Ad Hoc Networks:#

Architectures & Protocols

Dr Ljiljana Simić
iNETS, RWTH Aachen University
SS2016

Important  Note:  
These   course   notes   may   contain   some   copyrighted  
material.   The   copyright   of   this   material   covers   its  
use   in   class   and   for   educa7onal   purposes,   but   you  
are   not   allowed   to   distribute   this   course   material  
freely.   Under   the   code   of   appropriate   use,   please  
refrain   from   uploading   the   provided   source   files   or  
documents  to  any  publicly  accessible  system  outside  
RWTH  Aachen  University  without  prior  permission.  

1
 –WSN II
#

Many wireless design myths

§  once  WSN  research  was  started,  we  learned  that  

many  “intui1vely  clear”  design  rules  were  myths  

§  the  same  lessons  apply  to  ad  hoc  networks,  of  course  

2
 Multi-hop advantage myth
§  distributed  WSN:  mul7-­‐hop  RF  advantage  
§  RF  power  aGenua7on  near  ground:  
§  Prx  ∝ Ptx/rk,      k=3-­‐5;        equivalently,  Ptx  ∝ rkPrx  
§  power  advantage:  
Ptx(Nr)/(NPtx(r))  =  (NkrkPrx)/(NrkPrx)=  Nk-­‐1  
 

§  BUT  also  non-­‐RF  transceiver  circuit  energy  consump7on!  

Sensor node energy roadmap

§  Rehos7ng  to  Low  
10,000   Power  COTS  (10x)  
   Deployed  (5W)  
Average  Power  (mW)  

§  System-­‐On-­‐Chip  
1,000  
§  Adv  Power  Management  
  Algorithms  (50x)  
 PAC/C  Baseline  (.5W)  
100  
   (50  mW)  
10  
     
1   (1mW)  
 
.1  
2000    2002    2004  

Source:  ISI  &  DARPA  PAC/C  Program  

3
4
 Comparison of energy sources
Power (Energy) Density Source of Estimates
Batteries (Zinc-Air) 1050 -1560 mWh/cm3 (1.4 V) Published data from manufacturers
3
Batteries(Lithium ion) 300 mWh/cm (3 - 4 V) Published data from manufacturers
2
15 mW/cm - direct sun
Solar (Outdoors) 0.15mW/cm2 - cloudy day. Published data and testing.
2
.006 mW/cm - my desk
2
Solar (Indoor) 0.57 mW/cm - 12 in. under a 60W bulb Testing
3
Vibrations 0.001 - 0.1 mW/cm Simulations and Testing
2
3E-6 mW/cm at 75 Db sound level
2
Acoustic Noise 9.6E-4 mW/cm at 100 Db sound level Direct Calculations from Acoustic Theory
Passive Human
2
Powered 1.8 mW (Shoe inserts >> 1 cm ) Published Study.
Thermal Conversion 0.0018 mW - 10 deg. C gradient Published Study.
3
80 mW/cm
3
Nuclear Reaction 1E6 mWh/cm Published Data.
3
300 - 500 mW/cm
3
Fuel Cells ~4000 mWh/cm Published Data.

with  aggressive  energy  management,  WSN  nodes  might  be  

powered  off  the  environment  
Source:  UCB  

Low Energy Adaptive Clustering Hierarchy

(LEACH)
[W.  R.  Heinzelman  et  al,  “Energy-­‐efficient  communica7on  protocol  for  wireless  
microsensor  networks”,  in  Proc.  33rd  Hawaii  Int.  Conf.  on  System  Sciences,  2000.]  
 

§  clustering  protocol  for  WSN  that  minimizes  energy  consump7on  

§  clustering  +  data  aggrega7on  +  MAC  (TDMA  inter-­‐cluster  &  
CDMA  each  CH  to  BS)  +  simple  rou7ng  protocol  

§  clustered  WSN  architecture:  

§  sensor  nodes  organize  into  clusters  
§  one  node  acts  as  “local  base  sta7on”  cluster-­‐head  (CH)  
§  nodes  send  messages  to  their  local  CH  
§  CH  sends  messages  to  data-­‐gathering  base  sta7on  (AP  ac7ng  as  external  
gateway)  
§  useful  for  WSN  due  to  inherent  suitability  for  data  fusion    
§  in  self-­‐organizing  WSN,  cluster  forma1on  &  elec1on  of  CH  should  be  a  
distributed,  autonomous  process  

