and receive of multimedia messages Pre-paid and post-paid payments and payment reservations Management of accounts of prepaid customers Obtain status such as reachable.com/portal/tws_gsma/Resources for more information about OneAPI. 32 See http://oneapi. An increasing number of applications. 32 It also defines a REST (Representational State Transfer) interface for most functions as an alternative to using the Web service. however. RESTful interfaces are simpler for developers to work with and experiment with than Web services. unreachable or busy Obtain location of terminal Control by application for call handling of specific numbers Control for media to be added/dropped during call Create multimedia conferences including dynamic management of participants Manage subscriber groups Provide presence information Send messages to all users in specified area Obtain location address of subscriber Control quality of service of end-user connection Obtain device capability information and be able to push device configuration to device Control multimedia streaming to device Control multicast sessions. September 2009 Page 27 . This can be a challenge for developers. Rysavy Research / 3G Americas.aepona. Regardless of whether operators deploy with Parlay X or OneAPI. but that prioritizes implementation based on expected market demand. HSPA to LTE-Advanced. multimedia stream and obtain channel presence information A related project is GSMA OneAPI. OneAPI defines a simplified Web service for most functions that is essentially a subset of the related Parlay X Web service.

offline operation. Symbian and Windows Mobile. an increasing number of applications can be Web hosted.0. Spectrum has always been a major consideration for deploying any wireless network. Unlike the desktop market. as well as more powerful. These are software infrastructures that consist of a client component that operates on the mobile device. Examples include Mobile Service Architecture (MSA) for predictable capability and Mobile Information Device Profile 3 for multi-tasking. These include:  Mobile Middleware. and network topology. but it is particularly important when looking at high-performance broadband systems. Java Developments. Mobile Web 2. This result is best achieved by employing looser reuse. LiMo. enabled by Mobile Web 2. video capabilities. and which then operates on multiple device types. Palm Pre. Fortunately.0. Mobile Web 2. These factors include the amount of spectrum available. Rysavy Research / 3G Americas. September 2009 Page 28 . there are new capabilities that will result in more consistent. It was challenging enough for GSM operators to obtain UMTS spectrum. device execution environments. Vendors provide tools with which developers can develop an application in a platform-neutral manner. To achieve higher data rates requires wider radio channels. making the applications available from diverse platforms. Each of the device platforms comes with its own application development environment. however. HSPA and HSPA+ can deliver high throughput rates on the downlink and uplink with low latency in 5 MHz channels when deployed in single frequency (1/1) reuse. an OFDMA approach in a 5 MHz radio channel yields only a small performance advantage.because there are a number of different mobile platforms now available including Android. Others. backhaul. will play as important a role for mobile systems as for desktops. HSPA to LTE-Advanced. and mashups (combining data from multiple Web sources). and developers must face a learning curve to become adept at programming for any specific platform. however. Mobile browsers are adopting many of the same sophisticated capabilities as desktop browsers. such as having every sector use only one-third of the available radio channels (1/3 reuse). If delivering very high data rates is the objective. Mobile middleware is mostly used for business applications. yet other factors are equally important in determining the services and capabilities of a wireless network. RIM BlackBerry. As previously discussed.   Broadband-Wireless Deployment Considerations Much of the debate in the wireless industry is on the merits of different radio technologies. The 10 MHz radio channel could now demand as much as 30 MHz of available spectrum. have access to this much spectrum.0 technologies include items such as Ajax. such as 10 or 20 MHz wide channels. Some developers may be content targeting specific platforms. there are various developments that address the fragmentation challenge. Cloud computing. we mean that every cell sector (typically three per cell) in every cell uses the same radio channel(s). in combination with emerging OFDMA radio technologies. may need their applications to operate across multiple platforms. then the system must minimize interference. Apple iPhone. Combined with networks that have higher throughputs and lower latency. By this. fast JavaScript execution. Very few operators today. Though Java itself has presented a challenge through inconsistent implementation on devices. and a server component that acts as a proxy for the client. the mobile device market has become quite fragmented.

Additionally. One way to increase performance is by using flatter architectures. The mismatch between backhaul capabilities and radio performance in some networks is one reason that user rates on some 3G systems are lower than theoretical rates. September 2009 Page 29 . Rysavy Research / 3G Americas. in turn. especially with respect to latency. meaning a less hierarchical network with more direct routing from mobile device to end system. Finally. With many cell sites today serviced by just a small number of T1/E1 circuits. as well as UMTS equipment. all-new MSC and/or SGSN products are capable of supporting both GSM and UMTS-HSPA radio-access networks. HSUPA. a system claiming spectrum efficiency of 1. GSM operators have largely upgraded data service to EDGE. and any new GSM network includes EDGE capability. transceivers. and so forth—need. Since installing GPRS. Operators have deployed UMTS-HSPA worldwide. Any constraint on the transport system below 100 Mbps will restrict the range of achievable throughput and. any technology’s ability to reach its peak spectrum efficiency is somewhat contingent on the system’s ability to reach the instantaneous peak data rates allowed by that technology. new HSPA equipment will be upgradeable to LTE through a software upgrade. because of the way it affects TCP/IP traffic. much of the GSM/GPRS core network can be used. at most. several factors facilitate deployment. the circuits connecting the cell sites to the core network must be able to handle the increased load. it can be misleading to say that one wireless technology outperforms another without a full understanding of how that technology will be deployed in a complete system that also takes spectrum into account. Feature and Network Roadmap GSM operators first enhanced their networks to support data capability through the addition of GPRS infrastructure with the ability to use existing cell sites. Second. As the throughput of the radio link increases. the overall network topology also plays an important role. This means that all core-network elements above the Serving GPRS Support Node (SGSN) and Mobile Switching Center (MSC)—the Gateway GPRS Support Node (GGSN). and MBMS are being designed so that the same upgraded UMTS radio channel can support a mixture of terminals including those based HSPA to LTE-Advanced. the Home Location Register (HLR). a software upgrade to support 3G UMTS-HSPA.5 bps/Hz (as described above) might rely on the ability to reach 100 Mbps instantaneously to achieve this level of spectrum efficiency. And while early 3G deployment used separate 2G/3G SGSNs and MSCs. For example. most UMTS cell sites can be collocated in GSM cell sites enabled by multi-radio cabinets that can accommodate GSM/EDGE. and there are many available and emerging wireline technologies such as VDSL and optical Ethernet. New features such as HSDPA. How traffic routes through the core network—how many hops and nodes it must pass through—can influence the overall performance of the network. An OFDMA system with 1. operators will have to significantly upgrade backhaul capacity to obtain the full benefit of next-generation wireless technologies. Although UMTS involves a new radioaccess network. First.5 bps per hertz (Hz) of spectral efficiency in 10 MHz on three sectors has up to 45 Mbps average cell throughput. each able to carry only 1. The core EPC/SAE network for 3GPP LTE emphasizes just such a flatter architecture. Operators are actively enhancing their backhaul approaches.0 Mbps. Low latency is critical to achieving very high data rates. impact the spectral efficiency of the system. In summary. and interconnection facilities. billing and subscriber administration systems. Similarly. as well as competitive point-to-point microwave systems that make this possible. service platforms.Backhaul is another factor.5/2.

The upgrade to LTE will be relatively straightforward. so that operators can use empty space in an old base station for LTE baseband cards. with new LTE infrastructure having the ability to reuse a significant amount of the UMTS-HSPA cell site and base station including using the same shelter. Note also that most UMTS terminals today support GSM. This flexibility assures the maximum degree of forward. tower. HSUPA) operating in a Release 5 mode. Release 5.1 GHz AWS band and the recently auctioned 700 MHz bands in the US.html.qualcomm. antennas. as mentioned previously. multi-mode chipsets will enable devices to easily operate across UMTS and LTE networks. which is described in detail later in this paper.com/press/releases/2008/080207_Qualcomm_to_Ship. because their multimode GSM-HSPA devices can seamlessly hand off between networks. Alternatively. New multistandard radio units (HSPA and LTE). as well as LTE-only baseband cards. For example. and Release 6 terminals. Release 5.on 3GPP Release 99. HSPA+. This will result in both reduced deployment costs and reduced latency. EV-DO Rev B. EV-DO Rev B and LTE. Rysavy Research / 3G Americas. Different vendors have different so-called “zero-footprint” solutions allowing operators to use empty space to enable re-use of existing sites without the need for any new floor space. HSPA+ and LTE. are mechanically compatible with older building practices. the intent is to reduce the number of nodes that packets must traverse. thus facilitating use across large coverage areas and multiple networks. HSPA to LTE-Advanced. Vendors and operators are planning LTE commercial deployments beginning in 2010. thus enabling re-use of existing sites without the need for any new floor space. HSDPA) can support Release 99. and Release 6. September 2009 Page 30 . and UMTS.7/2. because these networks share many of the same aspects including:       Packet-data architecture Cell sites Antenna systems Backhaul circuits Subscriber account management Service platforms Users largely don’t even need to know to what type of network they are connected.and backwardcompatibility. a network supporting Release 6 features can support Release 99. The key enabling technology is EPC/SAE. such as HSPA. On the device side.33 One important and interesting aspect of technology deployment is that an advanced technology such as LTE enables operators to upgrade prior technologies. and Release 6 terminals (for example. a network supporting Release 5 features (for example. one chipset vendor has announced a series of chips that support the following combination of technologies: UMTS. An operator can add LTE capability simply by adding an LTE baseband card. Release 5. Base station equipment is available for many bands including the 1. The changes being planned for the core network are another aspect of evolution. Examples include: 33 http://www. operators can minimize the costs of managing GSM/EDGE and UMTS networks. In other words. power supply and climate control. Once deployed. Here.

once it is deployed. integrated voice/multimedia/presence/location) 2010 Evolved EDGE capabilities available to significantly increase EDGE throughput rates HSPA+ peak speeds further increased to peak rates of 42 Mbps based on Release 8 LTE introduced for next-generation throughput performance using 2X2 MIMO Advanced core architectures available through EPC/SAE. Today. supporting voice on HSPA via VoIP will be a much simpler task as it can share the same core IP network as LTE. Table 5: Expected UMTS/LTE Feature and Capability Availability Year 2009 Features Networks and devices capable of Release 7 HSPA+. showing an approximate doubling of throughput per year. the separate GSM/EDGE Access Network (GERAN). Rysavy Research / 3G Americas. For actual users with multimode devices. Device processing power. Figure 9 presents the continuing advances in HSPA and LTE. Since LTE uses an IP core. plotted over time.g.. September 2009 Page 31 . 30 Mbps or higher). primarily for LTE but also for HSPA+. thus lowering total network cost and improving integrated operation of the separate access networks.. VoIP for HSPA. providing benefits such as integration of multiple network types and flatter architectures for better latency performance Most new services implemented in the packet domain over HSPA+ and LTE 2011 and later 2012 LTE enhancements such as 4X2 MIMO and 4X4 MIMO LTE-Advanced specifications completed LTE-Advanced potentially deployed in initial stages Over time. nearly all UMTS phones and modems support GSM /EDGE.g. 50 Mbps or higher) will provide sufficient processing in the device to also support very high HSPA rates (e. UMTS Access Network (UTRAN). boosting HSPA peak speeds to 28 Mbps Enhanced IMS-based services (for example.  Table 5 shows the rollout of EDGE/HSPA/LTE features over time. including MIMO. Supporting the high throughput rates with LTE (e. and core-infrastructure elements will undergo consolidation. the networks they access will be largely transparent. HSPA to LTE-Advanced.

Release 99 specifications were completed in 2000. most networks and devices will be tri-mode—supporting GSM. By then. The history of wireless-network deployment provides a useful perspective.6M HSUPA 5. even as operators start to deploy LTE networks at the end of this decade and the beginning of the next. During these years.Figure 9: Peak Rates for Downlink and Uplink over Time34 DL LTE(20MHz) 300M Downlink Speeds DL LTE(20MHz) 140M 100 Mbps MIMO/64QAM 42M MIMO 2x2 28M Mbps UL LTE (10MHz) 50M UL LTE (10MHz) 25M HSDPA 14. however. Only over many years. most worldwide subscribers will still be using GSM. most new subscribers will be taking advantage of UMTS.4M 10 Mbps HSUPA/16QAM 11M HSDPA 7. and LTE. will most network usage migrate to UMTS. 34 Source: A 3G Americas’ member company. which in 2009 is still growing its subscriber base. Rysavy Research / 3G Americas. Figure 10 shows the relative adoption of technologies over a multi-decadal period and the length of time it takes for any new technology to be adopted widely on a global basis. HSPA to LTE-Advanced.2M HSDPA 3.8M HSUPA 1. only now is UMTS deployment and adoption starting to surge. GSM. obtain devices with corresponding capabilities and interoperability tests DL R’99-384k UL R’99 384k 100 kbps 2004 kbps 2005 2006 2007 2008 2009 2010 2011 2012 2013 Despite rapid UMTS deployment. and HSPA+ specifications were completed in 2007. September 2009 Page 32 . Although it’s been more than a decade since work began on the technology.6M Mbps Uplink Speeds HSDPA 1. UMTS. market momentum means that even by the end of the decade. it will probably be the middle of the next decade before a large percentage of subscribers are actually using LTE.5M 1 Mbps • HSPA DL and UL peak throughputs expected to double every year on average  • Limitations not induced by the technology itself but time frames required to upgrade infrastructure and transport networks. The UMTS Task Force established itself in 1995. Similarly. as subscribers upgrade their equipment. was specified in 1990 with initial networks deployed in 1991.

Figure 10: Relative Adoption of Technologies35 LTE Relative Subscriptions UMTS/HSPA GSM/EDGE 1990 2000 2010 2020 2030 One option for GSM operators that have not yet committed to UMTS. GSM and WiMAX. EV-DO Rev A/B and LTE and. could migrate 1xRTT users to LTE including voice service. September 2009 Page 33 . in the long term. a CMDA2000 operator in scenario A could in the medium term deploy a combination of 1xRTT. Deployment Scenarios There are many different scenarios that operators will use to migrate from their current networks to future technologies such as LTE. Figure 11 presents various scenarios including operators who today are using CDMA2000. Rysavy Research / 3G Americas. In scenario B. as shown in the first bar. in the long term. and do not have an immediate pressing need to do so. could migrate EV-DO data traffic to LTE. is to migrate directly from GSM/EDGE or Evolved EDGE to LTE with networks and devices supporting dual-mode GSM-EDGE/LTE operation. 35 Source: Rysavy Research projection based on historical data. HSPA to LTE-Advanced. For example. UMTS. a CDMA2000 operator with just 1xRTT could introduce LTE as a broadband service and.

2009.cdg. 1xRTT is currently the most widely deployed CDMA2000 version.Figure 11: Different Deployment Scenarios for LTE36 3GPP and 3GPP2 both have specified detailed migration options for current 3G systems (UMTS-HSPA and EV-DO) to LTE. This section of the paper looks at the relationship between GSM/UMTS/LTE and some of these other technologies. CDMA2000 CDMA2000. 0 networks and 62 EV-DO Rev. operators are deploying other wireless technologies to serve both wide and local areas. A number of operators have deployed or are deploying 1xEV-DO where a radio carrier is dedicated to high-speed data functions. In June 2009.org. 0 or Rev. there were 106 EV-DO Rel. Source: www. A networks deployed worldwide. A radio-interface specifications. Page 34 HSPA to LTE-Advanced. which has spectral 36 37 Source: A 3G Americas’ member company. June 5. Competing Technologies Although GSM-HSPA networks are dominating global cellular-technology deployments. is the other major cellular technology deployed in many parts of the world. September 2009 . 3GPP operators are likely to have a competitive cost advantage over 3GPP2 operators.37 Currently deployed network versions are based on either Rel. Data-Optimized (1xEV-DO) versions. EV-DO Rev. A incorporates a more efficient uplink. consisting principally of 1xRTT and One Carrier-Evolved. Due to economies of scale for infrastructure and devices. Rysavy Research / 3G Americas.

although average throughputs for high level network loading are similar. 2005. A achieves a lower typical performance level than HSPA. HSPA to LTE-Advanced. in which data only constitutes a small percentage of total network traffic. and the 1xRTT channels offer only mediumspeed data. this is not a key issue. Under low. Operators have quoted 400 to 700 kilobits per second (kbps) typical downlink throughput for EV-DO Rev.efficiency similar to that of HSUPA. efficient scheduling. because of the lower peak achievable data rates. EV-DO uses many of the same techniques for optimizing spectral efficiency as HSPA including higher order modulation. Source: Sprint press release January 30. this limitation will cause suboptimal use of radio resources. Operators started to make EV-DO Rev. July 29. The EV-DO channel is not available for circuit-switched voice.39 One challenge for EV-DO operators is that they cannot dynamically allocate their entire spectral resources between voice and high-speed data functions.25 MHz Channels Another limitation of using a separate channel for EV-DO data services is that it currently prevents users from engaging in simultaneous voice and high-speed data 38 39 Source: Verizon Broadband Access Web page.25 MHz radio channels. it achieves spectral efficiency that is virtually the same as HSPA.4 Mbps for EV-DO Rev.to medium-load conditions. Rysavy Research / 3G Americas. Figure 12: Radio Resource Management 1xRTT/1xEV-DO versus UMTS-HSPA Three 1. Figure 12 illustrates this severe limitation. September 2009 One 5 MHz Channel Page 35 . A commercially available in 2007. But as data usage expands. For these reasons. resulting in lower theoretical peak rates. turbo-coding. and adaptive modulation and coding. 038 and between 600 kbps and 1. A. compared to the 5 MHz channels UMTS uses. EV-DO or EVDO Rev. 2007. In the current stage of the market. The 1x technologies operate in the 1.

Tight coupling. Such an approach by itself does not necessarily increase overall capacity. to its credit. where the amount of power each uses can be dynamically adjusted. but it does offer users higher peak-data rates. Even then. operators will face difficult choices: How many radio channels at each base station should be made available for 1xRTT to support legacy terminals versus how many radio channels should be allocated to EV-DO. Time Warner Cable. WiMAX WiMAX has emerged as a potential alternative to cellular technology for wide-area wireless networks. WiMAX has gained the greatest traction in developing countries as an alternative to wireline deployment. Based on OFDMA and recently accepted by the International Telecommunications Union (ITU) as an IMT-2000 (3G technology) under the name OFDMA TDD WMAN (Wireless Metropolitan Area Network). vendors have continued to enhance HSPA. Sprint Nextel and others (Intel. however. and perceived WiMAX advantages are no longer apparent. A. Beyond Rev. however. many of its innovations have been brought to market ahead of competing technologies. and protocol optimizations to reduce packet overhead. More likely. Loose coupling. as well as addressing problems such as jitter. B. has come closer to reality. The two systems intercommunicate with network-controlled make-before-break handovers. Tight coupling allows maintenance of data sessions with the same IP address. Rysavy Research / 3G Americas. Other enhancements are planned for EV-DO. this network was available in Atlanta. This involves little or no inter-system functionality. Google.services. Comcast. and many CDMA2000 operators including Verizon have announced their intentions to migrate to LTE. EV-DO will eventually provide voice service using VoIP protocols through EV-DO Rev.25 MHz radio channels in 20 MHz—significantly boosting peak theoretical rates to 73. UMTS allows both circuit-switched and packet-switched traffic to occupy the same radio channel. 3GPP2 had defined an OFDM-based technology called Ultra Mobile Broadband (UMB). This will likely involve a more complex implementation than loose coupling. HSPA to LTE-Advanced. But as WiMAX. Clearwire. A. This work item. MIMO and 64 QAM in the downlink. WiMAX is trying to challenge existing wireless technologies—promising greater capabilities and greater efficiencies than alternative approaches such as HSPA. for example—and being able to initiate and receive phone calls while maintaining their data sessions. This makes it simple to migrate users over time from circuit-switched voice to packet-switched voice. and resources are released in the source system prior to handover execution. with performance characteristics similar to LTE. particularly mobile WiMAX. an operator would combine three radio channels in 5 MHz. 3GPP2 has defined technical means to integrate CDMA2000 networks with LTE along two available approaches: 1.5 Mbps. 2. and 16 QAM in the uplink. September 2009 Page 36 . whereas this is possible with UMTS and HSPA. In June 2009. In the United States. There are also a number of planned improvements for CDMA2000 1xRTT that will result in increased voice capacity. Many users enjoy having a tethered data connection from their laptop—by using Bluetooth. 3GPP2 has defined EV-DO Rev. Instead. Baltimore and Portland. has been terminated as the standard had no commercial support. and Bright House Networks) have created a joint venture to deploy a nationwide WiMAX network. which can combine up to 15 1. including femtocell support. which includes a higher speed uplink. CDMA2000 is clearly a viable and effective wireless technology and. In contrast. QoS mechanisms in the network.

0.5 enhancements include MAC overhead reductions for VoIP (persistent scheduling). IEEE 802. the WiMAX Forum has defined a new profile called WiMAX Release 1. universities. thus.16-2004-certified equipment. In addition to operator-hosted access solutions. Current WiMAX profiles emphasize TDD operation.16e-2005 (referred to as mobile WiMAX) makes the most sense in licensed bands.16-2004 standard. Mobile WiMAX 2.40 Like GSM-HSPA. The IEEE has also completed a mobile-broadband standard—IEEE 802. location-based services support. 2009”. sub-10 GHz operation. IEEE 802. The original specification. load balancing. 5. that it would launch 10 additional markets on September 1. Clearwire announced on August 3. including one based on OFDM-256 and one based on OFDMA. August 3rd. Beyond Release 1. Unlike IEEE 802. 2009 HSPA to LTE-Advanced.16-2004 networks. Release 1. handover optimizations. and uplink MIMO.16-2004 also supports point-tomultipoint communications. will be designed to address the performance requirements being developed in the ITU IMT-Advanced Project and will be standardized 40 Source: Clearwire Press Release. 2009. The next major step in the evolution of IEEE 802. This standard does not compete directly with cellular-data and private Wi-Fi networks. was completed in 2001 and intended primarily for telecom backhaul applications in point-to-point.16 occurred in 2004 with the release of the IEEE 802. and cellular backhaul for connections from cellular base stations to operator infrastructure networks. although there has been little or no development in this area. Potential applications include wireless Internet Service Provider (ISP) service.16. private entities such as municipal governments. FDD operation.8 GHz) for local connectivity.0. closedloop MIMO (FDD mode only). IEEE 802. handovers across base stations. downlink adaptive modulation and coding. which operates in both licensed and unlicensed bands. more formally. September 2009 Page 37 . Vendors are now delivering IEEE 802. Like the original version of the standard. Rysavy Research / 3G Americas. 64 QAM in the uplink. it can provide complementary services. “Clearwire to Officially Launch CLEAR 4G Service in 10 Markets on September 1. and non-line-of-sight communications. line-of-sight configurations using spectrum above 10 GHz. Vendors can design equipment for either licensed or unlicensed bands. Initial mobile WiMAX networks will be deployed using 2X2 MIMO. and corporations will be able to use this version of WiMAX in unlicensed bands (for example.5 includes various refinements intended to improve efficiency and performance and could be available for deployment in a similar timeframe as LTE.16-2004.OR.16 uses a radio interface based on a single-carrier waveform.0.16e-2005—that adds mobility capabilities including support for radio operation while mobile. WiMAX is not a single technology. TDD and 10 MHz radio channels in a profile defined by the WiMAX Forum known as WiMAX Wave 2 or. meaning that subscriber stations are typically immobile. as WiMAX System Profile 1. It added multiple radio interfaces. This original version of IEEE 802. Operators have begun limited mobile WiMAX network deployments in 2009. Mobile WiMAX release 1. A subsequent version. 2009. Mobile WiMAX networks are not backward-compatible with IEEE 802. operation is fixed. and handovers across operators. local telephony bypass (as an alternative to cable modem or DSL service).5 with product certification expected by the end of 2009. it is a family of interoperable technologies.