5
 LEACH
§  randomly  selects  nodes  as  CH  
 
§  performs  periodic  re-­‐elec7on  
èhigh-­‐energy  dissipa7on  of  CHs  spread  across  network  
 
§  each  itera7on  of  CH  selec7on  =  “round”  
 
§  a  round  split  into  2  phases:  
1.)  set-­‐up  phase  
2.)  steady  phase  

LEACH
1.)  set-­‐up  phase  
§  each  node  chooses  random  number  in  range  [0-­‐1]  
§  node  becomes  CH  if  this  random  number  <  T(n)  
§  threshold  for  node  n:  
{  
T(n)  =          P/(1-­‐P[r  x  mod(1/P)])
                       0    
 if  n  ∈  G  
     otherwise  
§  P  =  desired  %  CH  nodes  
§  r  =  current  round  
§  G  =  set  of  nodes  which  have  not  been  CH  in  past  1/P  rounds  
§  ensures  all  nodes  eventually  spend  equal  energy  
 
§  ager  selec7on,  CHs  adver7se  to  all  nodes  
§  nodes  choose  their  nearest  CH,  based  on  the  
adver1sement  heard  with  max  signal  strength  
§  CH  assigns  TDMA  schedule  for  cluster  members  

6
 LEACH
2.)  steady  phase  
§  longer  dura7on,  to  minimise  O/H  of  cluster  forma7on  
§  data  transmission  within  each  cluster  using  TDMA  
§  enables  nodes  to  sleep  
§  CH  perform  data  aggrega7on/fusion  
§  BS  receives  only  aggregated  data  from  CHs  
§  CHs  use  unique  CDMA  codes  to  avoid  inter-­‐cluster  
interference  
§  ager  certain  period  of  7me,  CHs  re-­‐elected  via  set-­‐up  phase  

50 1.1

45 1

40 0.9

Normalized energy dissipation

35 0.8

30 0.7

LEACH 25

20
0.6

0.5

§  dynamic  cluster  forma7on:   15 0.4

[Heinzelman  2000]  
10 0.3

5 0.2

  0
−25 −20 −15 −10 −5 0 5 10 15 20 25
0.1
0 10 20 30 40 50 60
Percent of nodes that are clus
50 50 1.1

45 45 1

40 40 0.9 Figure 8. Normalized total sy

sipated versus the percent o
Normalized energy dissipation

35 35 0.8
cluster-heads. Note that dir
30 30 0.7
is equivalent to 0 nodes bei
25 25 0.6
or all the nodes being cluste
20 20 0.5

15 15 0.4

Direct Trans
10 10 0.3
LEACH
5 5 0.2
in the system to be large. If there ar
0
−25 −20 −15 −10 −5 0 5 10 15 20 25
0
−25 −20 −15
0.1
−100
heads,
−510
100 020
the distance nodes have to tran
530 1040 1550 2060 2570 80 90
Percent of nodes that are cluster heads
50
est cluster-head does not reduce sub
45 more cluster-heads that have to tran
40
Figure 8.
Figure 7. Dynamic Normalized
clusters: total system energydistances
(a) cluster-head dis- to the base station, and the
nodes = sipated
at time versus
(b) the percentnodes
cluster-head of nodes thatbeing are performed locally. For our s
35
= at timecluster-heads.
. All nodes Note that direct
marked with transmission
a topology, %.
30
given symbol is belong
equivalentto theto same
0 nodes being
cluster, andcluster-heads Figure 8 also shows that LEACH
25
the cluster-heador allnodes
the nodes being cluster-heads.
are marked with a . tor of 7 reduction in energy dissipat
20
communication with the base station
15 mal number of cluster-heads. The m
10 the LEACH protocol is due to combi
network and thus nosystem
extra negotiation with the data routing. There is clea
5
in the to be large.isIfrequired
there aretomore deter-than cluster-
mine the cluster-heads. the quality of the output and the a
0 heads, the distance nodes have to transmit to reach the near-
−25 −20 −15 −10 −5 0 5 10 15 20 25
The systemestcancluster-head
determine, does
a priori, achieved. In this case, some data fr
not thereduceoptimal number yet there
substantially, are 7 in a subs
of clusters to have nals is lost, but this results
morein the system. that
cluster-heads Thishave will todepend
transmit on data
sev- the long-haul
eral parameters, such astothethe
network topology overall energy dissipation of the syst
Figure 7. Dynamic clusters: (a) cluster-head distances base station, andand theretheisrela-
less compression
nodes = at time tive costs of computation
(b) cluster-head nodes versuslocally.
being performed communication.
For our system We sim- parametersWeandsimulated LEACH (with 5%
= at time ulated
. All nodes marked with a the LEACH protocol for%.
topology, the random network shown cluster-heads) using MATLAB wit
given symbol belong to the same cluster,inand Figure 3 using the radio parameters in Table 1 and a com-
Figure 8 also shows that LEACH can achieve over a fac-
shown in Figure 3. Figure 9 shows
the cluster-head nodes are marked with aputation
. cost of 5 nJ/bit/message to fuse 2000-bit messages
tor of 7 reduction in energy dissipation compared to direct
compare using nJ/bit as
while varying the percentage of total nodes that are cluster- work is increased. This plot shows
 LEACH
1

0.9

Direct §  energy  reduc7on  due  to:  

1
Total energy dissipated in system (Joules)

Total energy dissipated in system (Joules)

0.8
MTE 0.8

0.7 LEACH §  local  communica1on  within  cluster   0.6 Direct −>
0.6
§  data  aggrega1on  at  CH   0.4

§  dras7cally  reduces  number  of  messages  that  must  be  

0.5

0.2 LEACH

sent  on  the  longer  communica7on  link  to  BS  

0.4 v|

0
0.3 1
0.8 200
0.2 0.6 150
−7
x 10 0.4

§  useful  network  life1me  increased  by  distribu7ng  energy  

100
0.1 0.2 50
0 0

dissipa7on  evenly  through  network  

Electronics energy (Joules/bit)
0 Network diameter (m)
0 20 40 60 80 100 120 140 160 180 200
Network diameter (m)

 
Figure 9. Total system energy dissipated us-
Total energy dissipated in system (Joules)

0.6

ing direct communication, MTE routing and 0.5

LEACH for the 100-node random network 0.4 <− MTE

shown in Figure 3. nJ/bit, 0.3

pJ/bit/m , and the messages are 2000 bits. 0.2

0.1 LEACH
v|
0
1
0.8 200
0.6 150
−7
x 10 0.4 100
In addition to reducing energy dissipation, LEACH suc- 0.2 50
0
cessfully distributes energy-usage among the nodes in the Electronics energy (Joules/bit) 0
Network diameter (m)

network such that the nodes die randomly and at essentially

the same rate. Figure 11 shows a comparison of system Figure 10. Total system energy dissipated
lifetime using LEACH versus direct communication, MTE using (a) direct communication and LEACH
routing, and a conventional static clustering protocol, where
the cluster-heads and associated clusters are chosen initially
and (b) MTE routing and LEACH for the ran- LEACH
dom network shown in Figure 3.
and remain fixed and data  fusion is performed at the cluster- pJ/bit/m , and the messages are 2000 bits.
heads, for the network energy  
§  shown &  network  
in Figure 3. For this life7me  
exper- performance:  
[Heinzelman  
iment, each node was initially given 0.52000]  
J of energy. Fig-
ure 11 shows that LEACH more than doubles the useful sys-
100 1
tem lifetime compared with the alternative approaches. We
ran similar experiments with different energy thresholds and 90 0.9
1
Direct
Total energy dissipated in system (Joules)