Finally. IEEE 802. that IEEE 802. For example. WiMAX has the following technical disadvantages: 5 msec frames instead of 1 msec frames. 2008. OFDMA)” above. Source: Informa WiMAX projection.in a new IEEE standard. 2009-2014. Even over the next five years. Very few operators today have access to spectrum for WiMAX that would permit them to provide widespread coverage. “WiMAX: Mobilizing the Internet”. HSPA to LTE-Advanced. and Hybrid Automatic Repeat Request (HARQ).16m will be available in 2011. 2008. As discussed in the section “Technical Approaches (TDMA. It is not possible for one cell site to be transmitting and an adjacent cell site to be receiving at the same time. the performance will likely be somewhat less than HSPA due to increased overhead and other design issues. IEEE International Symposium on Personal. the number of WiMAX subscribers is likely to be quite low. The principal difference from HSPA is IEEE 802. while WiMAX uses OFDMA. Sprint Nextel. OFDM provides a potential implementation advantage for wide radio channels (for example. for example. September 2009 Page 38 . Relative to LTE. by then.16e exploits significant radio innovations similar to HSPA+ and LTE. adaptive modulation and coding.1 million by 201343 while Maravedis predicts a lower 75 million WiMAX subscribers by the end of 201444. 41 Ali Tabassi. In reference to economies of scale. including high-order modulation. supplied to 3G Americas. from a technology standpoint. it faces challenges such as economies of scale and technology maturity. Chase combining instead of incremental redundancy. March 5. Different operators in the same band must either coordinate their networks or have guard bands to ensure that they don’t interfere with each other. CDMA. But by the time it becomes widely available. GSM-HSPA subscribers number in the billions.5 GHz band in the United States may be used for both TDD and FDD operation. According to Sprint Nextel.16e-2005 contains some aspects that may limit its performance.16e-2005’s use of OFDMA. IEEE 802. Indoor and Mobile Radio Communications: Anders Furuskär et al “The LTE Radio Interface – Key Characteristics and Performance”.” Maravedis. LTE will not be that far from deployment. 42 43 44 Source: “WiMAX and Broadband Wireless Access Equipment Market Analysis. This may introduce problems as more operators introduce networks in the same spectrum band. Rysavy Research / 3G Americas. 2009. Thus. the 2. and some aspects—such as power control of the mobile station—are provided using MAC signaling messages rather than the fast power control used in WCDMA and other technologies. particularly in scenarios in which a sector contains a large number of mobile users. 10 to 20 MHz). mobile WiMAX on paper may be slightly more capable than today’s available versions of HSPA. Trends and Forecasts. Informa projects 82.41 WiMAX employs many of the same mechanisms as HSPA to maximize throughput and spectral efficiency. Although IEEE 802. June 2009. In 5 to 10 MHz radio channels. coarser granularity for modulation and coding schemes and vertical coding instead of horizontal coding. mobile WiMAX will actually have to compete against evolved HSPA systems that will offer largely similar capabilities.16m. efficient coding. It should be noted.42 One deployment consideration is that TDD requires network synchronization. however. The performance of the MAC layer is inefficient when scheduling large numbers of users. Fierce Wireless Webcast. June 1. there is no evidence indicating that WiMAX will have any performance advantage compared to HSPA+. Further.

45 Arthur D. May 2007”. Little Limited.77 dB higher transmit power. First. With a 2:1 TDD system. “HSPA. and capacity. January 18. complemented by new dual-technology devices or significant operator investments. “HSPA. 3G systems have a significant link budget advantage over mobile WiMAX because of softhandoff diversity gain and an FDD duplexing advantage over TDD. Second. HSPA+ in Release 8— with a peak rate of 42 Mbps—exceeds mobile WiMAX in 10 MHz in TDD 2:1 DL:UL using 2X2 MIMO with a peak rate of 40 Mbps. Building new networks for broadband wireless mandates substantial capacity per subscriber. Some have cited intellectual property rights as an area in which WiMAX has an advantage. the undisputed choice for mobile broadband. Little reports that the radii of typical HSPA cells will be two to four times greater than typical mobile WiMAX cells for high-throughput operation.7 times more cell sites than HSPA. to support such claims. Matching the cellular footprint with WiMAX will require national roaming arrangements. 2006. Finally. Rysavy Research / 3G Americas. It is not clear how easily the available revenue per subscriber will be able to finance large-scale deployment of network capacity. however.47 Given that many real world deployments of HSPA will occur at frequencies such as 850 MHz.outlook4mobility. Today’s cellular networks can finance the deployment of data capabilities through a successful voice business.5 GHz will be at a significant disadvantage. As for data performance.46 One vendor estimates that for the same power output. To obtain the same cell edge data rates. May 2007”. Despite numerous attempts. 27 March 2007. 46 45 Source: "HSPA and mobile WiMAX for Mobile Broadband Wireless Access". And if the future is in multimedia services such as movie downloads. the mobile WiMAX industry is in its infancy. commentary: “Will Data-Only Networks Ever Make Money?” http://www. the large HSPA vendors have invested heavily in these technologies— hopefully giving them significant leverage with which to negotiate reasonable intellectual property rights (IPR) rates with other vendors. no terrestrial wireless-data-only network has ever succeeded as a business.49 Although there is discussion of providing voice services over WiMAX using VoIP. mobile-voice users demand ubiquitous coverage—including indoor coverage. and LTE at 700 MHz. 47 Source: Ericsson public white paper.48 The sometimes-quoted peak rate of 63. In fact. With respect to spectral efficiency. and there is considerable lack of clarity when it comes to how different companies will assert and resolve IPR issues. September 2009 Page 39 . Consumers who download 1 gigabyte of data each month represent a ten times greater load on the network than a 1. frequency. WiMAX deployments at 2. 48 49 Source: Andy Seybold. Arthur D. publicly available information. the mobile system must transmit at 4.One specific area in which WiMAX has a technical disadvantage is cell size.000-minute-a-month voice user. the undisputed choice for mobile broadband.com/commentary2006/jan1806. it is important to recognize that downloading a single DVD-quality movie—even with advanced compression— consumes approximately 2 gigabytes. wireless-data business models must also be considered. the reverse link only transmits one third of the time. Source: Ericsson public white paper. There is little substantial. as discussed in the section “Spectral Efficiency” that follows. mobile WiMAX requires 1.htm HSPA to LTE-Advanced. WiMAX is comparable to HSPA+.4 Mbps for mobile WiMAX in 10 MHz assumes no bandwidth applied to the uplink.

The latest 802.20 has not gained any momentum at this point in time. Various organizations are looking at integrating WLAN service with GSM-HSPA data services. and hotels. fast-food restaurants. However. and IEEE 802. whereas 3G systems can provide access over much larger coverage areas. because the hotspot can provide broadband services in extremely dense user areas and the cellular network can provide broadband services across much larger areas. operators—including cellular operators—are offering hotspot service in public areas such as airports. it is not clear whether there is sufficient momentum in this standard to make it a viable technology.11i enables robust security.20 is a mobile-broadband specification developed by the Mobile Broadband Wireless Access Working Group of the IEEE that was completed in 2008. UMA.11 family of technologies has experienced rapid growth. metro Wi-Fi and 3G are more likely to be complementary in nature. Even then. The IEEE is developing a mesh networking standard—IEEE 802.IEEE 802. Complementary standards increase the attraction of the technology.4 GHz radio channel. and 802. 802. Many early network deployments have experienced poorer coverage than initially expected. Although some industry observers are predicting that these systems will have an adverse effect on 3G data services.    HSPA to LTE-Advanced. Coverage in most metro systems is designed to provide an outdoor signal. With vendors focused heavily on LTE and WiMAX for next-generation wireless services.20 is very similar to UMB. Wi-Fi can generally provide better application performance over limited coverage areas. wherein access points forward packets to nodes that have backhaul connections. For the most part. At this time. UMB has been cancelled.11s—but this may not be ready until 2010. and 3GPP has multiple initiatives that address WLAN integration into its networks.11n offers users throughputs in excess of 100 Mbps and improved range through use of MIMO. Operation is in unlicensed bands in the 2.20 IEEE 802. Note that 802. hotspots are complementary with cellular-data networks. no operator has committed to the possible standard. The GSM Association has developed recommendations for SIM-based authentication of hotspots. including 3GPP System to WLAN Interworking. 802. mainly in private deployments. and the number of recommended access points per square mile has increased steadily. These systems are based on a mesh technology. interference between public and private systems is inevitable. Rysavy Research / 3G Americas. Leveraging this success. the IEEE 802. IMS.11e provides quality-of-service enabling VoIP and multimedia. and EPC/SAE. 802. Wi-Fi and Municipal Wi-Fi Systems In the local area. it is not clear that vendors will adopt this standard for outdoor systems. necessary for voice handover across access points. Metro systems today are still quite immature and face the following challenges:  Many city projects have been discontinued due to the difficulty of providing a viable business model. the signal does not penetrate many buildings in the coverage area and repeaters are needed to propagate the signal indoors.4 GHz. As such. Many cities are now deploying metro Wi-Fi systems that will provide Wi-Fi access in downtown areas. Given only three relatively non-overlapping radio channels at 2.11r provides fast roaming. Today’s mesh systems are all proprietary. September 2009 Page 40 .11 standard.

Another method is to disclose throughputs actually measured in deployed networks with applications such as File Transfer Protocol (FTP) under favorable conditions. Comparison of Wireless Technologies This section of the paper compares the different wireless technologies looking at throughput. Data Throughput Data throughput is an important metric performance. Nevertheless. throughput rates are lower. there are still considerable expenses and networking considerations in backhauling a large number of outdoor access points. Table 6 presents the technologies in terms of peak network throughput rates. because they depend on a multitude of operational and network factors. Except when the network is congested. however. The projected typical rates for HSPA+ and LTE show a wide range. Rysavy Research / 3G Americas. primarily in the US. however. and as more devices support both 3G and Wi-Fi. Unfortunately. This paper refers to this rate as the “peak user rate. Even if the radio network can deliver this speed. users can look forward to multiple access options. the values provide a good indication of what users can typically expect. This is because these technologies are designed to exploit favorable radio conditions to achieve very high throughput rates. peak userrates (under favorable conditions) and typical rates. latency. actual application throughput may be 10 to 20 percent lower (or more) than this layer-2 value.” Average rates. the paper presents a table that summarizes the competitive position of the different technologies across multiple dimensions. Under poor radio conditions. which assume light network loading (as low as one active data user in the cell sector) and favorable signal propagation. also quote typical throughput rates. These rates are based on throughput tests the operators have done across their operating networks and incorporate a higher level of network loading.” This refers to the fastest possible transmission speed over the radio link. Finally. Technical issues will likely be resolved over time. One method of representing a technology’s throughput is what people call “peak throughput” or “peak network speed. and some number of projects are still proceeding. metro networks have attracted considerable interest. Peak network speed is also usually quoted at layer 2 of the radio link. the ways in which statistics vary tremendously. This often results paper is to realistically represent the capabilities for quantifying network throughput various organizations quote throughput in misleading claims. Because of protocol overhead. September 2009 Page 41 . Although the operators do not disclose the precise methodology they use to establish these figures. Although mesh architecture simplifies backhaul. HSPA to LTE-Advanced. actual capability of the technology. spectral efficiency. are lower than this peak rate and difficult to predict. Some operators. however. and market position. the majority of users should experience throughput rates higher than one-half of the peak-achievable rate. It omits values that are not yet known such as those associated with future technologies. and it is generally based on the highest order modulation available and the least amount of coding (error correction) overhead. This number is useful because it demonstrates the highend. The intent of this of these technologies. other aspects of the network— such as the backhaul from base station to operator-infrastructure network—can often constrain throughput rates to levels below the radio-link rate.

Type 2 mobile. 16 slots downlink (dual carrier) at DBS-12 (118. 4 slots uplink.4 kbps/slot). 8 slots uplink. September 2009 Page 42 .6 kbps52 400 kbps peak 150 to 300 kbps typical expected 947. 54 55 HSPA to LTE-Advanced. Type 1 mobile.2 kbps55 768 kbps 350 kbps peak 200 to 300 kbps typical 384 kbps 350 kbps peak 200 to 300 kbps typical 384 kbps 384 kbps > 5 Mbps 2 Mbps > 1. 51 52 53 A type 2 Evolved EDGE MS can receive on up to 16 timeslots using two radio channels and can transmit on up to eight timeslots in one radio channel using 32 QAM modulation (with turbo coding in the downlink). UBS-12 (118.4 kbps/slot). Rysavy Research / 3G Americas.6 kbps 236.4 kbps/slot).5 Mbps 350 kbps peak Peak and/or Typical User Rate 473. Type 2 mobile. Type 1 mobile.Table 6: Throughput Performance of Different Wireless (Blue Indicates Theoretical Peak Rates. 10 slots downlink (dual carrier). DBS-12(118.048 Mbps 384 kbps HSDPA Initial Devices (2006) HSDPA HSPA56 Initial Implementation 1.2 Mbps > 1 Mbps peak 50 A type 1 Evolved EDGE MS can receive on up to ten timeslots using two radio channels and can transmit on up to four timeslots in one radio channel using 32 QAM modulation (with turbo coding in the downlink). Green Typical) Technologies Downlink Peak Network Speed EDGE (type 2 MS) EDGE (type 1 MS) (Practical Terminal) Uplink Peak and/or Typical User Rate Peak Network Speed 473.8 kbps Evolved EDGE (type 1 MS)50 1184 kbps51 1 Mbps peak 350 to 700 kbps typical expected (Dual Carrier) Evolved EDGE (type 2 MS)53 UMTS WCDMA Release 99 UMTS WCDMA Release 99 (Practical Terminal) 1894.8 kbps 200 kbps peak 70 to 135 kbps typical 473. UBS-12 (118.4 kbps/slot).4 Mbps 7.454 kbps 2.8 Mbps 14.6 kbps 200 kbps peak 70 to 135 kbps typical 236.

UL 16 QAM) HSPA+ (2X2 MIMO. DL 64 QAM.5 Mbps 130 kbps peak 86 Mbps 153 kbps 307 kbps 153 kbps 1.5 Mbps 1 Mbps to 4 Mbps 7.76 Mbps 1.2 Mbps 14. September 2009 Page 43 . DL 16 QAM. UL 16 QAM) Uplink Peak Network Speed Peak and/or Typical User Rate peak 500 kbps to 1.8 Mbps 150 kbps peak > 1 Mbps 130 kbps peak CDMA2000 1XRTT CDMA2000 1XRTT CDMA2000 EV-DO Rel 0 CDMA2000 EV-DO Rev A 56 High Speed Packet Access (HSPA) consists of systems supporting both High Speed Downlink Packet Access (HSDPA) and High Speed Uplink Packet Access (HSUPA). Rysavy Research / 3G Americas. UL 16 QAM) HSPA+ (2X2 MIMO. 59 HSPA to LTE-Advanced.” January 15.1 Mbps > 1 Mbps peak > 1.7 Mbps typical57 HSPA Current Implementation HSPA HSPA+ (DL 64 QAM. Downlink throughput will be about half in a 2 x 10 MHz deployment.5 Mbps 23 Mbps 173 Mbps 4 Mbps to 24 Mbps (in 2 x 20 MHz)59 58 Mbps LTE (4X4 MIMO) 326 Mbps 153 kbps 307 kbps 2.4 Mbps 21. UL 16 QAM. “Vodafone Trials HSPA+ Mobile Broadband at Speeds of Up To 16Mbps. Typical downlink and uplink throughput rates based on AT&T press release. DL 64 QAM.6 Mbps HSPA+ (2X2 MIMO. Dual Carrier) LTE (2X2 MIMO) 28 Mbps 42 Mbps 84 Mbps 11. 2008 57 58 Source: Vodafone press release.Downlink Peak Network Speed Peak and/or Typical User Rate peak 700 kbps to 1.5 Mbps to 7 Mbps 13 Mbps peak58 11. June 4.5 Mbps 11.76 Mbps 5.4 Mbps 3.2 Mbps typical 5. 2009.

0 (10 MHz TDD.4 Mbps 27 Mbps 9.5 IEEE 802. Significantly. the 75 percentile is at 5 Mbps. DL/UL=3. Rysavy Research / 3G Americas. 2x2 MIMO) WiMAX Release 1.Downlink Peak Network Speed Peak and/or Typical User Rate peak 600 kbps to 1.16m 46 Mbps 2 to 4 Mbps average 61 4 Mbps TBD TBD TBD TBD HSDPA Throughput in Representative Scenarios It is instructive to look at actual HSDPA throughput in commercial networks. meaning that 75% of samples are below 5 Mbps and 25% are above. Source: Sprint web page.5 Mbps WiMAX Release 1. Figure 13 shows the throughputs measured in one network with voice and data in one Western European country across three larger cities.3 Mbps 73. half of all the measurements showed 4 Mbps or higher throughput. The data shows the percentage of samples on the X axis that fall below the throughput shown on the Y axis. 60 61 Typical downlink and uplink throughput rates based on Sprint press release January 30. For example. 2007. HSPA to LTE-Advanced. September 2009 Page 44 . June 2009.4 Mbps typical60 CDMA2000 EV-DO Rev B (3 radio channels MHz) CDMA2000 EV-DO Rev B Theoretical (15 radio channels) Uplink Peak Network Speed Peak and/or Typical User Rate peak 300 to 500 kbps typical 5.

September 2009 Page 45 . HSPA to LTE-Advanced.8 Mbps with poor coverage.0 0.0 5.0 2. 62 Source: 3G Americas’ member company contribution.9 Mbps when mobile.0 Throughput [Mbps] 4.0 10 0% 85 % 90 % 70 % 75 % 95 % 55 % 80 % 50 % 60 % 35 % 65 % 40 % 20 % 45 % 25 % 10 % 30 % 15 % 5% 0% In another network study. It results in a median throughput of 1. and 3.Figure 13: HSDPA Throughput Distribution in Deployed Networks62 6.0 1.0 3.8 Mbps with good coverage. Figure 14 shows the downlink throughput performance of a 7.2 Mbps device (peak data rate capability). Rysavy Research / 3G Americas. 1.

September 2009 Page 46 . the Y axis shows the cumulative distribution function.2 Mbps Device in a Commercial Network63 Good Coverage Bad Coverage Median bitrate 3. operators increased speeds to 384 kbps peak rates with peak user-achievable rates of 350 kbps. Testers measured average FTP downlink application throughput of 2.8 Mbps Median bitrate 1.2 Mbps HSDPA.0 Mbps.64 Release 99 and HSUPA Uplink Performance HSUPA dramatically increases uplink throughputs over 3GPP Release 99. this increased to 128 kbps. Many networks were initially deployed with a 64 kbps uplink rate.8 Mbps -106 dBm Mobile Median bitrate 1. and 1.Figure 14: HSDPA Performance of a 7.9 Mbps in the second network. have seen significant uplink increases. The anticipated 1 Mbps achievable uplink throughput with HSUPA can be seen in the measured throughput of a commercial network as documented in Figure 15.1 Mbps in the first network. HSPA to LTE-Advanced. Source: 3G Americas’ member company contribution. 63 64 Source: 3G Americas’ member company contribution. however. Later. and the bars show the number of samples obtained for that throughput rate on a relative basis. Rysavy Research / 3G Americas.9 Mbps These rates are consistent with other vendor information for two deployed HSPA networks that supported 7. The median bit rate is 1. Later still. The X axis shows throughput rate. Even Release 99 networks.

Until operators actually deploy complete networks. typical rates will not be available. but the data suggests that users should be able to obtain throughputs an order of magnitude higher than today’s 3G networks.2 Mbps. Source: 3G Americas’ operator member observation for 2009 conditions.68 LTE Throughput As part of the LTE/SAE/EPC Trial Initiative (LSTI).0 Mbps 80 70 60 50 40 30 20 10 0 0 70 140 280 350 490 560 630 700 840 910 1050 1120 1190 1260 210 420 770 980 1330 These rates are consistent with other vendor information for a deployed HSPA network that supported 2. a mean of 78 Mbps and a minimum of 16 Mbps. vendors are testing LTE technology. Source: 3G Americas’ member company contribution. Testers measured average FTP downlink application throughput of 1. cells are only lightly loaded.Figure 15: Uplink Throughput in a Commercial Network65 Mobile 100 90 Median bitrate 1. 2 x spreading factor (2xSF2) code configuration. HSPA to LTE-Advanced. because for a large percentage of cells and for a large percentage of time. 65 66 67 68 Source: 3G Americas’ member company contribution. peak rates are often higher than the stated typical rates.0 Mbps HSUPA66 uplink speed. Figure 16 shows LTE throughputs in a 2X2 MIMO trial network reaching a maximum of 154 Mbps. September 2009 1400 Page 47 .67 One operator has noted that in its networks. Rysavy Research / 3G Americas.

showing layer 1 throughput measured at 10 MHz bandwidth using the Extended Vehicular A 3 km/hour channel model. 69 Source: 3G Americas’ member company contribution. September 2009 Page 48 . Rysavy Research / 3G Americas. L1 Throughput Max: 154 Mbps Mean: 78 Mbps Min: 16 Mbps User Speed Max: 45 km/h Mean: 16 km/h Min: 0 km/h Suburban area with lineof-sight: less than 40% of the samples Heights of surrounding buildings: 15-25 m 100 meters 123 97 74 54 37 23 12 Figure 17 provides additional insight into LTE downlink throughput.Figure 16: LTE Measured Throughput in Test Network69 154 Base station located at x. The figure shows the increased performance obtained with the addition of different orders of MIMO. HSPA to LTE-Advanced.