Total energy dissipated in system (Joules)

found that no matter how much energy MTE each node is given, 80 0.8
0.8

it takes approximately 8 times longerLEACH for the first node to 70 0.7

Direct
Number of sensors still alive

0.6 Direct −>

die and approximately 3 times longer for the last node to 60 0.6

die in LEACH as it does in any of the other protocols. The MTE 0.4

50 0.5
data from these experiments is shown in Table 2. The ad- Static Clus 0.2 LEACH
0.4 v|
vantage of using dynamic clustering (LEACH) versus static 40
LEACH 0
clustering can be clearly seen in Figure 11. Using a static 30 0.3 1
0.8 200
clustering algorithm, as soon as the cluster-head node dies, 20 0.2
−7
0.6 150
x 10 0.4
all nodes from that cluster effectively die since there is no 10 0.1 0.2 50
100

way to get their data to the base station. While these simu- 0 0
Electronics energy (Joules/bit)
0 0
Network diameter (m)
0 20 40 60 80 100 120 140 160 180 200
lations do not account for the setup time Networkto configure
diameter (m) the 0 200 400 600 800
Time steps (rounds)
1000 1200 1400

dynamic clusters (nor do they account for any necessary

routing start-up costs or updates
Figure 9. Totalassystem
nodes die), energy they dissipated
give us-
Figure 11. System lifetime using direct trans-
Total energy dissipated in system (Joules)

0.6
a good first order ing
approximation of the lifetime extension
direct communication, MTE routing and mission, MTE routing, static clustering, and 0.5
we can achieve using LEACH.
LEACH for the 100-node random network LEACH with 0.5 J/node. 0.4 <− MTE

Another important
shownadvantage of LEACH,
in Figure 3. illustratednJ/bit,
in 0.3

pJ/bit/m
Figure 12, is the fact that nodes ,die
andinthe messages
essentially are 2000 bits.
a “ran- 0.2

0.1 LEACH
v|
0
7 1
0.8 200

8
0.6 150
−7
x 10 0.4 100
In addition to reducing energy dissipation, LEACH suc- 0.2 50
0
cessfully distributes energy-usage among the nodes in the Electronics energy (Joules/bit) 0
Network diameter (m)

network such that the nodes die randomly and at essentially

the same rate. Figure 11 shows a comparison of system Figure 10. Total system energy dissipated
lifetime using LEACH versus direct communication, MTE using (a) direct communication and LEACH
routing, and a conventional static clustering protocol, where and (b) MTE routing and LEACH for the ran-
 – Radios for Ad Hoc Networks
#

Standards vs. non-standards

§  boils  down  quickly  to  the  tough  ques7on  on  the  value  of  standards  
and  availability  of  a  mass-­‐market  (read:  cheap)  chipsets  
§  like  always,  no  simple  answer  
§  standards:  
§  IEEE  802.11a,b,g  (PHY),  IEEE  802.11e,k,  (MAC)  …  
§  IEEE  802.15.4  &  ZigBee  
§  Bluetooth  
§  Bluetooth-­‐LE  
§  non-­‐standards:  
§  cc2400-­‐series  (Texas  Instruments)  
§  ultra-­‐low  power,  burst  radio  designs,…  

9
 IEEE 802.11: architecture of an
infrastructure network
802.11  LAN   §  sta1on  (STA)  
802.x  LAN   terminal  with  access  mechanisms  to  
wireless  medium  &  radio  contact  to  AP  
STA1   §  basic  service  set  (BSS)  
BSS1  
Portal   group  of  sta7ons  using  the  same  radio  
Access   frequency  (&  same  AP)  
 Point  
§  access  point  (AP)  
Distribu1on  System  
sta7on  integrated  into  the  wireless  LAN  
ESS   Access   &  the  distribu7on  system  
 Point  
§  portal  
BSS2   bridge  to  other  (wired)  networks  
§  distribu1on  system  
interconnec7on  network  to  form  one  
logical  network  (EES:  extended  service  
STA2   802.11  LAN   STA3   set)  based  on  several  BSS  

IEEE 802.11: architecture of an#

ad-hoc network
direct  communica7on  within  
802.11  LAN   a  limited  range  
 

STA1   §  sta1on  (STA)  

IBSS1   STA3   terminal  with  access  mechanisms  
to  wireless  medium    
§  independent  basic  service  set  
STA2   (IBSS)  
group  of  sta7ons  using  the  same  
radio  frequency    

IBSS2  

STA5  
STA4   802.11  LAN  

10
 IEEE 802.11 standard
fixed  
mobile  terminal   terminal  

infrastructure  
network  
access  point  
applica7on   applica7on  
TCP   TCP  
IP   IP  
LLC   LLC   LLC  
802.11  MAC   802.11  MAC   802.3  MAC   802.3  MAC  
802.11  PHY   802.11  PHY   802.3  PHY   802.3  PHY  

IEEE 802.11: protocol stack

11
 IEEE 802.11
   
§  operates  in  unlicensed  bands  (2.4  GHz  &  5  GHz)  
 

IEEE 802.11: physical (PHY) layer

§   3  versions  in  original  standard  
§  (FHSS  or  DSSS)  radio  in  2.4  GHz  ISM  band  
§  infrared    
§  data  rates  1  or  2  Mbps  
§  FHSS  (Frequency  Hopping  Spread  Spectrum)  
§  two-­‐level  GFSK  modula7on  
§  typically  1  Mbps,  min.  2.5  frequency  hops/s  (USA)  
§  DSSS  (Direct  Sequence  Spread  Spectrum)  
§  DBPSK  modula7on  for  1  Mbps  (Differen7al  Binary  Phase  Shig  Keying),  
DQPSK  for  2  Mbps  (Differen7al  Quadrature  PSK)  
§  preamble  and  header  of  a  frame  is  always  transmiGed  with  1  Mbps  
§  chip  sequence:  +1,  -­‐1,  +1,  +1,  -­‐1,  +1,  +1,  +1,  -­‐1,  -­‐1,  -­‐1  (Barker  code)  
§  IR  (infrared)  
§  850-­‐950  nm,  diffuse  light,  typ.  10  m  range  

12
 IEEE 802.11: physical (PHY) layer
§  IEEE  802.11a  in  5  GHz  unlicensed  band  
§  OFDM  (orthogonal  frequency  division  mul7plexing)  
§  data  rate:  6,  9,  12,  18,  24,  36,  48,  54  Mbps,  depending  on  SNR  
§  frequency  
§  5.15-­‐5.25,  5.25-­‐5.35,  5.725-­‐5.825  GHz  unlicensed  band  
§  5  GHz  band  less  crowded  
§  but  higher  frequency  =  shorter  transmission  range  vs.  2.4  GHz  

§  IEEE  802.11g  

§  essen7ally  IEEE802.11a  at  2.4  GHz  
 
§  IEEE  802.11n  (ac,  ad,  …)  
§  adds  MIMO,  wider  channels,  higher  PHY  rates  
§   forget  it  for  very  low-­‐power,  mobile  applica7ons!  
 

IEEE 802.11: MAC layer

§  traffic  services  
§  asynchronous  data  service  (mandatory)  
§  exchange  of  data  packets  based  on  “best-­‐effort”  
§  support  of  broadcast  and  mul7cast  
§  7me-­‐bounded  service  (op7onal)  
§  implemented  using  PCF  (Point  Coordina7on  Func7on)    

13
 IEEE 802.11: MAC layer
access  methods:  
§  DFWMAC-­‐DCF  CSMA/CA  (mandatory)  
§  collision  avoidance  via  randomized  “back-­‐off”  mechanism  
§  minimum  distance  between  consecu7ve  packets  
§  ACK  packet  for  acknowledgements  (not  for  broadcasts)  
§  DFWMAC-­‐DCF  w/  RTS/CTS  (op7onal)  
§  avoids  hidden  terminal  problem  
§  DFWMAC-­‐  PCF  (op7onal)  
§  access  point  polls  terminals  according  to  a  list  

-  DFWMAC  =  distributed  founda7on  wireless  MAC  

-  DCF  =  distributed  coordina7on  func7on  
-  PCF  =  point  coordina7on  func7on  

IEEE 802.11: MAC layer

priori1es:  
§  defined  through  different  inter  frame  spaces  
§  no  guaranteed,  hard  priori7es  
§  SIFS  (Short  Inter  Frame  Spacing)  
§  highest  priority,  for  ACK,  CTS,  polling  response  
§  PIFS  (PCF  IFS)  
§  medium  priority,  for  7me-­‐bounded  service  using  PCF  
§  DIFS  (DCF,  Distributed  Coordina1on  Func1on  IFS)  
§  lowest  priority,  for  asynchronous  data  service  