Figure 18 shows how throughput rates can vary by number of active users and radio conditions. the throughput rates will go down as more users simultaneously use the network. This is largely a linear degradation. actual data payload throughput rates may be lower by an approximate 5% to 20% amount. RF Conditions and User Speed. “Initial Field Performance Measurements of LTE”. 3. throughput rates will be lower. 5 MHz) and at least a doubling of spectral efficiency. September 2009 . which one can anticipate from radio channels that are four times wider (20 MHz vs. Jonas Karlsson. Rysavy Research / 3G Americas. 2.g. 3 2008. Page 49 HSPA to LTE-Advanced. Peak rates are generally stated for the physical layer.. file downloads) result in a lower overhead ratio. Actual throughput rates that users will experience will be lower than the peak rates and will depend on a variety of factors including: 1. Like all wireless systems. Protocol Overhead. it is reasonable to expect an order of magnitude higher performance than HSPA.ericsson. http://www. Due to overhead at other layers. Peak rates depend on optimal conditions. Mathias Riback.com/ericsson/corpinfo/publications/review/2008_03/files/LTE. The higher curves are for better radio conditions.pdf. Larger packets (e.Figure 17: LTE Throughput in Various Modes70 For typical and average throughputs. Ericsson Review No. Network Loading. Under suboptimal conditions. 70 Source:. such as being at the edge of the cell or if the user is moving at high speed. The precise amount depends on the size of packets.

Page 50 HSPA to LTE-Advanced.72 Latency Just as important as throughput is network latency. Feb 2009. Each successive data technology from GPRS forward reduces latency. HSUPA brings latency down even further. Source: http://gigaom. with HSDPA networks having latency as low as 70 milliseconds (msec). 71 72 Source: LTE/SAE Trial Initiative. “Latest Results from the LSTI. Rysavy Research / 3G Americas.lstiforum.com/2009/05/15/verizons-lte-plans-get-real/. September 2009 .org. Ongoing improvements in each technology mean all these values will go down as vendors and operators fine tune their systems. defined as the round-trip time it takes data to traverse the network. Figure 19 shows the latency of different 3GPP technologies.” www. as will 3GPP LTE.Figure 18: LTE Actual Throughput Rates Based on Conditions71 Verizon Wireless has stated that it expects its LTE network to deliver 8 to 12 Mbps of throughput.

Figure 19: Latency of Different Technologies73

700 600 500

Milliseconds

400 300 200 100

GPRS Rel’97

EDGE Rel’99

EDGE WCDMA Evolved HSDPA EDGE Rel’4 Rel’99

HSPA

LTE

The values shown in Figure 19 reflect measurements of commercially deployed technologies. Some vendors have reported significantly lower values in networks using their equipment, such as 150 msec for EDGE, 70 msec for HSDPA, and 50 msec for HSPA. With further refinements and the use of 2 msec Transmission Time Interval (TTI) in the HSPA uplink, 25 msec roundtrip is a realistic goal. LTE will reduce latency even further, to as low as 10 msec in the radio-access network.

Spectral Efficiency
To better understand the reasons for deploying the different data technologies and to better predict the evolution of capability, it is useful to examine spectral efficiency. The evolution of data services will be characterized by an increasing number of users with ever-higher bandwidth demands. As the wireless-data market grows, deploying wireless technologies with high spectral efficiency will be of paramount importance. Keeping all other things equal such as frequency band, amount of spectrum, and cell site spacing, an increase in spectral efficiency translates to a proportional increase in the number of users supported at the same load per user—or, for the same number of users, an increase in throughput available to each user. Delivering broadband services to large numbers of users can best be achieved with high spectral-efficiency systems, especially

73

Source: 3G Americas' member companies. Measured between subscriber unit and Gi interface, immediately external to wireless network. Does not include Internet latency. Note that there is some variation in latency based on network configuration and operating conditions.

HSPA to LTE-Advanced, Rysavy Research / 3G Americas, September 2009

Page 51

because the only other alternatives are using more spectrum or deploying more cell sites. Increased spectral efficiency, however, comes at a price. It generally implies greater complexity for both user and base station equipment. Complexity can arise from the increased number of calculations performed to process signals or from additional radio components. Hence, operators and vendors must balance market needs against network and equipment costs. One core aspect of evolving wireless technology is managing the complexity associated with achieving higher spectral efficiency. The reason technologies such as OFDMA are attractive is that they allow higher spectral efficiency with lower overall complexity; thus their use in technologies such as LTE and WiMAX. The roadmap for the EDGE/HSPA/LTE family of technologies provides a wide portfolio of options to increase spectral efficiency. The exact timing for deploying these options is difficult to predict, because much will depend on the growth of the wireless data market and what types of applications become popular. When determining the best area on which to focus future technology enhancements, it is interesting to note that HSDPA, 1xEV-DO, and IEEE 802.16e-2005 all have highly optimized links—that is, physical layers. In fact, as shown in Figure 20, the link layer performance of these technologies is approaching the theoretical limits as defined by the Shannon bound. (The Shannon bound is a theoretical limit to the information transfer rate [per unit bandwidth] that can be supported by any communications link. The bound is a function of the Signal to Noise Ratio [SNR] of the communications link.) Figure 20 also shows that HSDPA, 1xEV-DO, and IEEE 802.16e-2005 are all within 2 to 3 decibels (dB) of the Shannon bound, indicating that there is not much room for improvement from a link layer perspective. Note that differences do exist in the design of the MAC layer (layer 2), and this may result in lower than expected performance in some cases as described previously.

HSPA to LTE-Advanced, Rysavy Research / 3G Americas, September 2009

Page 52

Figure 20: Performance Relative to Theoretical Limits for HSDPA, EV-DO, and IEEE 802.16e-200574

6 Shannon bound Shannon bound with 3dB margin HSDPA EV-DO IEEE 802.16e-2005

5 Achievable Efficiency (bps/Hz)

4

3

2

1

0 -15

-10

-5

0 5 Required SNR (dB)

10

15

20

The curves in Figure 20 apply to an Additive White Gaussian Noise Channel (AWGN). If the channel is slowly varying and the effect of frequency selectivity can be overcome through an equalizer in either HSDPA or OFDM, then the channel can be known almost perfectly and the effects of fading and non-AWGN interference can be ignored—thus justifying the AWGN assumption. For instance, at 3 km per hour, and fading at 2 GHz, the Doppler spread is about 5.5 Hz. The coherence time of the channel is thus 1 sec/5.5 or 180 msec. Frames are well within the coherence time of the channel, because they are typically 20 msec or less. As such, the channel appears “constant” over a frame and the Shannon bound applies. Much more of the traffic in a cellular system is at slow speeds (for example, 3 km/hr) rather than at higher speeds. Thus, the Shannon bound is relevant for a realistic deployment environment. As the speed of the mobile station increases and the channel estimation becomes less accurate, additional margin is needed. This additional margin, however, would impact the different standards fairly equally. The Shannon bound only applies to a single user; it does not attempt to indicate aggregate channel throughput with multiple users. It does indicate, however, that link layer performance is reaching theoretical limits. As such, the focus of future technology enhancements should be on improving system performance aspects that maximize the experienced SNRs in the system rather than on investigating new air interfaces that attempt to improve the link layer performance.

74

Source: A 3G Americas’ member company.

HSPA to LTE-Advanced, Rysavy Research / 3G Americas, September 2009

Page 53

8 1. MRxD. some of the values shown are lower (for all technologies) than the values indicated in other papers and publications. September 2009 Page 54 .7 0.3 1. Equalizer LTE 2X2 MIMO Future improvements Future improvements Rev B Cross-Carrier Scheduling Rev A. Figure 21 compares the spectral efficiency of different wireless technologies based on a consensus view of 3G Americas contributors to this paper. 5+5 MHz for UMTS-HSPA/LTE and CDMA2000. Rysavy Research / 3G Americas.5 2.8 0. because the per-antenna transfer rate (that is.9 1.2 0. Note that MIMO techniques using spatial multiplexing to potentially increase the overall information transfer rate by a factor proportional to the number of transmit or receive antennas do not violate the Shannon bound.3 2.7 1. 64 QAM HSPA+ 2X2 MIMO HSDPA MRxD. Figure 21: Comparison of Downlink Spectral Efficiency75 2.1 1.0 0.2 1. the per-communications link transfer rate) is still limited by the Shannon bound. HSPA to LTE-Advanced. The values shown are conservative and intended to be reasonably representative of real-world conditions.9 0. Equalizer Rel 1.4 2.5 0.5 1. Mix of mobile and stationary users.4 0. Most simulation results produce values under idealized conditions.3 0. and 10 MHz DL/UL=29:18 TDD for WiMAX.1 UMTS/HSPA Future improvements LTE 4X4 MIMO Spectral Efficiency (bps/Hz/sector) LTE 4X2 MIMO Future improvements HSPA+ SIC.0 2X2 MIMO HSDPA EV-DO Rev 0 UMTS R’99 LTE CDMA2000 WiMAX 75 Joint analysis by 3G Americas’ members. For instance.6 1. 3GPP studies indicate higher HSDPA and LTE spectral efficiencies than those shown below.0 1.Examples of technologies that improve SNR in the system are those that minimize interference through intelligent antennas or interference coordination between sectors and cells.5 4X2 MIMO Rel 1.5 2X2 MIMO Rel 1. It shows the continuing evolution of the capabilities of all the technologies discussed.1 2. as such.2 2.6 0.4 1.

Methods like successive interference cancellation (SIC) and 64 QAM allow gains in spectral efficiency as high as 1. The main reason that HSPA+ with MIMO is shown as more spectrally efficient than WiMAX Release 1. but EDGE itself is spectrally efficient at 0. the figure does not necessarily show the actual progression of technologies that operators will deploy to increase spectral efficiency. but are based on proprietary implementations and. some enhancements. The figure does not include EDGE.0 spectral efficiency. Downloadable codebooks in Release 9 LTE provide one avenue for such additional gains.0 with MIMO is because HSPA MIMO supports closed-loop operation with precode weighting and multi-codeword MIMO. and that WiMAX has larger control overhead in the downlink than HSPA.5. September 2009 Page 55 . such as 64 QAM. which further increases spectral efficiency by about 20 percent and exceeds WiMAX Release 1. Higher gains are possible with more advanced adaptive antenna and beam-forming algorithms. they are representative milestones in ongoing improvements in spectral efficiency. dual-carrier operation can be combined with MIMO.76 With Release 8. will be simpler for some operators to deploy than other enhancements such as 2X2 MIMO. although most of the gain is realized at 10 MHz. Relative to WCDMA Release 99. An uplink MAP zone in the downlink channel does this scheduling. Rysavy Research / 3G Americas. HSPA to LTE-Advanced. initially with 2X2 MIMO. Thus. like MRxD and MIMO.3 bps/Hz/sector. LTE is even more spectrally efficient with wider channels. Other reasons are that HSPA supports incremental-redundancy HARQ. OFDMA technology requires scheduling to avoid two mobile devices transmitting on the same tones simultaneously. whereas the latter requires additional hardware at the base station. Beyond HSPA. For instance. Mobile WiMAX also experiences gains in spectral efficiency as various optimizations.0 includes 2X2 MIMO. operators can deploy either MIMO or dual-carrier operation. 3GPP LTE will also result in further spectral efficiency gains. The gain for 4X2 MIMO is shown at a modest 15% increase for LTE. while WiMAX supports only Chase combining HARQ. Vendor estimates for spectral-efficiency gains from dualcarrier operation range from 5% to 20%. the actual gains will depend on implementation. which enables the use of SIC receivers. because the uplink in WiMAX is fully scheduled. there are terminals that employ mobile-receive diversity but not equalization. With respect to actual deployment. HSPA+ in Release 7 includes 2X2 MIMO. thus. which is close to LTE performance in 5+5 MHz channel bandwidth. With Release 9. This assumes a simplified switchedbeam approach defined in Release 8. are applied. 76 Source: 3G Americas’ member analysis.The values shown in Figure 21 are not all the possible combinations of available features. Rather. 4X2 MIMO and 4X4 MIMO. In more realistic operating scenarios. Lower spectral efficiency gains are due to full-buffer traffic assumptions. such as 10 and 20 MHz. Dual-carrier HSPA will offer a further modest gain in spectral efficiency from cross-carrier scheduling with possible gains of about 10%. HSDPA increases capacity by almost a factor of three. The former can be done as a software upgrade. WiMAX Release 1. gains will be significantly higher. Terminals with SIC can also be used with Release 7 systems.3 bps/Hz/sector. The same is true for WiMAX. Enhancements to WiMAX will come with Release 1. as well as other future enhancements. Similar gains to those for HSPA and LTE are available for CDMA2000. and then optionally with SIC. Type 3 receivers that include Minimum Mean Square Error (MMSE) equalization and Mobile Receive Diversity (MRxD) will effectively double HSDPA spectral efficiency.

September 2009 Page 56 . Finer granularity of modulation and coding schemes. Expected features include reduced MAC overhead. adaptive modulation and coding.0 OFDM in downlink and uplink WiMAX Release 1.LTE has higher spectral efficiency than WiMAX Wave 2 for a number of reasons 77:       Closed-loop operation with precoded weighting. One available improvement for LTE spectral efficiency not shown in the figure is successive interference cancellation. Modulation and Coding Scheme Granularity Fine granularity (1-2 dB apart) Coarse granularity (2-3 dB apart) 77 IEEE International Symposium on Personal. HSPA to LTE-Advanced. The following table summarizes the most important features of LTE and WiMAX technology that impact spectral efficiency.5 addresses some of these items and will thus have increased spectral efficiency. Indoor and Mobile Radio Communications: Anders Furuskär et al “The LTE Radio Interface – Key Characteristics and Performance”. This will result in a gain of 5% in a low-mobility environment and a gain of 10 to 15% in environments such as picocells in which there is cell isolation. Discrete Fourier Transform (DFT)spread OFDM in uplink Fractional pathloss compensation WiMAX Release 1. 2008. Fractional path-loss compensation enables flexible tradeoff between average and cell-edge data rates Access to the frequency domain yields larger scheduling gains Uplink Power Control Full path-loss compensation Full path-loss compensation Scheduling Channel dependent in time and frequency domains Multi-codeword (horizontal). Finer granularity enables better link adaptation precision. closed loop with pre-coding Channel dependent in time domain Channel dependent in time and frequency domains MIMO Scheme Single codeword (vertical) Single codeword (vertical).5 OFDM in downlink and uplink Impact DFT-spread OFDM reduces the peak-to-average power ratio and reduces terminal complexity. and other physical-layer enhancements. WiMAX Release 1. Incremental redundancy in error correction. Table 7: LTE and WiMAX Features Feature Multiple Access LTE OFDM in downlink. requires one-tap equalizer in base station receiver. with rank-adaptive MIMO (TDD) and with closed-loop pre-coding (FDD) Coarse granularity (2-3 db apart) Horizontal encoding enables per-stream link adaptation and successive interference cancellation receivers. Multi-codeword MIMO. Lower Channel Quality Indicator delay through use of 1 msec frames instead of 5 msec frames. Greater control channel efficiency. Rysavy Research / 3G Americas. which enables the use of SIC receivers.

HSPA to LTE-Advanced. and 10 MHz DL/UL=29:18 TDD for WiMAX.5 Chase combining Impact Incremental redundancy is more efficient (lower SNR required for given error rate) Shorter subframes yield lower user plane delay and reduced channel quality feedback delays Lower overhead improves performance 1 msec subframes 5 msec subframes 5 msec subframes Overhead / Control Channel Efficiency Relatively low overhead Relatively high overhead Relatively high overhead apart from reduction in pilots Figure 22 compares the uplink spectral efficiency of the different systems.0 Chase combining WiMAX Release 1. Mix of mobile and stationary users.8 0.7 0. Figure 22: Comparison of Uplink Spectral Efficiency78 1. 5+5 MHz for UMTS-HSPA/LTE and CDMA2000.2 0.4 0.Feature Hybrid Automatic Repeat Request (ARQ) Frame Duration LTE Incremental redundancy WiMAX Release 1. September 2009 Page 57 .5 1X4 Receive Diversity Future Improvements EV-DO Rev B.0 LTE 1X2 Receive Diversity EV-DO Rev A EV-DO Rev 0 LTE CDMA2000 WiMAX UMTS/HSPA 78 Joint analysis by 3G Americas’ members.1 HSUPA Rel 6 UMTS R’99 to Rel 5 Future Improvements HSPA+ Interference Cancellation.3 0. 16 QAM LTE 1x4 Receive Diversity Future Improvements Rel 1.5 1X2 Rx Div Rel 1.6 0.0 Spectral Efficiency (bps/Hz/sector) 0. Interference Cancellation Rel 1. Rysavy Research / 3G Americas.5 0.9 0.

0 EVRC 8 kbps UMTS/HSPA LTE CDMA2000 WiMAX Figure 23 shows UMTS Release 99 with both AMR 12. compared to Rel. UMTS has dynamic adaptation between vocoder rates.2 kbps Future Improvements Rel 1. static. B of 1xEV-DO. September 2009 Page 58 . Rysavy Research / 3G Americas. 10 + 10 MHz for UMTS-HSPA/LTE and CDMA2000.95 kbps Future Improvements Interference Cancellation EVRC-B 6 kbps EVRC-B 6 kbps EV-DO Rev A EVRC 8 kbps 1xRTT EVRC 8 kbps Erlangs. 10+10 MHz 400 350 300 250 200 150 100 50 Future Improvements Interference Cancellation AMR 5. It assumes a round-robin type of scheduler. which should increase spectral efficiency by 50%. Spectral efficiency gains. however. Initial systems will employ 1X2 receive diversity (two antennas at the base station) and later 1X4 diversity. but this improvement depends on factors such as the scheduling efficiency and the exact deployment scenario.2 kbps and 7. enabling enhanced voice quality compared to EVRC at the expense of capacity in situations that are not capacity limited. It is also possible to employ Multi-User MIMO (MU-MIMO).g.9 kbps Rel 7. With LTE. 79 Source: Joint analysis by 3G Americas’ members. VoIP AMR 7..95 kbps UMTS R’99 AMR 12. Figure 23: Comparison of Voice Spectral Efficiency79 500 450 Future Improvements LTE AMR 5. because it does not require an additional transmitter in the mobile device.9 kbps Rel 7 VoIP AMR 5. Mix of mobile and stationary users. easier to implement than true MIMO. in fact. with MU-MIMO are not as great as with the receive diversity schemes. as does Rev.9 kbps LTE VoIP AMR 7. OFDM-based systems can exhibit improved uplink capacity relative to CDMA technologies. HSPA to LTE-Advanced. 0. The AMR 12. as opposed to a proportional-fair scheduler that is normally used for asynchronous data. indoors) channel conditions.5 EVRC-B 6kbps Rel 1. spectral efficiency gains increase by use of receive diversity. A and Rev. and 20 MHz DL/UL=29:18 TDD for WiMAX.2 kbps vocoder provides superior voice quality in good (e.95 kbps UMTS R’99 AMR 7. Figure 23 compares voice spectral efficiency. which allows simultaneous transmission by multiple users on the uplink on the same physical resource to increase spectral efficiency and is.The implementation of HSUPA in HSPA significantly increases uplink capacity.95 kbps vocoders.

Mo-Han Fong and Robert Novak. voice capacity could double over existing circuit-switched systems. Rysavy Research / 3G Americas. No.5. will number in the tens of millions. There are a number of planned improvements for CDMA2000 1xRTT that will result in increased voice capacity. however. rather. The figures show performance assuming specific codecs at representative bit rates. providing an additional gain of 12%.0 and 210 Erlangs for Release 1. Vol 24. However. the difference in volume involved including subscribers and the amount of infrastructure required. however. EV-DO technologies could possibly exhibit a slightly higher spectral efficiency for VoIP than HSPA technologies (although not for packet data in general). may not be practical if the same carrier frequency must support both voice and data. since now 18 data symbols per frame are allocated for the UL compared to 12. however. This difference should translate to dramatically reduced costs for the highest volume solutions. If adjacent carriers are available. It should be noted. the bit rate shown is an average value. a variety of codecs can be used. Qi Bi. Based on projections and numbers already presented in this paper. namely.81 Changing this ratio. With respect to codecs. other benefits of VoIP are driving the migration to packet voice. Alternatively. gains relate to advances in radio techniques applied to the data channels. that the gains are not related specifically to the use of VoIP. A spectral efficiency gain of 50% is available by changing the DL:UL ratio from 29:18 to 23:24. Intel Corporation. See also IEEE Journal on Selected Areas in Communication. September 2009 Page 59 . whereas subscribers to emerging wireless technologies. however. There is a point of comparison. Huawei Technologies. in which the differences between the technologies diverge tremendously. See Figure 24 for details. Sean McBeath.” October. January. For codecs such as EVRC (Enhanced Variable Rate Codec).1. “Improved VoIP Capacity in Mobile WiMAX Systems Using Persistent Resource Allocation. voice and data may be placed on different carriers using different TDD ratios. as they operate purely in the packet domain and do not have circuit-switched control overhead. This is for the simple reason that they employ many of the same approaches. such as WiMAX. this paper has compared wireless technologies on the basis of technical capability and demonstrated that many of the different options have similar technical attributes. Volume and Market Comparison So far. Transmit Power Control (TPC) bits on the uplink Dedicated Physical Control Channel DPCCH in UMTS R’99. of being able to support simultaneous circuit-switched and packet-switched users on the same radio channel. 81 80 Source: IEEE Communications Magazine. specifically GSM-HSPA.80 Until VoIP over EV-DO becomes available. A further gain of 15% is available through the use of persistent scheduling and changing the DL:UL from 23:24 to 20:27. “An Analysis of VoIP Service Using 1 EV-DO Revision A System”.Opportunities will arise to improve voice capacity using VoIP over HSPA channels. Roshni Srinivasan. seven CDMA2000 carriers can be deployed in 10 MHz of spectrum. 3G subscribers on UMTS networks will number in the many hundreds of millions by the end of this decade. HSPA to LTE-Advanced. Depending on the specific enhancements implemented. in VoIP systems such as LTE and WiMAX. HSPA will have the significant advantage. Among these benefits are a consolidated IP core network for operators and sophisticated multimedia applications for users. 2008. 2006. Cost. WiMAX voice spectral efficiency is shown at 180 Erlangs for Release 1. Nortel Networks. Many of these same advances may also be applied to current circuit-switched modes.