DIFS   DIFS  
PIFS  
medium  busy   SIFS   conten1on   next  frame  
direct  access  if     t  
medium  is  free  ≥  DIFS  

14
 IEEE 802.11: CSMA/CA method I
conten7on  window  
DIFS   DIFS   (randomized  back-­‐off  
mechanism)  
medium  busy   next  frame  
direct  access  if     t  
medium  is  free  ≥  DIFS   slot  7me  

§  sta7on  ready  to  send  starts  sensing  the  medium  

§  physical  carrier  sense  based  on  CCA,  Clear  Channel  Assessment  
§  if  medium  free  for  dura1on  of  IFS  
§  sta7on  can  start  sending  (IFS  depends  on  service  type)  
§  if  medium  busy  
§  sta7on  has  to  wait  for  a  free  IFS,  then  must  addi1onally  wait  a  random  
back-­‐off  1me  (collision  avoidance,  mul7ple  of  slot-­‐7me)    
§  if  another  sta7on  occupies  the  medium  during  the  back-­‐off  7me  of  the  
sta7on,  the  back-­‐off  7mer  stops  (fairness)  

IEEE 802.11: competing stations (simple version)

DIFS   DIFS   DIFS   DIFS  
sta1on1   boe   bor   boebo
  r   boe   busy  

sta1on2   boe   busy  

sta1on3   busy  

sta1on4   boe  busy   boe   bor  

sta1on5   boe  bor   boe   busy   boe   bor  

t  

busy   medium  not  idle  (frame,  ack  etc.)     boe   elapsed  backoff  7me  

packet  arrival  at  MAC   bor   residual  backoff  7me  

15
 IEEE 802.11: CSMA/CA method II
sending  unicast  packets  
§  sta7on  has  to  wait  for  DIFS  before  sending  data  
§  receivers  acknowledge  at  once  (aaer  wai1ng  for  SIFS)  if  the  packet  
was  received  correctly  (CRC)  
§  automa7c  retransmission  of  data  packets  in  case  of  transmission  
errors  

sender   DIFS  
data  

receiver   SIFS  
ACK  
other   DIFS  
sta1ons   data  
t  
wai7ng  7me   conten7on  

IEEE 802.11: DCF w/ RTS/CTS

sending  unicast  packets  
§  sta7on  can  send  RTS  with  reserva7on  parameter  ager  wai7ng  for  DIFS  
(determines  amount  of  7me  the  data  packet  needs  the  medium)    
§  acknowledgement  via  CTS  ager  SIFS  by  receiver  (if  ready  to  receive)  
§  sender  can  now  send  data  at  once,  acknowledgement  via  ACK  
§  other  sta7ons  store  medium  reserva7ons  distributed  via  RTS  and  CTS    
         i.e.  via  NAV  (virtual  carrier  sense)  

sender   DIFS  
RTS   data  
SIFS   SIFS  
receiver   CTS   SIFS   ACK  

other  
sta1ons   NAV  (RTS)   DIFS  
NAV  (CTS)   data  
t  
defer  access   conten7on  

16
 Bluetooth
idea:  
§  universal  radio  interface  for  ad-­‐hoc  wireless  connec7vity  
§  interconnec7ng  computer  and  peripherals,  handheld  
devices,  PDAs,  cell  phones  –  replacement  of  IrDA  
§  embedded  in  other  devices,  goal:  5€/device    
§  short  range  (10  m),  low  power  consump7on,  license-­‐exempt  
2.45  GHz  ISM  
§  voice  and  data  transmission,  about  1  Mbps  gross  data  rate  

One  of  the  first  modules  

(Ericsson)  

17
 Bluetooth
history:  
§  1994:  Ericsson  (Mavson/Haartsen),  “MC-­‐link”  project  
§  renaming  of  the  project:  Bluetooth  ager  Harald  “Blåtand”  Gormsen  
[son  of  Gorm],  King  of  Denmark  in  the  10th  century  
§  1998:  founda1on  of  Bluetooth  SIG,  www.bluetooth.org  
§  1999:  erec7on  of  a  rune  stone  at  Ericsson/Lund  ;)  
§  2001:  first  consumer  products  for  mass  market,  spec.  version  1.1  
released  

§  Special  Interest  Group  

§  (founding)  Ericsson,  Intel,  IBM,  Nokia,  Toshiba;  (added)  3Com,  
Agere  (was:  Lucent),  Microsog,  Motorola  
§  >  2500  members  
§  common  specifica7on  and  cer7fica7on  of  products  

History and hi-tech …

1999:  
Ericsson mobile communications
AB reste denna sten till minne av
Harald Blåtand, som fick ge sitt
namn åt en ny teknologi för
trådlös, mobil kommunikation.