4 B 957 M 4. it is highly likely that LTE will match WiMAX subscriptions and by the end of the decade. there doesn’t seem to be any inherent cost advantage—even on an equal-volume basis.6  Billion Total 649 M 3.5 2000 131.5 0.Figure 24: Relative Volume of Subscribers Across Wireless Technologies82 Subscriptions Millions 7. HSPA to LTE-Advanced. June 2009 supplied to 3G Americas. June 2009. September 2009 Page 60 .000  438 M 4.5 0.1 44.0 B 3. Source: Informa WiMAX and LTE projections. when parity between the two technologies is expected. Ovum states “By 2014. 82 83 84 Source: Informa Telecoms & Media.7 B 3.7 3.7 B 6. mobile WiMAX will have almost 55 million connections.8 B 3. July 2.5 957 37. In comparison. This is in stark contrast to 2013. WCIS+. LTE will have 109 million connections worldwide. And when factoring in the lower volumes.000  ‐ UMTS‐HSPA GSM CDMA Other Dec 09 Dec 10 Dec 11 Dec 12 Dec 13 Dec 14 In the chart above. any real-world cost advantage is debatable.5 2700 By mid next decade. the small “Other” category represents both WiMAX and LTE.4 B 2.9 B 4.0 438 7.1 1400 82.0 304 2.000  3.5 649 16.000  2. Rysavy Research / 3G Americas.0 B 2.”84 Although proponents for technologies such as mobile WiMAX point to lower costs for their alternatives.8 0. 2009. Informa projections on HSPA.000  1. Telecom and Software News. will significantly exceed them.1 13. Source: Ovum.7 B 2.000  6.000  5.3 Billion Total   Subscriptions in  2014 1. LTE and WiMAX in millions of subscribers are as follows:83 2008 2009 2010 2011 2012 2013 2014 WiMAX LTE HSPA 0.

backhaul. slower growth expected than GSM-HSPA Extremely mature 88 WiMAX 82 million anticipated by 2013 Maturity Emerging/immature 85 86 87 88 Source: 3G Americas’ member analysis. This cost.86 Even LTE is on the road to a robust wireless ecosystem and significant economies of scale. Peter Meissner. NTT DoComo. Telecom Italia. http://www.org/content/view/2479/172/. HSPA versus WiMAX) only applies to the software supported by the digital cards at the base station. Telenor. As for the rest of the network including construction. RF cards. Its operator members include: Alltel. T-Mobile and Vodafone. Source: CDG. power amplifiers. spectrum sold at 3. Telefonica. As for UMTS-HSPA versus CDMA2000. is only a small fraction of the base station cost with the balance covering antennas. Telus. AT&T. the Next Generation Mobile Networks (NGMN) alliance confirmed its selection of LTE. Spectrum costs for each technology can differ greatly depending on a country’s regulations and the spectrum band.umts-forum. cables. Royal KPN. research and development amortization results in a four-to-one difference in base station costs. For example. 4 billion expected by 2010 Extremely mature CDMA2000 455 million today. TeliaSonera. Operating Officer of NGMN announced that “based on intensive and detailed technology evaluations. China Mobile.From a deployment point of view. racks.5 GHz will cost much less than spectrum sold at 850 MHz (all other things being equal). France Telecom. and core-network components.85 Similarly. higher deployment—by a factor of five—could translate to significant cost savings. June 2009 for Q4 2008. however. MSV Mobile Satellite Ventures. As a general rule in most parts of the world. HSPA to LTE-Advanced. Developments such as single-chip UMTS complementary metal oxide semiconductor (CMOS) transceivers could be particularly effective in making UMTS/HSDPA devices more affordable to the mass market. September 2009 Page 61 . 2007. Table 8 summarizes the competitive position of the different technologies discussed. 29 vendor sponsors and 3 University research institutes.7 billion today. Dr. costs are similar regardless of Radio Access Network (RAN) technology. In June of 2008. just as GSM handsets are considered much less expensive than 1xRTT handsets. 3GPP LTE/SAE is the first technology which broadly meets its recommendations and is approved by its Board. Table 8: Competitive Position of Major Wireless Technologies Technology Subscribers EDGE/HSPA/LTE Over 3. Reliance Communications. Telstra. SK Telecom. Competitive Summary Based on the information presented in this paper. UMTS-HSPA wholesale terminal prices could be the market leader in low-cost or mass-market 3G terminals. the type of technology used (for example. Source: Qualcomm press release Feb 13. Rysavy Research / 3G Americas.”87 The NGMN is comprised of 18 mobile network operators.

5 2 to 4 Mbps average. Initial devices emphasize data Optimized OFDMA in Release 1. With the application of Dual Transfer Mode. Voice coverage will be much more limited than cellular Potentially available. matches OFDMA approaches in 5 MHz with HSPA+ Peak downlink user-achievable rates of over 4 Mbps today. smoothest migration to VoIP of any technology Many more cell sites required at 2.5 GHz.0.5 Mbps.0. Broad selection of 1xRTT/EV-DO devices Highly optimized CDMA for Rev 0/A/B. but not higher than HSPA+ for Release 1. and not higher than LTE for Release 1. highly optimized CDMA for HSPA. with significantly higher rates in the future Extremely efficient circuit-voice available today. with significantly higher rates in the future Extremely efficient circuit-voice available today EV-DO radio channels with VoIP cannot support circuit-voice users Not available today Available with VoIP Voice Capability Relatively inefficient VoIP initially. June 2009. Rysavy Research / 3G Americas.5 Devices Radio Technology Spectral Efficiency Very high. Broad selection of GSM/EDGE/UMTS/ HSPA devices Highly optimized TDMA for EDGE. Global with the general exception of Western Europe Fewer cell sites required at 700 and 850 MHz. HSPA to LTE-Advanced. 12 Mbps peak89 Throughput Capabilities Peak downlink user-achievable rates of over 1. highly optimized OFDMA for LTE Very high with HSPA. though initial services will emphasize data Simultaneous Voice and Data Available with GSM90 and UMTS today 89 90 Sprint web page. highly optimized OFDMA for Rev C Very high with EVDO Rev A/B WiMAX Limited to date Very limited Deployment Fewer cell sites required at 700 and 850 MHz.Technology Adoption Coverage/Footprint EDGE/HSPA/LTE Cellular operators globally Global CDMA2000 Cellular operators globally. September 2009 Page 62 . more efficient in later stages. but lower than LTE. More optimized in Release 1.

and increased revenues. HSPA offers the highest peak data rates of any widely available. the EDGE/HSPA/LTE family of technologies has proven itself as the predominant wireless network solution and offers operators and subscribers a true mobile-broadband advantage. Benefits include the ability to roam globally. Unlike some competing technologies. It achieves high spectral efficiency and data performance that today supports a wide range of applications. spectral efficiency will improve. EDGE has proven to be a remarkably effective and efficient technology for GSM networks. the advantages offer a best-of-breed. greater network usage. It also allows operators to support voice and data across their entire available spectrum. long-term solution that matches or exceeds the performance of competing approaches. and call control. and an astonishing variety of competitive handsets and other devices. the different radio-access technologies can coexist using the same core architecture. as well as halving latency and increasing spectral efficiency—making the technology viable for many years to come. The result is support for more users with more supported applications. huge economies of scale. as well as an effective technology solution for greenfield operators. Evolved EDGE will greatly enhance EDGE capabilities—doubling and. Many networks are now being upgraded to include HSUPA providing users with uplink rates in excess of 1 Mbps. UMTS today offers users simultaneous voice and data. HSPA to LTE-Advanced. UMTS-HSPA and/or LTE offer an excellent migration path for GSM operators. quadrupling throughputs. the migration path from EDGE to HSPA then to LTE is proving inevitable. video. potentially. Greater efficiencies and capabilities translate to more competitive offers. complementary services such as messaging and multimedia. The scope of applications will also increase as new services through standardized network interfaces become available such as location information. Currently more than 264 commercial UMTS-HSPA networks are already in operation. Whereas EDGE is efficient for narrowband data services. the UMTS-HSPA radio link is efficient for wideband services. the technologies’ advantages provide for broadband services that will deliver increased data revenue and provide a path to all-IP architectures. Because of practical benefits and deployment momentum. With LTE. With continued evolution. wide-area wireless technology. widespread acceptance by operators. and latency will decrease. With UMTS-HSPA. Today.Technology Efficient Spectrum Usage EDGE/HSPA/LTE Entire UMTS radio channel available for any mix of voice and highspeed data CDMA2000 Radio channel today limited to either voice/medium speed data or highspeed data only WiMAX Currently only efficient for datacentric networks Conclusion Through constant innovation. HSPA has significantly enhanced UMTS by providing a broadband data service with userachievable rates that often exceed 1 Mbps on the downlink in initial deployments and that now exceed 4 Mbps in some commercial networks. The continued use of GSM and EDGE technology through ongoing enhancements allows operators to leverage existing investments. In all cases. now the most widely chosen technology platform for the forthcoming decade and with deployment imminent. Rysavy Research / 3G Americas. peak data rates will continue to increase. September 2009 Page 63 .

Rysavy Research / 3G Americas. and allow a common core network to support both LTE and legacy GSM-HSPA systems. multimedia messaging. with peak rates of 42 Mbps. and location services. wireless data has slowly. September 2009 Page 64 . powerful new handheld-computing platforms. 3GPP adopted OFDMA with 3GPP LTE. EDGE/HSPA/LTE provides one of the most robust portfolios of mobile-broadband technologies.Not only expected continual improvements in radio technology. Other innovations. With the continued growth in mobile computing. will largely match the throughput and capacity of OFDMA-based approaches in 5 MHz. enable all services in the IP domain. HSPA to LTE-Advanced. and it is an optimum framework for realizing the potential of this market. Evolved HSPA+ systems. which will provide a growth platform for the next decade. speed applications. but inevitably. such as MIMO and higher order modulation are now being deployed. an increasing amount of mobile content. simplify deployment. but improvements to the core network through flatter architectures—particularly EPC/SAE—will reduce latency. mobile commerce. become a huge industry.

In these cases.Appendix: Technology Details The EDGE/HSPA/LTE family of data technologies provides ever-increasing capabilities that support ever more demanding applications. therefore. HSUPA). but as critical for effective application performance. reduced latency. These are currently available at a rudimentary level (e.g. With HSPA. Under favorable conditions. devices will be multi-modal supporting multiple types of wireless technologies. Wireless Application Protocol [WAP]): 8 to 128 kbps Multimedia messaging: 8 to 64 kbps Video telephony: 64 to 384 kbps General-purpose Web browsing: 32 kbps to more than 1 Mbps Enterprise applications including e-mail. are the needs for low latency. in particular. It is also helpful to specifically note the throughput requirements necessary for different applications:      Microbrowsing (for example. and efficient access to their enterprise applications. QoS control. be granted quite different levels of connectivity ranging from a dense urban environment where they may obtain the latest wireless technology to slower speeds in a rural network deployment or when roaming in a visited network. Latency has continued to improve. LTE. 3G). Rysavy Research / 3G Americas. HSPA to LTE-Advanced. With UMTS and HSPA.. The obvious needs are broad coverage and high data throughput. consumer applications. It is important to understand the needs enterprises and consumers have for these services. high-fidelity music. e-mail. we consider different technical approaches for wireless and the parallel evolution of 3GPP technologies. and increased spectral efficiency. EDGE. and spectral efficiency. September 2009 Page 65 . Web browsing. is of paramount concern. UMTS-HSPA. applications operate faster and the range of supported applications expands even further. richer multimedia applications. We then provide details on EDGE. In this section. enhanced security. Evolved EDGE vs. improved QoS. EDGE delivers peak user-achievable throughput rates close to 200 kbps and initial deployments of HSPA deliver peak user-achievable downlink throughput rates of well over 1 Mbps. users are enjoying videophones.g. The combination of low latency and high throughput translates to a broadband experience for users in which applications are extremely responsive. and VPNs: 32 kbps to more than 1 Mbps Video and audio streaming: 32 kbps to 2 Mbps  Note that EDGE already satisfies the demands of many applications. database access. with HSPA networks today having round-trip times as low as 70 msec. too. details how the progression from EDGE to HSPA to LTE is one of increased throughput. but future improvements will enable display of additional details (e. Users equipped with such multimode devices may. and supporting technologies such as IMS. users will benefit from knowing what service level to expect such as from indications on the device screen.. Spectral efficiency. 2G vs. Increasingly. The discussion below. HSPA+. Less obvious for users. messaging. already makes a wealth of applications feasible including enterprise applications. now available globally. easily meeting the demands of many applications. because it translates to higher average throughputs (and thus more responsive applications) for more active users in a coverage area. and even some multimedia applications. EDGE. which examines each technology individually.

2. GPRS provides a packet-based IP connectivity solution supporting a wide range of enterprise and consumer applications. 91 “Peak user-achievable” means users. under favorable conditions of network loading and signal propagation. Average rates depend on many factors and will be lower than these rates. It is an enhancement applicable to GPRS. The SGSN forwards/receives user data to/from the GGSN.. and it is often collocated with the MSC. Internet) EDGE is essentially the addition of a packet-data infrastructure to GSM. Figure 25: GSM/GPRS/EDGE Architecture Mobile Station Mobile Station Mobile Station Base Transceiver Station Base Transceiver Station Circuit-Switched Traffic Base Mobile Station Switching Controller Center IP Traffic Home Location Register Public Switched Telephone Network GPRS/EDGE Data Infrastructure Serving GPRS Support Node Gateway GPRS Support Node External Data Network (e. The term GPRS may also be used to refer to the initial radio interface. users have the same effective access speed as a modem. The SGSN performs the types of functions for data that the MSC performs for voice. GSM networks with EDGE operate as wireless extensions to the Internet and give users Internet access. HSPA to LTE-Advanced. now supplanted by EDGE. as depicted in Figure 25. which is the original packet data service for GSM networks. can achieve this rate as measured by applications such as file transfer. the Internet). beginning with the architecture of GSM and EDGE. this same data architecture is preserved in UMTS and HSPA networks. The GGSN also manages IP addresses. Each serving area has one SGSN. To understand the evolution of data capability. With peak user-achievable91 throughput rates of up to 200 kbps with EDGE using four timeslot devices. September 2009 Page 66 . The base station controller directs/receives packet data to/from the SGSN.g. dynamically assigning them to mobile stations for their data sessions. which can be viewed as a mobile IP router to external IP networks. Typically. as well as to GSM circuitswitched services. but with the convenience of connecting from anywhere. the latter not being considered further in this document. Functions of the data elements are as follows: 1. most GSM networks support EDGE. and it is technically referred to as GPRS for the core-data function in all these networks. In fact. 3. an element that authenticates and tracks the location of mobile stations. Rysavy Research / 3G Americas.EDGE/EGPRS Today. there is one GGSN per external network (for example. we briefly examine how these data services operate. as well as access to their organizations from anywhere.

Example: WAP notification message delivered via SMS. In the radio link. The network can have multiple radio channels (referred to as transceivers) operating in each cell sector. circuit-switched functions like voice calls or data calls. each data timeslot can deliver peak user-achievable data rates of up to about 50 kbps. The network assigns different functions to each timeslot such as the Broadcast Control Channel (BCCH). which stores users’ account information for both voice and data services. 92 93 Source: 3G Americas’ member company contribution. users can accept an incoming voice call. may alternate between frames for half-rate PDTCH: Packet Data Traffic Channel – carries packet data traffic for GPRS and EDGE PBCCH: Packet Broadcast Control Channel – additional signalling for GPRS/EDGE. an operator can accommodate customers by assigning an increasing number of channels for data service that is limited only by that operator’s total available spectrum and radio planning. Users can also receive SMS messages and data notifications93 while on a voice call. The network can aggregate up to four of these timeslots on the downlink with current devices. The network can dynamically adjust capacity between voice and data functions. and it can also reserve minimum resources for each service. Figure 26: Example of GSM/EDGE Timeslot Structure92 4. Rysavy Research / 3G Americas. and packet data channels. however. Of significance is that this same data architecture supports data services in GSM and in UMTS-HSPA networks. In most networks. September 2009 Page 67 . thereby maximizing overall use of the network. This enables more data traffic when voice traffic is low or. and then resume their data session automatically when the voice session ends. As demand for data services increases. users with DTM-capable devices can engage in simultaneous voice/data operation. may be set up on a timeslot of a TDMA frame when justified by the volume of data traffic. GSM uses radio channels of 200 kilohertz (kHz) width. used only if needed EDGE offers close coupling between voice and data services. For example. they share the available data channels. thereby simplifying operator network upgrades. With respect to data performance. HSPA to LTE-Advanced.6 msec. the PBCCH.615 ms per frame of 8 timeslots 577 S per timeslot 0 1 TCH 1 TCH 2 TCH 2 TCH 3 TCH 3 PDTCH 4 TCH 4 PDTCH 5 PDTCH 5 PDTCH 6 PDTCH 6 PDTCH 7 PDTCH 7 PDTCH Possible BCCH carrier configuration Possible TCH carrier configuration BCCH 0 PBCCH BCCH: Broadcast Control Channel – carries synchronization. If multiple data users are active in a sector. paging and other signalling information TCH: Traffic Channel – carries voice traffic data. likewise. With networks supporting DTM. which suspends the data session. as shown in Figure 26. while in a data session. divided in time into eight timeslots comprising 577 microseconds (s) that repeat every 4.Another important element is the HLR. which expands the capabilities of the normal BCCH. the optional Packet Broadcast Control Channel (PBCCH). more voice traffic when data traffic is low.

during a voice call. do multimedia messaging. for example when a LTE user HSPA to LTE-Advanced. For example. Many operators that originally planned to use only UMTS for next-generation data services have deployed EDGE as a complementary 3G technology. it is less costly than upgrading to UMTS. There are a number of networks and devices now supporting DTM. Also. A GPRS network using the EDGE radio interface is technically called an Enhanced GPRS (EGPRS) network. users will be able to retrieve e-mail. Typically.EDGE is an official 3G cellular technology that can be deployed within an operator's existing 850. and cable-modem speeds are realistically achievable. Operators also continue to make improvements in how EDGE functions including network optimizations that boost capacity and reduce latency. This is particularly useful when connecting phones to laptops via cable or Bluetooth and using them as modems. and do Internet conferencing. Evolved EDGE Recognizing the value of the huge installed base of GSM networks. which also includes voice enhancements not discussed in this paper. any application developed for GPRS will work with EDGE. it makes good sense to continue to evolve EDGE capabilities. It is fully backward-compatible with older GSM networks. which is already equipped for simple roaming and billing. September 2009 Page 68 . and a GSM network with EDGE capability is referred to as GERAN. Dual Transfer Mode (DTM) devices. EDGE capability is now largely standard in new GSM deployments. Release 4 significantly reduced EDGE latency (network round-trip time)—from the typical 500 to 600 msec to about 300 msec. For example. because it involves fewer long-term capital investments to upgrade an existing system. In addition. many regions do not have licensed spectrum for deployment of a new radio technology such as UMTS-HSPA or LTE. It is important to note that EDGE technology is continuing to improve. already available from vendors. Rysavy Research / 3G Americas. browse the Web. Release 7’s Evolved EDGE also introduces significant new features. Evolved EDGE offers higher data rates and system capacity. DTM is a 3GPP-specified technology that enables new applications like video sharing while providing a consistent service experience (service continuity) with UMTS. EDGE has been an inherent part of GSM specifications since Release 99. and reduced latency. 3GPP has worked to improve EDGE capabilities for Release 7. This work was part of the GERAN Evolution effort. meaning that GPRS devices work on EDGE networks and that GPRS and EDGE terminals can operate simultaneously on the same traffic channels. meaning that a user will not have a hugely different experience when moving between environments. and it is highly asset-efficient. Although EDGE today already serves many applications like wireless e-mail extremely well. From an economic standpoint. With 85 percent of the world market using GSM. The impact for users is that EDGE networks today are more robust with applications functioning more responsively. 900. Evolved EDGE provides better service continuity between EDGE and HSPA or LTE. and 1900 MHz spectrum bands. Devices themselves are increasing in capability. In addition. 1800. it is easy to offer global service to subscribers. a DTM end-to-end solution requires only a software upgrade to the GSM/EDGE radio network. allow simultaneous voice and data communications. because most enhancements are designed to be software based.

    The methods standardized in Release 7 to achieve or surpass these objectives include:     Dual-Carrier Receiver A key part of the evolution of EDGE is the utilization of more than one radio frequency carrier. EDGE uses 8-PSK modulation. 16 QAM and 32 QAM. The A/Gb mode interface is part of the 2G core network. These enhancements will have a dramatic effect on throughput for many applications. Simulations have demonstrated sensitivity gains of 3 dB and a decrease in required C/I of up to 18 dB for a single co-channel interferer. with voice operation provided by GSM. Downlink diversity reception of the same radio channel to increase the robustness in interference and to improve the receiver sensitivity. Downlink dual-carrier reception to double the number of timeslots that can be received for a 100 percent increase in throughput. HSPA to LTE-Advanced. The addition of Quadrature Phase Shift Keying (QPSK). Although GSM and EDGE are already highly optimized technologies.moves to a GSM/Evolved EDGE network to establish a (circuit-switched) voice call94 or when leaving LTE coverage. advances in radio techniques will enable further efficiencies. A reduction in overall latency. Using two radio-frequency carriers requires two receiver chains in the downlink. To avoid impacts on infrastructure by enabling improvements through a software upgrade. so this goal is required for full backward-compatibility with legacy GPRS/EDGE.2x) and a new set of modulation/coding schemes that will increase maximum throughput per timeslot by up to 100 percent (EGPRS2-B). Some of the objectives of Evolved EDGE included:     A 100 percent increase in peak data rates. A 50 percent increase in spectral efficiency and capacity in C/I-limited scenarios. Significant increases in system capacity can be achieved. To coexist with legacy mobile stations by allowing both old and new stations to share the same radio resources. A sensitivity increase in the downlink of 3 dB for voice and data. Rysavy Research / 3G Americas. thus facilitating deployment in existing networks. as well as an increased symbol rate (1. September 2009 Page 69 . To be applicable to DTM (simultaneous voice and data) and the A/Gb mode interface. Currently. as explained below. thereby enabling support for conversational services such as VoIP and PoC. 94 Some initial LTE networks will be data-only. as shown in Figure 27. This overcomes the inherent limitation of the narrow channel bandwidth of GSM. This is achieved by lowering the TTI to 10 msec and by including the acknowledgement information in the data packet. A reduction of latency for initial access and round-trip time. Using two carriers enables the reception of twice (or more than twice for some multi-slot classes) as many radio blocks simultaneously. To achieve compatibility with existing frequency planning.

As network loading increases. it is statistically unlikely that contiguous timeslots will be available. the original number of radio blocks can be divided between the two carriers.Figure 27: Evolved EDGE Two-Carrier Operation95 Slot N Rx1 Rx2 Tx (1) Slot N + 1 (Idle Frame) Slot N + 2 Slot N + 3 Neighbor Cell Measurements Uplink Timeslot Downlink Timeslot Alternatively. HSPA to LTE-Advanced. which increases trunking efficiency. Figure 28: EDGE Multi-Carrier Receive Logic – Mobile Part96 Channel capacity with dual-carrier reception improves greatly. With today’s EDGE devices. Rysavy Research / 3G Americas. but because of statistical improvement in the ability to assign radio resources. it is not possible to change radio frequencies when 95 96 Source: 3G Americas’ member company contribution. not by increasing basic efficiencies of the air interface. September 2009 Page 70 . This eliminates the need for the network to have contiguous timeslots on one frequency. Source: 3G Americas’ member company contribution.

Rysavy Research / 3G Americas. the probability of being able to obtain 5 timeslots on the same radio carrier decreases dramatically. this becomes possible. which otherwise is not possible. thus enabling contiguous timeslots across different radio channels. Figure 29 shows why this is important. the probability of being assigned 1 timeslot decreases.) The BCCH carrier repeats over 12 cells in a 4/12 frequency reuse pattern. however. With an Evolved EDGE dual receiver. As this probability decreases (X axis). HSPA to LTE-Advanced. 97 Source: 3G Americas’ member company contribution. (BCCH refers to the Broadcast Control Channel and TCH refers to the Traffic Channel. As the network becomes busy. A fractionally loaded system may repeat f12 through f15 on each of the cells.going from one timeslot to the next. but also potentially high co-channel interference in loaded conditions.4 MHz for GSM. The result is that the system can allocate a larger set of time slots for data even if they are not contiguous. significantly improves the likelihood of obtaining the desired timeslots. Being able to obtain timeslots across two carriers in Evolved EDGE. which requires 2. Figure 29: Probabilities of Time Slot Assignments97 Mobile Station Receive Diversity Figure 30 illustrates how mobile-station receive diversity increases system capacity. This is a 1/1 frequency reuse pattern with higher system utilization. September 2009 Page 71 . however.

Newer techniques for demodulation. help enable their use. however. it is possible to increase the load on the 1/1 layer from 25 to 50 percent and possibly to as high as 75 percent. Rysavy Research / 3G Americas. Two different levels of support for higher order modulation are defined for both the uplink and the downlink: EGPRS2-A and EGPRS2-B. variability of interference. however. In today’s EDGE systems. and 16 QAM at the legacy symbol rate. thereby taking advantage of the geographical locations associated with probabilities of high C/I ratio and enabling very high data transfer rates whenever possible. In the uplink. This level of support reuses Modulation and Coding Schemes (MCSs) 1 through 6 from EGPRS and adds five new 16 QAM modulated schemes called uplink EGPRS2-A schemes (UAS). because factors such as processing power.f11 are associated with TCH carrier frequencies f12 – f15 Example of a 4/12 frequency reuse pattern used for BCCH carriers with a 1/1 frequency reuse pattern for TCH carriers. Using Evolved EDGE and receive-diversity-enabled mobile devices that have a high tolerance to co-channel interference. More bits-per-symbol means more data transmitted per unit time. EGPRS2-A level includes GMSK. 8-PSK. Use of higher order modulation exploits localized optimal coverage circumstances. September 2009 Page 72 . such as advanced receivers and receive diversity.Figure 30: Example of 4/12 Frequency Reuse with 1/1 Overlay98 f15 f14 f13 f12 f15 f14 f13 f12 f15 f14 f13 f12 0 9 10 11 6 8 2 1 3 7 5 4 BCCH carriers on f0 . An increase to 50 percent translates to a doubling of capacity on the 1/1 layer without requiring any new spectrum and to a 200 percent gain compared to a 4/12 reuse layer. These enhancements are only now being considered. Thus. 98 Source: 3G Americas’ member company contribution. with four of these frequencies. This yields a fundamental technological improvement in information capacity and faster data rates. it is possible to obtain 100 percent of the capacity of the frequencies in the 4/12 reuse layer or to double the capacity by adding 800 KHz of spectrum. and signal level made higher order modulations impractical for mobile wireless systems just a few years ago. Higher Order Modulation and Higher Symbol Rate Schemes The addition of higher order modulation schemes enhances EDGE network capacity with little capital investment by extending the range of the existing wireless technology. HSPA to LTE-Advanced. f12 through f15 in the 1/1 reuse layer can only be loaded to around 25 percent of capacity.

and eight new uplink EGPRS2-B schemes (UBS) are added. MCSs 1 through 4 from EGPRS are reused.2 70.2 435.6 118.2 44. Table 10: Uplink Modulation and Coding Schemes with Higher Symbol Rate Modulation and Coding Scheme Name MCS-1 MCS-2 MCS-3 MCS-4 UBS-5 UBS-6 UBS-7 UBS-8 UBS-9 UBS-10 UBS-11 UBS-12 Uplink EGPRS2 Support Level B Modulation Type GMSK GMSK GMSK GMSK QPSK QPSK 16 QAM 16 QAM 16 QAM 32 QAM 32 QAM 32 QAM Peak Throughput (kbps) – 4 slots 35.Table 9: Uplink Modulation and Coding Schemes Modulation and Coding Scheme Name MCS-1 MCS-2 MCS-3 MCS-4 MCS-5 MCS-6 UAS-7 UAS-8 UAS-9 UAS-10 UAS-11 Uplink EGPRS2 Support Level A Modulation Type GMSK GMSK GMSK GMSK 8-PSK 8-PSK 16 QAM 16 QAM 16 QAM 16 QAM 16 QAM Peak Throughput (kbps) – 4 slots 35.2 The second support level in the uplink includes QPSK.4 179.8 355.6 The first downlink support level introduces a modified set of 8-PSK coding schemes and adds 16 QAM.2 236.2x) symbol rate.2 473. and 32 QAM all at the legacy symbol rate. Turbo codes are used for all HSPA to LTE-Advanced.8 307.2 70.4 179.8 268. Rysavy Research / 3G Americas.2 44. and 32 QAM modulation as well as a higher (1.4 89.4 89.8 59.8 236. 16 QAM.8 59.6 118.8 268.2 204. September 2009 Page 73 .

6 262. Table 11: Downlink Modulation and Coding Schemes Modulation and Coding Scheme Name MCS-1 MCS-2 MCS-3 MCS-4 DAS-5 DAS-6 DAS-7 DAS-8 DAS-9 DAS-10 DAS-11 DAS-12 Downlink HOM/HSR Support Level A Modulation Type GMSK GMSK GMSK GMSK 8-PSK 8-PSK 8-PSK 16 QAM 16 QAM 32 QAM 32 QAM 32 QAM Peak Throughput (kbps) – 4 slots 35. Rysavy Research / 3G Americas.2x) symbol rate.6 108. MCSs 1 through 4 are reused. and eight new downlink EGPRS2-B level schemes (DBS) are defined. 16 QAM.4 393.2 44. September 2009 Page 74 .new modulations.2 179.4 89. and 32 QAM modulations at a higher (1.4 326.2 70.8 59.2 217.8 131.6 The second downlink support level includes QPSK. MCSs 1 through 4 are reused. and eight new downlink EGPRS2-A level schemes (DAS) are added. HSPA to LTE-Advanced.

HSPA to LTE-Advanced.2 435.4 kbps100 This translates to a peak network rate close to 2 Mbps and a user-achievable data rate of well over 1 Mbps! Evolved EDGE Implementation Table 13 shows what is involved in implementing the different features defined for Evolved EDGE.2 236.2 44.2 70.4 kbps Timeslots per carrier = 8 Carriers used in the downlink = 2 Total downlink data rate = 118. Evolved EDGE is compatible with legacy frequency planning. a Type 2 mobile device (one that can support simultaneous transmission and reception) using DBS-12 as the MCS and a dual-carrier receiver can achieve the following performance: Highest data rate per timeslot (layer 2) = 118.Table 12: Downlink Modulation and Coding Schemes with Higher Symbol Rate99 Modulation and Coding Scheme Name MCS-1 MCS-2 MCS-3 MCS-4 DBS-5 DBS-6 DBS-7 DBS-8 DBS-9 DBS-10 DBS-11 DBS-12 Downlink HOM/HSR Support Level B Modulation Type GMSK GMSK GMSK GMSK QPSK QPSK 16 QAM 16 QAM 16 QAM 32 QAM 32 QAM 32 QAM Peak Throughput (kbps) – 4 slots 35.6 The combination of Release 7 Evolved EDGE enhancements shows a dramatic potential increase in throughput. For all features. For a number of features.8 59. Rysavy Research / 3G Americas. September 2009 Page 75 .6 118.4 179. 99 These data rates require a wide-pulse shaping filter that is not part of Release 7. 100 For the near future.4 kbps X 8 X 2 = 1894. in the downlink. For example.8 268.4 89. two carriers will be a scenario more practically realized on a notebook computer platform than handheld platforms.2 473.8 355. there are no hardware changes required for the base transceiver station (BTS).

and it gives operators maximum flexibility in providing different services across their coverage areas (see Figure 31). UMTS technology is mature and benefits from research and development that began in the early 1990s. WCDMA.102 Compared to emerging wireless technologies. support for high-bandwidth data applications. Operators can also use their entire available spectrum for both voice and highspeed data services. Additionally. Rysavy Research / 3G Americas. September 2009 Page 76 . With the addition of HSPA for high-speed packet data services. UMTS employs a wideband CDMA radio-access technology. UMTS-HSPA is quickly emerging as the dominant global mobile-broadband network. and planned for. It has been thoroughly trialed. and commercially deployed. by GSM. UMTS deployment is now accelerating with stable network infrastructures and attractive. and evolutions of these technologies. tested. Source: Informa Telecoms & Media. UMTS-HSPA Technology UMTS has garnered the overwhelming majority of new 3G spectrum licenses with 283 commercial networks already in operation. high user densities that can be supported with low infrastructure costs. HSPA. 101 102 Source: 3G Americas’ member company contribution. and a clean migration to VoIP in the future.Table 13: Evolved EDGE Implementation101   In conclusion. “World Cellular Information Service. HSPA to LTE-Advanced. This is called the UMTS multi-radio network. it is interesting to note the sophistication and capability that is achievable with. simultaneous voice and data capability for users. operators can use a common core network that supports multiple radioaccess networks including GSM. reliable mobile devices that have rich capabilities. EDGE. The primary benefits of UMTS include high spectral efficiency for voice and data.” June 2009.

or code space. Streaming. based on time division in combination with frequency hopping. spreadspectrum system. This wideband technology approach results in the flexibility to manage multiple traffic types including voice. HSPA to LTE-Advanced. HLR.Figure 31: UMTS Multi-radio Network GSM/EDGE UMTS Core Network (MSC. of each channel every 10 msec with WCDMA Release 99 and every 2 msec with HSPA. Real-time interactive data with controlled bandwidth and minimum delay such as VoIP or video conferencing. When both GSM and WCDMA access networks are available. The RNCs connect to the core network as do the BSCs. WLAN Radio-Access Networks Other Cellular Operators External Networks The UMTS radio-access network consists of base stations referred to as Node B (corresponding to GSM base transceiver systems) that connect to RNCs (corresponding to GSM base station controllers [BSCs]). Conversational. WCDMA creates high-bandwidth traffic channels by reducing the amount of spreading (using a shorter code) with WCDMA Release 99 and higher-order modulation schemes for HSPA. GGSN) Packet-Switched Networks WCDMA. narrowband data. as well as in areas in which the operator has continuous GSM coverage. 103 Spread spectrum systems can either be direct sequence or frequency hopping. but has only deployed WCDMA in some locations. or the network can assign dedicated channels to users. Continuous data with controlled bandwidth and some delay such as music or video. UMTS employs a sophisticated QoS architecture for data that provides four fundamental traffic classes including: 1. WCDMA allocates different codes for different channels. and wideband data. September 2009 Page 77 . but it is the wideband nature of WCDMA that provides its greatest advantage—the ability to translate the available spectrum into high data rates. Packet data users can share the same codes as other users. Rysavy Research / 3G Americas. To further expand the number of effectively operating applications.. Whereas GSM can effectively operate like a spread-spectrum system103. WCDMA is spectrally more efficient than GSM.g. This is important for managing capacity. 2. HSDPA Circuit-Switched Networks Other e. whether for voice or data. the network can hand over users between these networks. SGSN. and it can adjust the amount of capacity. WCDMA is a direct-sequence.

packet-data service that delivers peak theoretical rates of 14 Mbps. This capability is essential for expanding the scope of supported applications. while the next section covers HSUPA. these services become much more efficient and more powerful with HSPA. Interactive. with about 100 to 200 msec measured in actual networks. For example. In UMTS Release 99. In advance of Release 6. Peak user-achievable throughput rates in 104 Initial UMTS networks had peak uplink rates of 64 kbps or 128 kbps. This section covers technical aspects of HSDPA.104 This satisfies many communications-oriented applications. Although UMTS Release 99 offers attractive data services. the network uses HSPA (HSDPA/HSUPA) for data. which included voice and data capabilities. is a high-performance. Back-and-forth data without bandwidth control and some delay such as Web browsing. WCDMA has lower network latency than EDGE. and the interfaces to external networks such as the Internet. whereas a 384 kbps data channel uses a downlink spreading factor of 8. the capabilities of devices and the number of users active in the network limit the peak throughput rates a user can achieve to about 350 kbps in commercial networks. applications can negotiate QoS parameters on an end-to-end basis between a mobile terminal and a fixed-end system across the Internet or private intranets. each with a spreading factor of 4. HSDPA HSPA refers to networks that support both HSDPA and HSUPA. the data stream has greater redundancy. involves negotiation and prioritization of traffic in the radio-access network. With more spreading. Background. The commonly quoted rate of more than 2 Mbps downlink throughput for UMTS can be achieved by combining three data channels of 768 kbps. but many deployments emphasize 384 kbps. Consequently. Although exact throughput depends on the channel sizes the operator chooses to make available. Voice channels use downlink spreading factors of 128 or 256. Release 5 has defined HSDPA and Release 6 has defined HSUPA. HSDPA. in 2008. 4. This QoS architecture. By the end of 2008. specified in 3GPP Release 5.3. and the operator can employ more channels. Lower priority data that is non-real-time such as batch transfers. Peak downlink network speeds are 384 kbps. the uplink in HSDPA (Release 5) networks uses the Release 99 approach. HSPA to LTE-Advanced. Rysavy Research / 3G Americas. Since then. UMTS Release 99 Data Capabilities Initial UMTS network deployments were based on 3GPP Release 99 specifications. AT&T upgraded most of its network to HSPA. particularly multimedia applications including packetized video telephony and VoIP. September 2009 Page 78 . In comparison. All new deployments today are HSPA. Channel throughputs are determined by the amount of channel spreading. and many operators have upgraded their HSDPA networks to HSPA. the maximum theoretical downlink rate is just over 2 Mbps. available through all HSPA versions. the core network. a high-speed data channel has less spreading and fewer available channels. HSPA was deployed throughout the Americas. With HSPA-capable devices. as in voice channels. Uplink peak-network throughput rates are also 384 kbps in newer deployments with user-achievable peak rates of 350 kbps. Operators with Release 99 networks are upgrading them to HSPA capability.

The network can then readjust how users are assigned to different HS-PDSCH every 2 msec. Up to 15 of these channels can operate in the 5 MHz WCDMA radio channel. The can simultaneously service UMTS voice and data users. HSPA to LTE-Advanced. HSDPA achieves its performance gains from the following radio features:       High-speed channels shared in both code and time domains Short TTI Fast scheduling and user diversity Higher order modulation Fast link adaptation Fast HARQ These features function as follows: High-Speed Shared Channels and Short Transmission Time Interval: First. and soft combining.initial deployments same radio carrier HSDPA data users. HSDPA uses high-speed data channels called High Speed Physical Downlink Shared Channels (HS-PDSCH). September 2009 Page 79 . as well as through the addition of powerful new techniques such as fast scheduling. The higher spectral efficiency and higher data rates not only enable new classes of applications. but also support a greater number of users accessing the network. Figure 32 illustrates different users obtaining different radio resources. HSDPA achieves its high speeds through techniques similar to those that push EDGE performance past GPRS including higher order modulation. as well as HSDPA also has significantly lower latency measured today on some 70 msec on the data channel. variable coding. Each uses a fixed spreading factor of 16. The result is that resources are assigned in both time (the TTI interval) and code domains (the HS-PDSCH channels). User transmissions are assigned to one or more of these channels for a short TTI of 2 msec. Rysavy Research / 3G Americas. networks as low as are well over 1 Mbps and as high as 4 Mbps in some networks.

the network achieves significant user diversity and significantly higher spectral efficiency. rather than in a round-robin fashion. HSPA to LTE-Advanced. The system also makes sure that each user receives a minimum level of throughput. Because channel conditions vary somewhat randomly across users. This approach is sometimes called proportional fair scheduling.Figure 32: High Speed–Downlink Shared Channels (Example) User 1 User 2 User 3 User 4 Channelization Codes 2 msec Time Fast Scheduling and User Diversity: Fast scheduling exploits the short TTI by assigning users channels that have the best instantaneous channel conditions. most users can be serviced with optimum radio conditions and thereby obtain optimum data throughput. With about 30 users active in a sector. Rysavy Research / 3G Americas. Figure 33 shows how a scheduler might choose between two users based on their varying radio conditions to emphasize the user with better instantaneous signal quality. September 2009 Page 80 .

reduces delays and further improves overall data throughput. Fast Link Adaptation: Depending on the condition of the radio channel.8 Mbps peak-rate devices. and “hybrid” refers to a process of combining repeated data transmissions with prior transmissions to increase the likelihood of successful decoding. and increased productivity. a greater range of applications that function well. This is done in close coordination with fast scheduling. Rysavy Research / 3G Americas. faster application performance. For example. and the combination of MIMO and 64 QAM became available this year in Release 8. September 2009 Page 81 . Data throughput is increased with 16 QAM. vendors measured consistent throughput rates in actual deployments of more than 1 Mbps. as described above. Field results validate the theoretical throughput results. Managing and responding to real-time radio variations at the base station. this translates to better network performance under loaded conditions. Using the approaches just described. as opposed to the BSC in GPRS/EDGE. an advanced modulation scheme—16 QAM. HSPA Evolution will add 64 QAM modulation to further increase throughput rates. Note that 64 QAM was available in Release 7. Fast Hybrid Automatic Repeat Request: Another HSDPA technique is Fast Hybrid Automatic Repeat Request (Fast Hybrid ARQ). With initial 1. The process of selecting and quickly updating the optimum modulation and coding rate is referred to as fast link adaptation. HSDPA maximizes data throughputs and capacity and minimizes delays. as opposed to an internal network node. “Fast” refers to the medium-access control mechanisms implemented in Node B (along with scheduling and link adaptation). under good radio conditions. while QPSK is available under adverse conditions. These rates rose to more than 2 Mbps for 3. For users. The benefit of 16 QAM is that 4 bits of data are transmitted in each radio symbol as opposed to 2 bits with QPSK.Figure 33: User Diversity User 1 Signal Quality High data rate User 2 Low data rate Time User 2 User 1 User 2 User 1 User 2 User 1 Higher Order Modulation: HSDPA uses both the modulation used in WCDMA—namely QPSK—and. different levels of forward-error correction (channel coding) can also be employed. a three-quarter coding rate means that three quarters of the bits transmitted are user bits and one quarter are error-correcting bits.6 Mbps devices and are close HSPA to LTE-Advanced.

as low as 2 msec. HSUPA can achieve different throughput rates based on various parameters including the number of codes used. Many operator networks support 7. the spreading factor of the codes. backhaul) can support the high throughput rates. The improved uplink mechanisms also translate to better coverage and. and Fast Hybrid ARQ also serves to reduce latency. typical devices supporting peak data rates of 3.2 Mbps became available. HSUPA Whereas HSDPA optimizes downlink performance. In 2008. and increased spectral efficiency. Such an improved uplink benefits users in a number of ways. which can benefit many applications as much as improved throughput. These improvements include higher throughputs.to 4 Mbps for 7. Among these technologies are mobile-receive diversity and equalization (for example. HSUPA—which uses the Enhanced Dedicated Channel (E-DCH)—constitutes a set of improvements that optimizes uplink performance. fast scheduling.6 Mbps or 7. This improvement enables not only higher peak HSDPA throughput speeds. by extension and for simplicity. HSUPA is standardized in Release 6. but makes these speeds available over a greater percentage of the coverage area.2 Mbps peak operation. MMSE).2 Mbps devices. It results in an approximately 85 percent increase in overall cell throughput on the uplink and more than 50 percent gain in user throughput. For instance. which allows the base station to efficiently allocate radio resources Fast Hybrid ARQ.4 Mbps. HSPA to LTE-Advanced. although it is likely that most networks will use the two approaches together. HSUPA can operate with or without HSDPA in the downlink. Networks and devices supporting HSUPA became available in 2007. the primary uplink traffic channel defined for HSUPA is a dedicated channel that could be used for services delivered through either the circuit-switched or the packetswitched domains. some user applications transmit large amounts of data from the mobile station such as sending video clips or large presentation files. Rysavy Research / 3G Americas. the TTI value. HSPA technology is not standing still. which improve the quality of the received radio signal prior to demodulation and decoding. and the transport block size in bytes. which improves the efficiency of error processing   The combination of TTI. September 2009 Page 82 . the WCDMA-enhanced uplink capabilities are often identified in the literature as HSUPA. larger cell sizes. a significant benefit resulting in much improved application performance on HSPA networks Although the primary downlink traffic channel supporting HSDPA serves is a shared channel designed for the support of services delivered through the packet-switched domain. HSUPA also reduces packet delays. and some even support the maximum rate of 14. which allows faster responses to changing radio conditions and error conditions Fast Node B-based scheduling. reduced latency. Nevertheless. Advanced radio technologies are becoming available. improvements will balance the capacity of the uplink with the capacity of the downlink. assuming other portions of the network (for example. HSUPA achieves its performance gains through the following approaches:   An enhanced dedicated physical channel A short TTI. For future applications like VoIP. for rural deployments.

and Type 3. which results in an almost doubling of downlink capacity when employed in conjunction with techniques such as channel equalization. And with a later introduction of a 2 msec TTI. The goal in evolving HSPA is to exploit available radio technologies—largely enabled by increases in digital signal processing power—to maximize CDMA-based radio performance. are specified in 3GPP Release 6. Wireless and networking technologists have defined a series of enhancements for HSPA. latency will fall below 50 msec.Initial devices enabled peak user rates of close to 2 Mbps as measured in actual network deployments. This problem can be solved by advanced-receiver architectures with channel equalizers that yield additional capacity gains over HSDPA with receive diversity. relative to current HSDPA networks at 70 msec. Hence. some of which are specified in Release 7 and some of which are being finalized in Release 8. Different vendors are emphasizing different approaches. although only with the addition of interference cancellation methods that boost SNR. September 2009 Page 83 . Continuous Packet Connectivity. because it results in a large capacity gain independent of the radio channel. Future devices will ultimately approach speeds close to 5 Mbps. Moreover. MIMO. Higher-Order Modulation and One Tunnel Architecture. which includes a combination of receive diversity and channel equalization. which uses channel equalization. These designs include Type 1. will employ interference cancellation. Beyond throughput enhancements. however. latency will be as low as 30 msec. This not only makes HSPA competitive. Note that the different types of receivers are releaseindependent. What makes such enhancements attractive is that the networks do not require any changes other than increased capacity within the infrastructure to support the higher bandwidth. As already discussed in this paper. HSUPA also significantly reduces latency. The performance requirements for advanced-receiver architectures. The antenna spacing yields signals that have somewhat independent fading characteristics. CDMA approaches can match OFDMA approaches in reduced channel bandwidths. But at higher speeds. which uses mobile-receive diversity. These include advanced receivers. it significantly extends the life of sizeable operator infrastructure investments. the network can support a combination of devices including both HSPA to LTE-Advanced. however. the combination of reduced symbol period and multipath interference results in inter-symbol interference and diminishes rake receiver performance. One important area is advanced receivers for which 3GPP has specified a number of designs. which are not yet available. Advanced Receivers. Type 3i receivers will work and provide a capacity gain in a Release 5 network. The combination of mobilereceive diversity and channel equalization (Type 3) is especially attractive. the combined signal can be more effectively decoded. Rysavy Research / 3G Americas. Receive diversity is effective even for small devices such as PC Card modems and smartphones. The first approach is mobile-receive diversity. For example. Current receiver architectures based on rake receivers are effective for speeds up to a few megabits per second. Type 2. Type 3i devices. Evolution of HSPA (HSPA+) OFDMA systems have attracted considerable attention through technologies such as 3GPP LTE and WiMAX. This technique relies on the optimal combination of received signals from separate receiving antennas. In optimized networks. Alternate advanced-receiver approaches include interference cancellation and generalized rake receivers (G-Rake).

The transmitter sends different data streams over each antenna.earlier devices that do not include these enhancements and later devices that do. Spatial multiplexing MIMO should also benefit HSPA “hotspots” serving local areas such as airports. campuses. MIMO Another standardized capability is MIMO.105 Although MIMO can significantly improve peak rates. often in combination with multiple radios and multiple parallel data streams. This results in multiple data paths effectively operating somewhat in parallel and. 3GPP has standardized spatial multiplexing MIMO in Release 7 using Double Transmit Adaptive Array (D-TxAA). In a fully loaded network with interference from adjacent cells. and malls. MIMO—as illustrated in Figure 34—actually exploits multipath. however. a technique that employs multiple transmit antennas and multiple receive antennas. Device vendors can selectively apply these enhancements to their higher performing devices. where the technology will increase capacity and peak data rates. The most common use of the term “MIMO” applies to spatial multiplexing. relying on signals to travel across different uncorrelated communications paths. in a multiplicative gain in throughput. September 2009 Page 84 . other techniques such as Space Division Multiple Access (SDMA)—also a form of MIMO—may be even more effective 105 For further details on these techniques. Relative to a 1x1 antenna system. through appropriate decoding. 2X2 MIMO can deliver cell throughput gains of about 80 percent. overall capacity gains will be more modest—in the range of 20 to 33 percent over mobile-receive diversity. Rysavy Research / 3G Americas. Whereas multipath is an impediment for other radio systems. Figure 34: MIMO Using Multiple Paths to Boost Throughput and Capacity Encoder Decoder Tests of MIMO have proven very promising in WLANs operating in relative isolation where interference is not a dominant factor. 3GPP Release 7 and Beyond. however. refer to the 3G Americas’ white paper “Mobile Broadband: The Global Evolution of UMTS-HSPA.” HSPA to LTE-Advanced.

HSPA+ Taking advantage of these various radio technologies. because the radio can turn off between VoIP packets. which consumes the most power. Rysavy Research / 3G Americas. HSPA uses 16 QAM on the downlink and QPSK on the uplink. But radio links can achieve higher throughputs—adding 64 QAM on the downlink and 16 QAM on the uplink—precisely what is added in HSPA+. in turn. The goals of HSPA+ are to:  Exploit the full potential of a CDMA approach before moving to an OFDM platform in 3GPP LTE. CPC is especially beneficial to VoIP on the uplink. Figure 35: Continuous Packet Connectivity Data Pilot Without CPC Data Pilot With CPC Higher Order Modulation Another way of increasing performance is to use higher order modulation. these capabilities are referred to as HSPA+. September 2009 Page 85 . which is enabled through other enhancements such as receive diversity and equalization. 3GPP has standardized a number of features in Release 7 including higher order modulation and MIMO. CPC enhancements reduce the uplink interference created by the dedicated physical control channels of packet data users when those channels have no user data to transmit. wherein the modem can turn off its receiver after a certain period of HSDPA inactivity. Release 8 will include further enhancements.than MIMO for improving capacity in high spectral efficiency systems employing a reuse factor of 1. Achieve performance close to LTE in 5 MHz of spectrum. Higher order modulation requires a better SNR. This. Collectively.  HSPA to LTE-Advanced. See Figure 35. increases the number of simultaneously connected HSUPA users. CPC allows both discontinuous uplink transmission and discontinuous downlink reception. Continuous Packet Connectivity In Release 7.

0 Uplink (Mbps) Peak Data Rate 5. Dual. HSPA+ will also have improved latency performance of below 50 msec and improved packet call setup time of below 500 msec.    Depending on the features implemented. Allow operation in a packet-only mode for both voice and data.4 21.5 11. The prior discussion emphasizes throughput speeds. and HSPA+ with 42 Mbps capability on the downlink and 11.and multi-carrier operation are explained further below. It is also partly because HSPA supports Incremental Redundancy (IR) and has lower overhead than WiMAX. operators may choose to leverage the EPC/SAE planned for LTE. Be backward-compatible with previous systems while incurring no performance degradation with either earlier or newer devices. as well as additional MIMO radios at cell sites.5 23. thereby facilitating the operation of both technologies.0 Beyond the peak rate of 42 Mbps defined in Release 8. enterprise campuses. Facilitate migration from current HSPA infrastructure to HSPA+ infrastructure. Sleep-to-data-transfer times of less than 200 msec will improve users’ “always-connected” experience. HSPA+ with 28 Mbps capability will be available for deployment by the end of 2009. and in-building networks.5 11. Future releases of HSPA+ could also use a quad-carrier approach for even higher throughputs. Table 14: HSPA Throughput Evolution Technology HSPA as defined in Release 6 Release 7 HSPA+ DL 64 QAM. With advances in backhaul transport like metropolitan Ethernet. UL 16 QAM Release 9 HSPA+ 2X2 MIMO. operators will be able to expand coverage.0 42. UL 16 QAM Release 7 HSPA+ 2X2 MIMO. Given the large amount of backhaul bandwidth required to support HSPA+. As such. HSPA+ can exceed the capabilities of IEEE 802. Release 9 may specify 2X2 MIMO in combination with dual-carrier operation. UL 16 QAM Release 8 HSPA+ 2X2 MIMO DL 64 QAM. Provide smooth interworking between HSPA+ and LTE. but HSPA+ will also more than double HSPA capacity as well as reduce latency below 25 msec. which would further boost peak network rates to 84 Mbps. however.1 28.16e-2005 (mobile WiMAX) in the same amount of spectrum. and HSPA to LTE-Advanced. Rysavy Research / 3G Americas. September 2009 Page 86 . This is mainly because MIMO in HSPA supports closed-loop operation with precode weighting. as well as multicode-word MIMO and enables the use of SIC receivers. Dual Carrier Downlink (Mbps) Peak Data Rate 14.76 11.5 Mbps on the uplink could be ready for deployment by 2009 or 2010. Table 14 summarizes the capabilities of HSPA and HSPA+ based on various methods. operators are likely to initially deploy HSPA+ in limited “hotspot” coverage areas such as airports.2 84. DL 16 QAM.

Under consideration for Release 10 is the use of four channels. it is possible to achieve a doubling of the 21 Mbps maximum rate available on each channel to 42 Mbps. This approach coordinates the operation of HSPA on two adjacent 5 MHz carriers so that data transmissions can achieve higher throughput rates. comparable to what can be obtained with 2X2 MIMO. For networks that have implemented uplink diversity in the base station. operators will be able to introduce HSPA+ capabilities through either a software upgrade or hardware expansions to existing cabinets to increase capacity. Same maximum-throughput rate of 42 Mbps as using MIMO.reduced power consumption with VoIP will result in talk times that are more than 50 percent higher. Multi-user diversity also improves because there are more users to select from. Figure 37 shows an analysis of dual-carrier performance using a cumulative distribution function. especially in lightlyloaded networks. as shown in Figure 36. The work item assumes two adjacent carriers. 2009. resulting in what is called trunking gain. 106 Source: "LTE for UMTS. September 2009 Page 87 . HSPA to LTE-Advanced.   By scheduling packets across two carriers. CDF indicates the probability of achieving a particular throughput rate and the figure demonstrates a consistent doubling of throughput. From a deployment point of view. In this configuration. those multiple antennas will facilitate MIMO deployment. but with a less expensive infrastructure upgrade. Wiley. Dual-Carrier HSPA 3GPP has defined a capability in Release 8 for dual-carrier HSPA operation. there is better resource utilization. downlink operation and no MIMO. upgrading to 64-QAM support will be easier to implement than 2X2 MIMO for many networks. For example. Rysavy Research / 3G Americas. Certain upgrades will be simpler than others. however. Significantly higher peak throughputs available to users. OFDMA and SC-FDMA Based Radio Access.” Harri Holma and Antti Toskala. Figure 36: Dual-Carrier Operation with One Uplink Carrier106 Uplink 1 x 5 MHz UE1 1 x 5 MHz UE2 2 x 5 MHz Downlink 2 x 5 MHz There are a number of benefits to this approach:  An increase in spectral efficiency of about 20%.

There is also an integrated RNC/NodeB option in which RNC functions are integrated in the Node B. 10% load CDF [%] 60 50 40 30 20 10 0 RAKE. are similar to the EPC/SAE architecture. In Release 7. there is the option of a one-tunnel architecture by which the network establishes a direct transfer path for user data between RNC and GGSN. HSPA to LTE-Advanced. These new architectures. single-carrier RAKE. The integrated RNC/NodeB for HSPA+ has been agreed-upon as an optional architecture alternative for packetswitched-based services. single-carrier GRAKE2. This is particularly beneficial in femtocell deployments. single-carrier GRAKE.Figure 37: Dual-Carrier Performance107 100 90 80 70 Ped A. This brings several benefits such as eliminating hardware in the SGSN and simplified engineering of the network. 107 Source: 3G Americas’ member company contribution. Rysavy Research / 3G Americas. multi-carrier GRAKE. September 2009 Page 88 . while the SGSN still performs all control functions. multi-carrier 0 5 10 15 20 25 30 35 40 Achievable bitrate [Mbps] One-Tunnel Architecture Another way HSPA performance can be improved is through a flatter architecture. as shown in Figure 38. as an RNC would otherwise need to support thousands of femtocells. especially on the packet-switched core network side where they provide synergies with the introduction of LTE. multi-carrier GRAKE2.

During the HS-FACH state. 2007. Rysavy Research / 3G Americas. 108 Source: 3G Americas’ white paper.Figure 38: HSPA One-Tunnel Architecture108 Traditional HSPA Architecture GGSN User Plane Control Plane RNC Node B SGSN HSPA with One-Tunnel Possible HSPA+ with Architecture One-Tunnel Architecture GGSN SGSN RNC Node B Node B GGSN SGSN HSPA. September 2009 Page 89 . reduces setup time to practically zero and provides a more efficient way of carrying application signaling for always-on applications. and other advanced functions provide a compelling advantage for UMTS over competing technologies: The ability today to support voice and data services on the same carrier and across the whole available radio spectrum. illustrated in Figure 39. HS-FACH In Release 7. a new capability called High-Speed Access Forward Access Channel (HSFACH). to deliver data at ever-increasing broadband rates. to offer these services simultaneously to users. “UMTS Evolution from 3GPP Release 7 to Release 8. The network accomplishes this by using the same HSDPA power/code resources for access requests (CELL_FACH state) as for dedicated packet transfer (CELL_DCH). This allows data transmission to start during the HS-FACH state with increased data rates immediately available to the user equipment. the network allocates dedicated resources for transitioning the user equipment to a dedicated channel state. HSPA+. and to do so in a spectrally efficient manner.” HSPA to LTE-Advanced.

Wiley. Rysavy Research / 3G Americas. HSPA to LTE-Advanced. FACH RB reconfiguration No data flow during transition >500 ms eFACH Data flows on HSFACH also during transition. DCH /HSPA HSPA DCH – Dedicated Channel FACH – Forward Access Channel RACH – Reverse Access Channel PCH – Paging Channel HS-FACH – High Speed FACH RB – Radio Bearer RRC – Radio Resource Control Figure 40 summarizes the capabilities and benefits of the features being deployed in HSPA+. September 2009 Page 90 . 109 Source: "LTE for UMTS.5 s R7/R8 solution Seamless transition Cell_FACH RACH + FACH 6-32 kbps Cell_DCH HSDPA + HSUPA >1 Mbps Cell_FACH Cell_DCH HSDPA + HSUPA >1 Mbps R99-RC RRC States PCH RB reconfiguration Cell update and C-RNTI allocation takes >300 ms R7/8 RRC States PCH Immediate transmission w/o cell update. 2009.” Harri Holma and Antti Toskala. No PCH required. OFDMA and SC-FDMA Based Radio Access.Figure 39: High-Speed Forward Access Channel109 R99-R6 solution Delay >0.

Figure 40: Summary of HSPA Functions and Benefits110 Uplink DTX + downlink DRX CS voice over HSPA Downlink 64QAM. but then to connect the CS voice traffic back into the existing CS infrastructure (MSCs. current networks support simultaneous voice and data operation. There is an elegant alternative: To packetize the circuit-switched voice traffic which is already in digital form. etc. which is why it makes sense to use them to carry voice communications. use the HSPA channels to carry the CS voice. The following figure shows the infrastructure changes required at the Node B and within the RNC. Moreover. 110 Source: 3G Americas’ member contribution. HSPA to LTE-Advanced. reasons to consider alternate approaches including reducing power consumption and being able to support even more users. The other is VoIP. September 2009 Page 91 . requires not only supporting packetized voice in the radio channel. Doing so with VoIP. This requires relatively straightforward changes in just the radio network and in devices. Rysavy Research / 3G Americas. however. There are. One approach is called circuitswitched voice over HSPA. however. CS Voice over HSPA HSPA channels employ many optimizations to obtain a high degree of data throughput.) immediately beyond the radio access network. MIMO and Dual carrier Uplink 16QAM L2 optimization (Flexible RLC) High speed FACH + High speed RACH Lower UE power consumption Higher voice capacity Higher downlink peak data rates and higher data capacity Higher uplink peak data rates Higher L2 throughput and less processing requirements Lower latency = better response times More efficient common channels = savings in channel elements Flat architecture optimization Less network elements HSPA Voice Support Voice support with WCDMA-dedicated channels in UMTS networks is spectrally very efficient. but also within the infrastructure network.

Significantly improves battery life with voice communications. 2007. Supports both narrowband and wideband codecs. IuCS Combined to one carrier HSPA scheduler HSPA IuPS PS R99 NodeB RNC With this approach. Transparent to existing CS infrastructure. which is using VoIP. VoIP is possible in Release 6. September 2009 Page 92 . but it is enhancements in 111 Source: 3G Americas’ white paper. Acts as a stepping stone to VoIP over HSPA/LTE in the future. listed below. The many benefits of this approach. This will eventually increase voice capacity. “UMTS Evolution from 3GPP Release 7 to Release 8. make it highly likely that operators will adopt it:        Relatively easy to implement and deploy. HSPA CS voice can be deployed with Release 7 or later networks. while new devices use HSPA channels. VoIP Once HSDPA and HSUPA are available. Provides a 50 to 100% capacity gain over current voice implementations.” HSPA to LTE-Advanced.Figure 41: Implementation of HSPA CS Voice111 Scheduler prioritizes voice packets CS mapped to R99 or HSPA bearer depending on terminal capability Transport queues etc CS R99 AMR adaptation possible AMR adapt. and enable innovative new applications that combine voice with data functions in the packet domain. legacy mobile phones can continue using WCDMA-dedicated traffic channels for voice communications. Rysavy Research / 3G Americas. operators will have another option of moving voice traffic over to these high-speed data channels. allow operators to consolidate their infrastructure on an IP platform. Enables faster call connections.

those based on IMS).8 PS Evolution VoIP capacity gains are quantified in detail in the main part of in this paper. One attractive aspect of deploying VoIP with HSPA is that operators can smoothly migrate users from circuit-switched operation to packet-switched operation over time.Release 7 that make it highly efficient and thus attractive to network operators. Moreover. Rysavy Research / 3G Americas. circuit-switched voice service and already allows simultaneous voice/data operation. and only over time will it transition to primary voice service. 3GPP LTE Although HSPA and HSPA+ offer a highly efficient broadband-wireless service that will enjoy success for the remainder of this decade and well into the next. 112 Source: 3G Americas’ member contribution.4 1.2 0 0 Power reserved for PS traffic (W) 2 4 6 8 10 12 14 Relative Capacity 0.4 0. September 2009 Page 93 . Figure 42 shows a system’s voice capacity with the joint operation of circuitswitched and IP-based voice services. 3GPP has completed the specification for Long Term Evolution as part of Release 8. it may not be used immediately for primary voice services. with HSPA+. packet voice requires a considerable amount of new infrastructure in the core network. HSPA to LTE-Advanced. Work on LTE began in 2004 with an official work item started in 2006 and a completed specification early 2009. packet voice will likely be used initially as part of other services (for example. As a result. Whereas packet voice is the only way voice will be supported in LTE. Figure 42: Ability for UMTS to Support Circuit and Packet Voice Users112 1.2 1 VoIP CS CS + VoIP 0. Initial deployments will occur in 2010. LTE will allow operators to achieve even higher peak throughputs in higher spectrum bandwidth. discussed later in this paper. some voice users can be on legacy circuit-switched voice and others can be on VoIP. VoIP will be implemented in conjunction with IMS. They range from 20 percent to as high as 100 percent with the implementation of interference cancellation and the minimization of IP overhead through a scheme called Robust Header Compression (ROHC).6 0. This is because UMTS already has a highly efficient. Because the UMTS radio channel supports both circuit-switched voice and packetswitched data.

and LTE will operate in various radio channel sizes ranging from 1. LTE uses an approach called SC-FDMA. radio-access technology that offers full vehicular speed mobility and that can readily coexist with HSPA and earlier networks. WCDMA radio technology is basically as efficient as OFDM for delivering peak data rates of about 10 Mbps in 5 MHz of bandwidth. Scalable bandwidth up to 20 MHz covering 1.4.LTE uses OFDMA on the downlink. September 2009 Page 94 . and 20 MHz in the study phase. This is where OFDM provides a practical implementation advantage. and preferences that will automate network planning and will result in lower operator costs. 5. would result in highly complex terminals.    LTE Throughput Rates The overall objective is to provide an extremely high-performance. Table 15 shows LTE peak data rates based on different downlink and uplink designs. Because of scalable bandwidth. The OFDMA approach is also highly flexible in channelization. and to less than 100 msec transition times from inactive to active. Hence.4 Mbps with 20 MHz bandwidth.4 Uplink (Mbps) Peak Data Rate 57. thereby boosting spectral efficiency.6 86. 3. a pure OFDMA approach results in high Peak to Average Ratio (PAR) of the signal. to 10 msec round-trip times between user equipment and the base station. but has a 2 to 6 dB PAR advantage over the OFDMA method used by other technologies such as WiMAX. Self-optimizing capabilities under operator control. On the uplink. 10. Achieving peak rates in the 100 Mbps range with wider radio channels. Table 15: LTE Peak Throughput Rates LTE Configuration Using 2X2 MIMO in the Downlink and 16 QAM in the Uplink Using 4X4 MIMO in the Downlink and 64 QAM in the Uplink Downlink (Mbps) Peak Data Rate 172. which compromises power efficiency and. ultimately. and it is not practical with current technology.8 326. 15. Operation in both TDD and FDD modes.4 HSPA to LTE-Advanced. Uplink peak data rates up to 86. however. Rysavy Research / 3G Americas. which is well suited to achieve high peak data rates in high-spectrum bandwidth.4 to 20 MHz. operators will be able to easily migrate their networks and users from HSPA to LTE over time. LTE capabilities include:     Downlink peak data rates up to 326 Mbps with 20 MHz bandwidth. which is somewhat similar to OFDMA. Reduced latency. Scheduling approaches in the frequency domain can also minimize interference. Increased spectral efficiency over Release 6 HSPA by a factor of two to four. battery life. however.

LTE is not only efficient for data but.200 in a 20 MHz channel. The different narrowband streams are generated in the frequency domain. Rysavy Research / 3G Americas.0 msec. and the IFFT conserves that characteristic. The system is called orthogonal. HSPA to LTE-Advanced. OFDM systems may lose their orthogonal nature as a result of the Doppler shift induced by the speed of the transmitter or the receiver. A minimum resource block that the system can assign to a user transmission consists of 12 subcarriers over 14 symbols in 1. and then combined to form the broadband stream using a mathematical algorithm called an Inverse Fast Fourier Transform (IFFT) that is implemented in digital-signal processors. the subcarriers have 15 kHz spacing from each other. which simplifies radio design. It results in an almost complete elimination of multipath induced Intersymbol Interference (ISI).4 MHz channel to 1. 3GPP specifically selected the subcarrier spacing of 15 kHz to avoid any performance degradation in high-speed conditions. The basic principle of OFDM is to split a high-rate data stream into a number of parallel low-rate data streams. WiMAX systems that use a lower subcarrier spacing (~11 kHz) will be more impacted in high-speed conditions than LTE. because of a highly efficient uplink.113 OFDMA and Scheduling LTE implements OFDM in the downlink. because the subcarriers are generated in the frequency domain (making them inherently orthogonal). September 2009 Page 95 . each a narrowband signal carried by a subcarrier. LTE maintains this spacing regardless of the overall channel bandwidth. LTE VoIP capacity will reach almost 500 users. Figure 43: OFDM Symbol with Cyclic Prefix Cyclic Prefix (4.7 sec) The multiple-access aspect of OFDMA comes from being able to assign different users different subcarriers over time. Figure 44 shows how the system can assign these resource blocks to different users over both time and frequency. The composite signal is obtained after the IFFT is extended by repeating the initial part of the signal (called the Cyclic Prefix [CP]). is extremely efficient for VoIP traffic.8 sec) Data (66. The CP is basically a guard time during which reflected signals will reach the receiver. This extended signal represents an OFDM symbol. In 10 MHz of spectrum. especially in supporting radio channels of different widths. The number of subcarriers ranges from 72 in a 1. In LTE. which otherwise makes extremely high data-rate transmissions problematic. 113 Source: 3GPP Multi-member analysis.

Rysavy Research / 3G Americas. as shown in Figure 45 can use those resource blocks that are not faded. This results in up to a 40% gain in average cell throughput for low user speed (3 km/hour). something that is not possible in CDMA-based systems. but by using a different one third of the subcarriers in each sector. but delivers high peak rates to users. The looser frequency reduces overall spectral efficiency. the system achieves a looser frequency reuse of 1/3. assuming a large number of users and no MIMO. HSPA to LTE-Advanced. LTE can easily control frequency reuse. September 2009 Page 96 . Since different frequencies may fade differently for different users. the system can allocate those frequencies for each user that result in the greatest throughput. By using all the subcarriers in each sector. frequency domain scheduling.Figure 44: LTE OFDMA Downlink Resource Assignment in Time and Frequency User 1 User 2 User 3 Frequency User 4 Time Minimum resource block consists of 14 symbols and 12 subcarriers By having control over which subcarriers are assigned in which sectors. The benefit decreases at higher user speeds. Beyond controlling frequency reuse. the system would operate at a frequency reuse of 1.

whereas MU-MIMO does not require any additional implementation at the device. A 3G Americas’ member analysis predicts 40% lower spectral efficiency with 1. identical symbol-level numerology. achieves nearly all of its efficiency with 5 MHz channels or wider. HSPA to LTE-Advanced. and similar synchronization and control channels. however. these include 2X2. the possibility of using similar reference signal patterns. and 4X4.Figure 45: Frequency-Domain Scheduling in LTE114 Carrier bandwidth Resource block Frequency Transmit on those resource blocks that are not faded LTE is specified for a variety of MIMO configurations. there are two possible approaches: single-user MIMO (SU-MIMO) and multi-user MIMO (MU-MIMO). This will simplify deployment of LTE TDD in 114 115 3G Americas’ member contribution. there is only one TDD variant. LTE is designed to operate in channel bandwidths from 1.115 The system.4 MHz to 20 MHz. The greatest efficiency. On the uplink. The first LTE release thus incorporates MU-MIMO with SU-MIMO deferred for the second LTE release. LTE TDD has been designed to co-exist with TDSCDMA and TD-CDMA/UTRA (both low-chip rate and high-chip rate versions). Rysavy Research / 3G Americas. Furthermore. 4X4 will be most likely used initially in femtocells. SU-MIMO is more complex to implement as it requires two parallel radio transmit chains in the mobile device. Also. however. 3G Americas’ member company analysis 2009. September 2009 Page 97 . occurs with higher bandwidth. On the downlink. LTE TDD achieves compatibility and co-existence with TD-SCDMA by defining frame structures where the DL and UL time periods can be time aligned to prevent BTS to BTS and UE to UE interference to support operation in adjacent carriers without the need for large guardbands between the technologies. 4X2 (four antennas at the base station).4 MHz radio channels and 13% lower efficiency with 3 MHz channels. TDD Harmonization 3GPP developed LTE TDD to be fully harmonized with LTE FDD including alignment of frame structures. Initial deployment will likely be 2x2.

. Minimum downlink peak spectral efficiency of 15 bps/Hz (assumes 4X4 MIMO). Page 98 HSPA to LTE-Advanced. September 2009 . Source: http://www. Minimum uplink peak spectral efficiency of 6.117 4G.org/workprogramme. 116 117 Source: A 3G Americas’ member company.2134. Encouragement to support wider bandwidths (e. The organization Next Generation Mobile Networks has a project for LTE TDD and FDD convergence.ngmn.html. Figure 46: TDD Frame Co-Existence Between TD-SCDMA and LTE TDD116 For LTE FDD and TDD to coexist. Table 16 shows the requirements for cell-spectral efficiency. IMT-Advanced and LTE-Advanced As introduced earlier in this paper. Some of the key requirements or statements include:     Support for scalable bandwidth up to and including 40 MHz.countries such as China that are deploying TD-SCDMA.g. 100 MHz). Rysavy Research / 3G Americas. Figure 46 demonstrates the synchronization between TC-SCDMA and LTE-TDD in adjacent channels.75 bps/Hz (assumes 2X4 MIMO). large guardbands will be needed to prevent interference. the term 4G will apply to networks that comply with the requirements of IMT-Advanced that are articulated in Report ITU-R M.

meaning LTE devices will operate in newer LTE-Advanced networks. a project that is a study item in 2009 with specifications expected in the second half of 2010 as part of Release 10. SAE/LTE and LTE-Advanced. and LTE-Advanced devices will operate in older LTE networks. Rysavy Research / 3G Americas.and forwards-compatible with LTE.2135.Table 16: IMT-Advanced Requirements for Cell-Spectral Efficiency Test Environment118 Indoor Microcellular Base Coverage Urban High Speed Downlink (bps/Hz) 3. Coordinated multipoint transmission (CoMP) with two proposed approaches: coordinated scheduling and/or beamforming. thereby achieving higher system capacity and improving cell-edge data rates. Uplink MIMO (two transmit antennas in the device).25 1. The intent is to closely coordinate transmissions at different cell sites. 118 119 120 Test environments are described in IT Report ITU-R M.7. LTE-Advanced will be both backwards. and joint processing/transmission. refer to section 7.7 Table 17 shows the requirements for voice capacity.120 Figure 47 shows the carrier aggregation. with up to 100 MHz of bandwidth supported. 3GPP is studying the following capabilities for LTE-Advanced:     Wider bandwidth support for up to 100 MHz via aggregation of 20 MHz blocks.2 1. September 2009 Page 99 . Downlink MIMO of up to 8 by 8 as described below.6 2.5 of the 3G Americas’ white paper “The Mobile Broadband Evolution: 3G Release 8 and Beyond.8 1. For further details. HSPA+.2135.0 2. Test environments are described in IT Report ITU-R M.1 Uplink (bps/Hz) 2.4 0. Table 17: IMT-Advanced Requirements for Voice Capacity Test Environment119 Indoor Microcellular Base Coverage Urban High Speed Minimum VoIP Capacity (Active Users/Sector/MHz) 50 40 40 30 3GPP is addressing the IMT-Advanced requirements through a version of LTE called LTEAdvanced.” HSPA to LTE-Advanced.

2009. Figure 48: Carrier Aggregation at Different Protocol Layers122 Beyond wider bandwidths. conceivably supporting a peak rate of 1 Gbps in just 40 MHz and even higher rates in wider bandwidths. http://www. Stefan Parkvall and David Astely. LTE-Advanced will 121 Source: "LTE for UMTS. Ericsson Research.Figure 47: Release 10 LTE-Advanced Carrier Aggregation121 Release 10 LTE-Advanced UE resource pool Rel’8 Rel’8 Rel’8 Rel’8 Rel’8 100 MHz bandwidth 20 MHz Release 8 UE uses a single 20 MHz block Figure 48 shows the carrier aggregation operating at different protocol layers.academypublisher. LTE-Advanced will extend performance through more powerful multi-antenna capabilities. Rysavy Research / 3G Americas. This would require additional reference signals for channel estimation and for measurements such as channel quality to enable adaptive. OFDMA and SC-FDMA Based Radio Access. September 2009 Page 100 .” Harri Holma and Antti Toskala. the technology will be able to transmit in up to 8 layers using an 8X8 configuration for a peak spectral efficiency of 30 bps/Hz that exceeds the IMT-Advanced requirements. 122 Source: “The Evolution of LTE towards IMT-Advanced”.pdf HSPA to LTE-Advanced. Wiley. For the downlink.com/jcm/vol04/no03/jcm0403146154. multi-antenna transmission.

07 bps/Hz In all cases.also include four-layer transmission in the uplink resulting in spectral efficiency exceeding 15 bps/Hz. Rysavy Research / 3G Americas.75 bps/Hz 2.0 bps/Hz 0. 123 Spectral efficiency values based on 4 antennas at the base station and 2 antennas at the terminal.6 bps/Hz 2. Item Peak Data Rate Downlink Peak Data Rate Uplink Spectrum Allocation Latency User Plane Latency Control Plane Peak Spectral Efficiency DL123 Peak Spectral Efficiency UL Average Spectral Efficiency DL Average Spectral Efficiency UL Cell-Edge Spectral Efficiency DL Cell-Edge Spectral Efficiency UL Up to 40 MHz 10 msec 100 msec 15 bps/Hz 6.2 bps/Hz 1. The idea is to relay frames at an intermediate node. HSPA to LTE-Advanced. with better signal quality. Table 18 shows anticipated performance relative to IMT-Advanced Requirements. user rates will be much improved.06 bps/Hz 0. resulting in much better in-building penetration and.03 bps/Hz IMT-Advanced Requirements and Anticipated LTE-Advanced IMT-Advanced Requirement LTE-Advanced Projected Capability 1 Gbps 500 Mbps Up to 100 MHz 10 msec 50 msec 30 bps/Hz 15 bps/Hz 2. September 2009 Page 101 .09 bps/Hz 0. Table 18: Capability. projections of LTE-Advanced performance exceed that of the IMT-Advanced requirements. Another capability being planned for LTE-Advanced is relays as shown in Figure 49.4 bps/Hz 0.

September 2009 Page 102 . called UMTS TDD. it would not be able to operate effectively in both bands. especially if it includes wide radio channels. in which both transmit and receive functions alternate in time on the same radio channel. Many data applications. and the Pacific region have licensed spectrum available specifically for TDD systems. operators are likely to eventually upgrade their LTE networks to LTE-Advanced to obtain spectral efficiency gains and capabilities such as relaying. in which the operator uses different radio bands for transmit and receive. as new spectrum becomes available next decade. there is limited spectrum specifically allocated for TDD systems. In the United States. A TDD radio interface can dynamically adjust the downlink-to-uplink ratio accordingly. One consideration. LTE-Advanced will have tremendous capability. Though initial deployments of LTE will be based on Release 8. which offers improved performance. the same amount of capacity in both the uplink and the downlink. An alternate approach is TDD. For this spectrum. 3GPP specifications include a TDD version of UMTS. Asia. especially for applications like Web browsing or multimedia downloads. which need balanced links—namely. TDD does not provide any inherent advantage for voice functions. Various countries around the world including those in Europe. however. The UMTS TDD specification also includes the capability to use joint detection in receiver-signal processing. however.125 UMTS TDD is not a good choice in FDD bands.Figure 49: LTE-Advanced Relay124 Direct Link Relay Link Access Link As demonstrated in this section. are asymmetric. often with the downlink consuming more bandwidth than the uplink. Even in existing bands. 124 Source: 3G Americas’ member contribution. the higher spectral efficiency achievable in the downlink versus the uplink is critical in addressing the asymmetrical nature of most data traffic. hence balancing both forward-link and reverse-link capacity. Note that for UMTS FDD. thereby making the overall system efficiency relatively poor. It is also a good choice in any spectrum that does not provide a duplex gap between forward and reverse links. Rysavy Research / 3G Americas. is a good choice. UMTS TDD. relates to available spectrum. UMTS TDD Most WCDMA and HSDPA deployments are based on FDD. or in the future LTE in TDD mode. then LTE-Advanced will be the ideal technology for these new bands. HSPA to LTE-Advanced.

subscriber databases. Other organizations supporting IMS include the GSM Association (GSMA). Rather.84 Mcps for UMTS TDD. the ITU. The QoS mechanisms in UMTS will be an important component of some IMS applications. CableLabs. These include the Internet Engineering Taskforce (IETF). Other protocols include Real Time Transport Protocol (RTP) and Real Time Streaming Protocol (RTSP) for transporting actual sessions. the Telecoms and Internet Converged Services and Protocols for Advanced Networks (TISPAN). Specified through 3GPP as a variant of the UMTS TDD System and operating with a 1. 3GPP2. but has never been used. There has been little deployment of UMTS TDD.As discussed in more detail in the “WiMAX” section. This makes it a possible alternative for wireless local loops. TD-SCDMA technology is not as mature as UMTS and CDMA2000. and one initial application under consideration—PoC—is being specified by the Open Mobile Alliance. The Parlay Group. with 2008 being the first year of limited deployments in China in time for the Olympic Games. Operators already have IMS trails in progress. service-layer applications. mostly for deployment in China. The exact services will depend on cellular operators and the application developers that make these applications available to operators. Rysavy Research / 3G Americas. VoIP. September 2009 Page 103 . which specifies endto-end. and the Open Mobile Alliance. and likely will. the ETSI. PoC. IMS is relatively independent of the radio-access network and can. TDSCDMA uses the same core network as UMTS. Although originally specified by 3GPP. and the Java Community Process (JCP). numerous other organizations around the world are supporting IMS. the primary attribute of TD-SCDMA is that it is designed to support very high subscriber densities. The core networking protocol used within IMS is Session Initiation Protocol (SIP). Potential applications include video sharing. Other applications include picture and video sharing that occur in parallel with voice communications. Although there are no planned deployments in any country other than China. Operators looking to roll out VoIP over 125 The 1910-1920 MHz band targeted unlicensed TDD systems. which may be problematic in certain bands. be used by other radio-access networks or wireline networks. which includes the companion Session Description Protocol (SDP) used to convey configuration information such as supported voice codecs. and so forth. it provides a framework of application servers. TDD systems require network synchronization and careful coordination between operators or guardbands. which specifies key protocols such as SIP. TD-SCDMA could theoretically be deployed anywhere unpaired spectrum is available—such as the bands licensed for UMTS TDD—assuming appropriate resolution of regulatory issues. and gateways to make them possible. Future TDD deployments of 3GPP technologies are likely to be based on LTE. the American National Standards Institute (ANSI). IMS by itself does not provide all these applications. streaming video. TD-SCDMA TD-SCDMA is one of the official 3G wireless technologies being developed. IMS IMS is a service platform that allows operators to support IP multimedia applications. and it is possible for the same core network to support both UMTS and TD-SCDMA radio-access networks.28 megachips per second (Mcps) chip rate against 3. HSPA to LTE-Advanced. interactive gaming.

tighter integration with the Internet. Because multiple users in a cell are tuned to the same content. and a lower cost infrastructure that is based on IP building blocks used for both voice and data services. wherein multiple users receive the same information using the same radio resource. only users with subscriptions receive the information. broadcasting and multicasting result in much greater spectrum efficiency for services such as mobile TV. IMS applications can reside either in the operator’s network or in third-party networks including those of enterprises. Figure 50: IP Multimedia Subsystem IMS Home Subscriber Server (HSS) DIAMETER SIP Application Server SIP Media Resource Function Control Media Resource Gateway Control Call Session Control Function (CSCF) (SIP Proxy) UMTS/HSPA Packet Core Network DSL Wi-Fi Multiple Possible Access Networks The benefits of using IMS include handling all communication in the packet domain. This allows operators to potentially deliver data and voice services at lower cost. Service areas for both broadcast and multicast can span either the entire network or a specific geographical area. Rysavy Research / 3G Americas. In a broadcast. This enables the receiver to constructively superpose HSPA to LTE-Advanced. every subscriber unit in a service area receives the information. Release 7 includes optimizations through a solution called multicast/broadcast. By managing services and applications centrally—and independently of the access network—IMS can enable network convergence. As shown in Figure 50. Broadcast/Multicast Services An important capability for 3G and evolved 3G systems is broadcasting and multicasting. This creates a much more efficient approach for delivering content such as video programming to which multiple users have subscriptions.networks could also use IMS. and wireline networks. This allows operators to offer common services across 3G. 3GPP initially introduced IMS in Release 5 and has enhanced it in each subsequent specification release. 3GPP defined highly-efficient broadcast/multicast capabilities for UMTS in Release 6 with MBMS. whereas in a multicast. September 2009 Page 104 . single-frequency network operation that involves simultaneous transmission of the exact waveform across multiple cells. thus providing these services at lower prices and further driving demand and usage. IMS operates just outside the packet core. Wi-Fi.

One important performance aspect of EPC/SAE is a flatter architecture. WCDMA-based broadcast transmission technology that matches the benefits of OFDMA-based broadcast approaches. It will support service continuity across heterogeneous networks. EPC/SAE could also be deployed for use with HSPA+ where it could provide a stepping-stone to LTE. September 2009 Page 105 . EPC/SAE will be optimized for all services to be delivered via IP in a manner that is as efficient as possible—through minimization of latency within the system. the networks are highly optimized for broadcast. The eNodeB (base station) integrates the functions traditionally HSPA to LTE-Advanced. for example. called Evolved Node B (eNodeB) and the Access Gateway (AGW). As such. lower-latency. EPC/SAE 3GPP is defining EPC/SAE in Release 8 as a framework for an evolution or migration of the 3GPP system to a higher-data-rate. For packet flow. because the mobile system can combine the signal from multiple base stations and because of the narrowband nature of OFDM. EPC/SAE includes two network elements. the LTE broadcast capability is expected to be quite efficient. packet-optimized system that supports multiple radio-access technologies. OFDM is particularly well-suited for broadcasting. Although it will most likely be deployed in conjunction with LTE. which will be important for LTE operators who must simultaneously support GSM-HSPA customers. Figure 51: OFDM Enables Efficient Broadcasting An alternate approach for mobile TV is to use an entirely separate broadcast network with technologies such as Digital Video Broadcasting–Handheld (DVB-H) or Media Forward Link Only (MediaFLO). these signals would interfere with each other. Normally.multiple MBSFN cell transmissions. which various operators around the world have opted to do. The result is highly efficient. Although this requires a separate radio in the mobile device. The focus of this work is on the packetswitched domain with the assumption that the system will support all services—including voice—in this domain. Rysavy Research / 3G Americas. LTE will also have a broadcast/multicast capability.

The user plane functions consist of two elements: A serving gateway that addresses 3GPP mobility and terminates eNodeB connections. the AGW integrates the functions traditionally performed by the SGSN and GGSN. while the MME is centrally located to minimize complexity and cost. and PDN gateways can be collocated in the same physical node or distributed. September 2009 Page 106 . e. The MME. Figure 52: EPC/SAE Architecture GERAN Rel’7 Legacy GSM/UMTS SGSN UTRAN One-Tunnel Option Control MME PCRF Evolved RAN. IMS EPC/SAE Access Gateway Non 3GPP IP Access HSPA to LTE-Advanced. serving gateway.g. and a Packet Data Network (PDN) gateway that addresses service requirements and also terminates access by non-3GPP networks. It will also manage QoS across the whole system. EPC/SAE also allows integration of non-3GPP networks such as WiMAX. handled through the Mobile Management Entity (MME). based on vendor implementations and deployment scenarios. EPC/SAE will use IMS as a component. which will be essential for enabling a rich set of multimedia-based services. The EPC/SAE architecture is similar to the HSPA One-Tunnel Architecture discussed in the “HSPA+” section that allows for easy integration of HSPA networks to the EPC/SAE. and user plane (data communications) functions. The AGW has both control functions.performed by the radio-network controller. Meanwhile.. so that the EPC/SAE Access Gateway is located close to the RAN in a distributed fashion to reduce latency. Another architectural option is to reverse the topology. which previously was a separate node controlling multiple Node Bs. Figure 52 shows the EPC/SAE architecture. LTE User Plane Serving Gateway PDN Gateway IP Services. Rysavy Research / 3G Americas.

The QCI is a scalar denoting a set of transport characteristics (bearer with/without guaranteed bit rate. namely transporting circuitswitched voice over LTE. Guaranteed Bit Rate (GBR) and Maximum Bit Rate (MBR). allocates IP addresses.126 The rules provide for fixed devices and personal/portable devices. queue management thresholds. all devices must be able to sense the spectrum to detect both TV broadcasting and wireless microphone signals. scheduling weights. Each packet flow is mapped to a single QCI value (nine are defined in the Release 8 version of the specifications) according to the level of service required by the application. enforces policy. The need for supporting a broader variety of applications requiring higher bandwidth and lower latency led 3GPP to alleviate the existing (UMTS Release 99) Quality of Service (QoS) principles with the introduction for EPS of a QoS Class Identifier (QCI). packet error loss rate) and used to infer nodes specific parameters that control packet forwarding treatment (e. admission thresholds. September 2009 Page 107 .. The usage of the QCI avoids the transmission of a full set of QoS-related parameters over the network interfaces and reduces the complexity of QoS negotiation. This approach is not currently part of any 3GPP specifications. if applicable. does routing. The rules include transmit power limits and emission limits. However. determines the QoS associated to an EPS bearer. The QCI.Elements of the SAE architecture include:    Support for legacy GERAN and UTRAN networks connected via SGSN. Rysavy Research / 3G Americas. The MME that supports user equipment context and identity. called Voice over LTE Generic Access (VOLGA).). The FCC has suggested two usage types: broadband services to homes and businesses at a higher power level to fixed devices over larger geographical areas. etc. The Serving Gateway that terminates the interface toward the 3GPP radio-access networks. there is an alternative voice approach being discussed in the industry. HSPA to LTE-Advanced.g. White Space The FCC in the US has ruled that unlicensed devices that have mechanisms to not interfere with TV broadcast channels may use TV channels that are not in use. priority. In addition. To prevent interference with TV transmissions.    3GPP is planning to support voice in EPS through VoIP and IMS. as well as having the ability to access a database that lists permitted channels for a specific location. packet delay budget. as well as authenticating and authorizing users. together with Allocation-Retention Priority (ARP) and. 126 FCC-08-260: 2nd Report & Order. both device types must employ geolocation capability with 50-meter accuracy (although fixed devices can store their position during installation). The PDN gateway that controls IP data services. link-layer protocol configuration. The Policy Control and Charging Rules Function (PCRF) that manages QoS aspects. and provides access for non-3GPP access networks. and wireless portable devices at a lowpower level in indoor environments. A mapping between EPS and pre-Release 8 QoS parameters has been defined to allow proper interworking with legacy networks. Support for new radio-access networks such as LTE.

Rysavy Research / 3G Americas. has developed a standard.The frequency-sensing and channel-change requirements are not supported by today’s 3GPP. The IEEE. September 2009 Page 108 . IEEE 802. The industry is in the very early stages of determining the viability of using white-space spectrum and. 3GPP2 and WiMAX technologies. based on IEEE 802.22 is aimed at fixed or nomadic services such as DSL replacement. IEEE 802.16 concepts. HSPA to LTE-Advanced.22. however. that complies with the FCC requirements. at this time. there are no products or services available.

September 2009 Page 109 . Data Voice 1XRTT – One Carrier Radio Transmission Technology 2G – Second Generation 3G – Third Generation 3GPP – Third Generation Partnership Project 3GPP2 – Third Generation Partnership Project 2 4G – Fourth Generation (meeting requirements set forth by the ITU IMT-Advanced project) 8-PSK – Octagonal Phase Shift Keying AAS – Adaptive Antenna Systems ABR – Allocation Retention Priority AGW – Access Gateway AMR – Adaptive Multi Rate ANSI – American National Standards Institute ARP – Allocation Retention Priority ARQ – Automatic Repeat Request ARPU – Average Revenue Per User ATM – Asynchronous Transfer Mode AWGN – Additive White Gaussian Noise Channel BCCH – Broadcast Control Channel bps – bits per second BRS – Broadband Radio Service BSC – Base Station Controller BTS – Base Transceiving Station C/I – Carrier to Interference Ratio CAPEX.Acronyms The following acronyms are used in this paper.Capital Expenditure CDF – Cumulative Distribution Function CDMA – Code Division Multiple Access CMOS – Complementary Metal Oxide Semiconductor CP – Cyclic Prefix CPC – Continuous Packet Connectivity CRM – Customer Relationship Management DAS – Downlink EGPRS2-A Level Scheme dB – Decibel DBS – Downlink EGPRS2-B Level Scheme DC-HSPA – Dual Carrier HSPA DFT – Discrete Fourier Transform HSPA to LTE-Advanced. 1xEV-DO – One Carrier Evolved. Rysavy Research / 3G Americas. Data Optimized 1xEV-DV – One Carrier Evolved. Acronyms are defined on first use.

DSL – Digital Subscriber Line DTM – Dual Transfer Mode D-TxAA – Double Transmit Adaptive Array DVB-H – Digital Video Broadcasting Handheld E–DCH – Enhanced Dedicated Channel EBCMCS – Enhanced Broadcast Multicast Services EDGE – Enhanced Data Rates for GSM Evolution EGPRS – Enhanced General Packet Radio Service eNodeB – Evolved Node B EPS – Evolved Packet System ERP – Enterprise Resource Planning ETRI – Electronic and Telecommunications Research Institute ETSI – European Telecommunications Institute E-UTRAN – Enhanced UMTS Terrestrial Radio Access Network EV-DO – One Carrier Evolved.High Speed Physical Downlink Shared Channels HSPA to LTE-Advanced. Rysavy Research / 3G Americas. September 2009 Page 110 . Data Optimized EV-DV – One Carrier Evolved. Data Voice EVRC – Enhanced Variable Rate Codec FCC – Federal Communications Commission FDD – Frequency Division Duplex Flash OFDM – Fast Low-Latency Access with Seamless Handoff OFDM FLO – Forward Link Only FMC – Fixed Mobile Convergence FP7 – Seventh Framework Programme FTP – File Transfer Protocol G-Rake – Generalized Rake Receiver Gbps – Gigabits Per Second GBR – Guaranteed Bit Rate GERAN – GSM EDGE Radio Access Network GGSN – Gateway GPRS Support Node GHz — Gigahertz GMSK – Gaussian Minimum Shift Keying GPRS – General Packet Radio Service GSM – Global System for Mobile communications GSMA – GSM Association HARQ – Hybrid Automatic Repeat Request HD – High Definition HLR – Home Location Register HSDPA – High Speed Downlink Packet Access HS-FACH – High Speed Forward Access Channel HS-PDSCH .

Multimedia Broadcast/Multicast Service Mbps – Megabits Per Second MBR – Maximum Bit Rate Mcps – Megachips Per Second MCS – Modulation and Coding Scheme MediaFLO – Media Forward Link Only MHz – Megahertz MIMO – Multiple Input Multiple Output mITF – Japan Mobile IT Forum MMDS – Multichannel Multipoint Distribution Service MME – Mobile Management Entity MMSE – Minimum Mean Square Error MRxD – Mobile Receive Diversity MS – Mobile Station MSA – Mobile Service Architecture MSC – Mobile Switching Center HSPA to LTE-Advanced. September 2009 Page 111 .HSPA – High Speed Packet Access (HSDPA with HSUPA) HSPA+ – HSPA Evolution HSUPA – High Speed Uplink Packet Access Hz – Hertz ICT – Information and Communication Technologies IEEE – Institute of Electrical and Electronic Engineers IETF – Internet Engineering Taskforce IFFT – Inverse Fast Fourier Transform IM – Instant Messaging IMS – IP Multimedia Subsystem IMT – International Mobile Telecommunications IPR .Intellectual Property Rights IP – Internet Protocol IPTV – Internet Protocol Television IR – Incremental Redundancy ISI – Intersymbol Interference ISP – Internet Service Provider ITU – International Telecommunications Union JCP – Java Community Process kbps – Kilobits Per Second kHz — Kilohertz km – Kilometer LSTI – LTE/SAE Trial Initiative MAC – Medium Access Control MBMS . Rysavy Research / 3G Americas.

September 2009 Page 112 .MU-MIMO – Multi-User MIMO msec – millisecond NGMC – Next Generation Mobile Committee OFDM – Orthogonal Frequency Division Multiplexing OFDMA – Orthogonal Frequency Division Multiple Access PAR – Peak to Average Ratio PBCCH – Packet Broadcast Control Channel PCRF – Policy Control and Charging Rules Function PCS – Personal Communications Service PHY – Physical Layer PDN – Packet Data Network PoC – Push-to-talk over Cellular QAM – Quadrature Amplitude Modulation QCI – Quality of Service Class Identifier QoS – Quality of Service QPSK – Quadrature Phase Shift Keying RAB – Radio Access Bearer RAN – Radio Access Network RCS – Rich Communications Suite REST – Representational State Transfer RF – Radio Frequency RNC – Radio Network Controller ROHC – Robust Header Compression RTP – Real Time Transport Protocol RTSP – Real Time Streaming Protocol SC-FDMA – Single Carrier Frequency Division Multiple Access SAE – System Architecture Evolution SDMA – Space Division Multiple Access SDP – Session Description Protocol SGSN – Serving GPRS Support Node SIC – Successive Interference Cancellation SIP – Session Initiation Protocol SMS – Short Message Service SNR – Signal to Noise Ratio SU-MIMO – Single User MIMO TCH – Traffic Channel TDD – Time Division Duplex TDMA – Time Division Multiple Access TD-SCDMA – Time Division Synchronous Code Division Multiple Access TD-CDMA – Time Division Code Division Multiple Access HSPA to LTE-Advanced. Rysavy Research / 3G Americas.

Multi-vendor assessment. UMTS. HSPA and SAE/LTE”.” References 3G Americas: “Mobile Broadband: The Global Evolution of UMTS-HSPA – 3GPP Release 7 and Beyond.org. at info@3gamericas. June 2008 Update. and HSDPA deployments worldwide.” July 2006. available for free download on its Web site: http://www.” August 1. HSPA and SAE/LTE”. 3G Americas: “Global UMTS and HSPA Operator Status”. If there are any questions regarding the download of this information. July 11. Public Relations Administrator. VoIP. Downlink. September 2009 Page 113 . 2007.3gamericas. ABI Research: press release on study “Mobile Business Applications and Services. 3GPP: LTE Performance Summary. HSPA to LTE-Advanced. Uplink.TIA/EIA – Telecommunications Industry Association/Electronics Industry Association TISPAN – Telecoms and Internet converged Services and Protocols for Advanced Networks TTI – Transmission Time Interval UAS – Uplink EGPRS2-A Level Scheme UBS – Uplink EGPRS2-B Level Scheme UMA – Unlicensed Mobile Access UMB – Ultra Mobile Broadband UMTS – Universal Mobile Telecommunications System s – Microseconds UTRAN – UMTS Terrestrial Radio Access Network VDSL – Very High Speed DSL VoIP – Voice over Internet Protocol VOLGA – Voice over LTE Generic Access VPN – Virtual Private Network WAP – Wireless Application Protocol WCDMA – Wideband CDMA Wi-Fi – Wireless Fidelity WiMAX – Worldwide Interoperability for Microwave Access WLAN – Wireless Local Area Network WMAN – Wireless Metropolitan Area Network WRC-07 – World Radiocommunication Conference 2007 Additional Information 3G Americas maintains complete and current lists of market information including EDGE. 3G Americas: “UMTS Evolution from 3GPP Release 7 to Release 8.org. 2008. 3G Americas: “UMTS Evolution from 3GPP Release 7 to Release 8. July 2007. 2007. please call +1 425 372 8922 or e-mail Krissy Gochnour. Rysavy Research / 3G Americas.

submission to 3G Americas.” June 16. TX. No 3. http://www. 2008. ATIS 3GPP LTE Conference.” May 2007. January 26.com/ReportPDF/ProductSheet/BI-SOS-PS.org.cdg. Dow Jones NY: “3G Retains its Buzz but Potential Remains Unclear. September 2009 Page 114 . April 2008. February 2007. Mathias Riback. June 2006. Ericsson Review No. 3GPP LTE-Advanced Workshop. http://www. Ericsson: “The Evolution of LTE towards IMT-Advanced”. May 2009. 3. 2007.academypublisher. 2006. Ericsson: “Initial Field Performance Measurements of LTE.com/ericsson/corpinfo/publications/review/2008_03/files/LTE. submission to 3G Americas.” July 2007. 3. the Undisputed Choice for Mobile Broadband. Ericsson white paper: “HSDPA Performance and Evolution”. 2008.pdf Ericsson: Per Beming et al. http://www. Dallas. Rysavy Research / 3G Americas. submission to 3G Americas.com/jcm/vol04/no03/jcm0403146154. 2007. “LTE-SAE architecture and performance”. Ericsson Review No.pdf Ericsson: “HSPA and WiMAX Performance. 2009. AT&T: “Ideas on LTE-Advanced”. June 2009. Alcatel Lucent:”Technology Comparison”. May 2009.” Jonas Karlsson. Cisco: “Approaching the Zettabyte Era.” May 2009. 2008. Ericsson: “3GPP Improved UE Receiver Requirements. http://www. Alcatel Lucent: LTE Migration Scenarios. Ericsson white paper: “Basic Concepts of HSPA”. submission to 3G Americas. GSA: The Global mobile Suppliers Association.” June 2006. March 27.ericsson.” 2006. Berg Insight: Smartphone Operating http://www. “HSPA: Keys to a Successful Broadband Access Strategy”. AT&T: Tom Keathley. Systems. Sam Samra. Alcatel Lucent: “TV White Space. Ericsson: “The 3G Long-Term Evolution – Radio Interface Concepts and Performance Evaluation. 2009.gsacom. June 2008. July 2007. Ericsson Review No. June 5. 1. Senior Director-Technology Programs. submission to 3G Americas. Forecast Update. CDMA Developer Group: Deployments. HSPA to LTE-Advanced. 2008 Cisco: “Visual Networking Index: Global Mobile Data Traffic. Ericsson: Johan Bergman et al. Cingular Wireless: uplink throughput data. 2009. Ericsson: “HSPA Spectrum Efficiency Evolution”. Stefan Parkvall and David Astely. submission to 3G Americas.pdf.” January 29. Alcatel Lucent: LTE TDD Harmonization.berginsight.com. Arthur D Little: “HSPA and Mobile WiMAX for Mobile Broadband Wireless Access – An Independent Report Prepared for the GSM Association. submission to 3G Americas. submission to 3G Americas. “HSPA Evolution – Boosting the performance of mobile broadband access”. Ericsson white paper: “HSPA. Ericsson: “Cellular Evolution.” May 2006. 2006.Alcatel Lucent: “CDMA2000 Path to LTE”. CDMA Development Group. submission to 3G Americas. July 2007.” January 27.

” June 2006. 2000-2009 (4Q05 Update). IEEE: Communications Magazine.” Rajesh Chundury. submission to 3G Americas. June 2006. A detailed market assessment and forecast at the global." 2009 GSA: HSPA Devices Survey. Wireless Business Review. 2008. Ericsson: “GSM/EDGE Continued Evolution. CDMA Business View. 2008. submission to 3G Americas. No. Spring 2006.Ericsson: Erik Dahlman et al. IEEE: Journal on Selected Areas in Communication. 2. May 2009. http://www.com/press-releases/led-asiapacific-suppliers-cellular-modem-industry-will-exceed-200-million-units-2013.” Stefan Parkvall and David Astely.1. September 2009 Page 115 . Forward Concepts.” Ericsson Review No. Worldwide. Gartner Group: “Forecast: Mobile Terminals. Ericsson: “Mobile Broadband Backhaul: Addressing the Challenge. "LTE’s Five-Year Global Forecast: Poised to Grow Faster than 3G. press release. Rysavy Research / 3G Americas. Ericsson: “HSDPA performance . “Mobile data revenue to exceed US$200 billion in 2007.” Ericsson Review No. 2008. “Key features of the LTE radio interface”. IEEE International Symposium on Personal. Indoor and Mobile Radio Communications: Anders Furuskär et al “The LTE Radio Interface – Key Characteristics and Performance”.” January 12. supplied to 3G Americas. Vol 24. 2006. HSPA to LTE-Advanced. submission to 3G Americas. Mobile Internet Device and Chip Market Opportunities. Informa Telecoms & Media: World Cellular Information Service.” October. “An Analysis of VoIP Service Using 1 EV-DO Revision A System. Ericsson Review No. Roshni Srinivasan.” 2006. Ericsson white paper: “Technical Overview and Performance of HSPA and Mobile WiMAX”. Mo-Han Fong and Robert Novak. Nortel Networks. Intel Corporation. “Improved VoIP Capacity in Mobile WiMAX Systems Using Persistent Resource Allocation. Global Telecom Insider report. Huawei Technologies. Ericsson white paper: “WiMAX – Copper in the Air. Senza Fili Consulting: Press release of June 19. Informa Telecoms & Media. regional and country level (2006-2012)". 2008 Informa Telecoms & Media: WiMAX projection. June 2007. 2008 GSM Association: “Mobile phones on the catwalk” by Paul Rasmussen. June 2008.HSDPA and R99 on Same Carrier. 2009. Sean McBeath.” January 2006. 1. 2008. July 21.” April 2006. Ericsson: “Advanced Receivers for WCDMA Terminal Platforms and Base Stations. Ericsson Review No. Fierce Wireless: Press release. 2. 2007 describing the report "WiMAX: Ambitions and Reality. Qi Bi. Ericsson: HSPA voice migration. 2006.fiercewireless. Ericsson: “The Evolution of LTE towards IMT-Advanced. Ericsson: “WCDMA vs.” July 23. June 2009. 2006. 3. 2008.

Nokia: “VoIP over HSPA with 3GPP Release 7.” June 2006. June 11. submission to 3G Americas.” May 2006.” June 2006.com/story/16425. Feb 2009. HSPA VoIP Capacity Difference. Wiley. submission to 3G Americas.” www. Axel Klein. 2009.: “Mobile-Phone Premium Content Market to Reach $40 billion by 2010. Pyramid Research: “Europe to See Huge Growth in Mobile Broadband Services despite Recession. submission to 3G Americas. September 2009 Page 116 . HSPA to LTE-Advanced. submission to 3G Americas. Nokia: “Future Voice Traffic: Primary vs.cellular-news. 2007.” May 2009. Maravedis: “WiMAX and Broadband Wireless Access Equipment Market Analysis. Devices principal analyst.” June 2006.” 2009. Ovum Comment. LSTI Forum: "LTE/SAE Trial Initiative Latest Results from the LSTI”.” by Harri Holma. Arthur D. Qualcomm: press release. Harri Holma.” June 2007.” June 2006. 27 March 2007.” Harri Holma and Antti Toskala.” 2009. Nokia: “3GPP vs. Portio Research: “Mobile Data Services Markets 2008”. “Report ITU-R M. http://www. Nokia Siemens Networks: “HSPA/LTE Performance. OFDMA and SC-FDMA Based Radio Access. Little Limited: "HSPA and mobile WiMAX for Mobile Broadband Wireless Access". Ovum.html. 2008. 2008. 2009. “Smartphones: the silver lining of the declining handset market. May 15.” June 2009. LTE Report.com/press/releases/2008/080207_Qualcomm_to_Ship. Nokia Siemens Networks. March 8. Research in Motion: “Evolution of High Speed Wireless Data Standards in 3GPP”.” June 2006. Nokia: “WCDMA CS vs. et al. July 2. Nortel white paper: “GSM to LTE Evolution. Nokia: “HSUPA Simulation Results. 2009. "LTE for UMTS. Rysavy Research / 3G Americas. Adam Leach.php. Nokia: “SAE Evolved Architecture. July 2009. Juniper Research. Trends and Forecasts.” Danielle Morais de Andrade. 2009. Lehman Global Equity Research: Paul Wuh. Secondary?” June 2006 submission to 3G Americas. 2009-2014”. June 1.iSuppli Corp. Requirements related to technical performance for IMT-Advanced radio interface(s). “Global 3G Developments: 3G subs accelerate. submission to 3G Americas. 2006.qualcomm. submission to 3G Americas. Telecom and Software News. Nokia: “Overview on HSPA+. 2006. “Performance Evaluation on Dual-Cell HSDPA Operation.2134. submission to 3G Americas. more data revenue in ’09” May 23. Nokia: “HSDPA Performance Measurements with Commercial QPSK (CAT 12) and 16QAM (CAT 6). 3GPP2 Cellular VoIP Driver Comparison. International Telecommunications Union.

http://www. March 2008. Wireless Week: “One in Three Handsets Will Be a Smartphone http://www.” published by Datacomm Research. Wireless Intelligence and AT Kearney: news. Wikipedia: http://en. Value Partners: “Getting the Most Out of the Digital Divide – Allocating UHF Spectrum to Maximise the Benefits for European Society”.html.pdf.com/commentary2006/jan1806. Telstra: presentation.wirelessweek.aspx?id=158452. http://www1. http://www. July 2008. http://www.gov/ipc/www/idb/worldpopinfo. “Vodafone Trials HSPA+ Mobile Broadband at Speeds of Up To 16Mbps. Chetan Sharma: “US Wireless Data Market Update .spectrumstrategy. United 2009. July 31.asp. May 2007.census. “WiMAX: Mobilizing the Internet”. 2008.com/papers. Rysavy Research. 2005.outlook4mobility. 2008.wikipedia. “SNL Kagan Expects Wireless Data Revenue to Increase at a 16% Annual Rate over the Next Decade”. Rysavy Research / 3G Americas.com) and utilized a composite of statistical information from multiple resources. 2006 Verizon Wireless: Verizon Broadband Access Web page.html. “Mobile Broadband Spectrum Demand. March 5.com/story/31730.com/article.com/press/20080731. States Census Bureau. Vodafone: press release. Chetan Sharma: “Managing Growth and Profits in the Yottabyte Era”.Rysavy Research article: “Reach Me if You Can. 2009. July 2009. 2008 Sprint: Press release.php?source=newsletter. Andy Seybold: “Will Data-Only Networks Ever Make Money?” January 18.snl.” SNL Kagan: press release. HSPA to LTE-Advanced. Sprint Nextel: Ali Tabassi.org/wiki/World_population. http://www.” January 15. Rysavy Research: “Hard Numbers and Experts' Insights on Migration to 4G Wireless Technology.Q1 2009.” http://www.com/Pages/GB/perspectives/Spectrum-Getting-the-most-outof-the-digita-dividend-2008. 2008. commentary. July 29.htm. January 30. September 2009 Page 117 . February 2005. 2007. March 2008.rysavy.cellularby 2013”. This white paper was written for 3G Americas by Rysavy Research (http://www.rysavy.” December 2008. “HSPA as an Open Eco-System Today – Telstra Next G Network”. Fierce Wireless Webcast.

rysavy. Although Rysavy Research has exercised reasonable care in providing this information to you. analysis and conclusions of Rysavy Research and may not necessarily represent the comprehensive opinions and individual view points of each particular 3G Americas Board member company. or punitive arising out of or in any way related to the use of the information. incidental. September 2009 Page 118 . Rysavy Research provides this document and the information contained herein to you for informational purposes only. Rysavy Research disclaims and in no event shall be liable for any losses or damages of any kind.com HSPA to LTE-Advanced. Rysavy Research does not warrant that the information is error-free. consequential. Rysavy Research / 3G Americas. Rysavy Research provides this information solely on the basis that you will take responsibility for making your own assessments of the information. indirect.The contents of this paper reflect the research. All rights reserved. LLC. http://www. Copyright ©2009 Rysavy Research. whether direct.

Sign up to vote on this title
UsefulNot useful

Master Your Semester with Scribd & The New York Times

Special offer for students: Only $4.99/month.

Master Your Semester with a Special Offer from Scribd & The New York Times

Cancel anytime.