18
 …and the real rune stone
§  located  in  Jelling,  Denmark,  erected  by  
King  Harald  “Blåtand”  in  memory  of  his  
parents  
§  stone  has  three  sides  –  one  side  
showing  a  picture  of  Christ  
this  could  be  the  original  
colours  of  the  stone  
 
Inscrip7on:   “auk tani karthi
kristna”
"Harald king executes
(and made the
these sepulchral Danes Christians)
monuments after Gorm,
his father and Thyra, his
mother. The Harald who
won the whole of
Denmark and Norway and
turned the Danes to
Christianity."

Bluetooth: basic characteristics

§  2.4  GHz  ISM  band,  79  RF  channels,  1  MHz  carrier  spacing  
§  channel  0:  2402  MHz  …  channel  78:  2480  MHz  
§  G-­‐FSK  modula7on,  1-­‐100  mW  transmit  power  
§  FHSS  &  TDD  
§  frequency  hopping  with  1600  hops/s  
§  hopping  sequence  in  a  pseudo-­‐  random  fashion,  determined  by  master  
§  7me  division  duplex  for  send/receive  separa7on  
§  voice  link  –  SCO  (Synchronous  Connec1on  Oriented)  
§  FEC  (forward  error  correc7on),  no  retransmission,  64  kbps  duplex,  
point-­‐to-­‐point,  circuit  switched  
§  data  link  –  ACL  (Asynchronous  Connec1on-­‐Less)  
§  asynchronous,  fast  acknowledge,  point-­‐to-­‐mul7point,  up  to  433.9  kbps  
symmetric  or  723.2/57.6  kbps  asymmetric,  packet  switched  
§  topology  
§  overlapping  piconets  (stars)  forming  a  scaiernet  

19
 Bluetooth: piconet
§  collec7on  of  devices  connected  in  ad  hoc  
fashion   P  
S  
§  one  unit  acts  as  master  and  the  others   S  
as  slaves  for  the  life7me  of  the  piconet   M   P  

§  master  determines  hopping  paiern,   SB   S  

slaves  have  to  synchronize   P   SB  
§  each  piconet  has  a  unique  hopping  paiern  

§  par7cipa7on  in  a  piconet  =  synchroniza7on   M=Master   P=Parked  

to  hopping  sequence   S=Slave   SB=Standby  

§  each  piconet  has  one  master  and  up  to  

7  simultaneous  slaves  
(>  200  could  be  parked)  

Bluetooth: forming a piconet

§  all  devices  in  a  piconet  hop  together  
§  master  gives  slaves  its  clock  and  device  ID  
§  hopping  paGern:  determined  by  device  ID  (48  bit,  unique  worldwide)  
§  phase  in  hopping  paGern  determined  by  clock  
§  addressing  
§  Ac7ve  Member  Address  (AMA,  3  bit)  
§  Parked  Member  Address  (PMA,  8  bit)  
»  
·   P   »  
SB   ¸   S  
SB   »  
½   S   »   »  
SB   »   ¿   M   P  
SB   SB  
¸   »  
¸   ¾   SB   S  
SB   SB   »   Á  
¹   Á   P   SB  
SB   SB  

20
 Bluetooth: scatternet
§  linking  of  mul7ple  co-­‐located  piconets  through  the  
sharing  of  common  master  or  slave  devices  
§  devices  can  be  slave  in  one  piconet  and  master  of  another  
§  communica7on  between  piconets  
§  devices  jumping  back  and  forth  between  the  piconets  
piconets  

P  
S   S  
S  
P  
P  
M  
M=Master   M  
S=Slave   SB   S  
P=Parked  
SB=Standby  
P   SB   SB  
S  

Bluetooth protocol stack

audio  apps.   NW  apps.   vCal/vCard   telephony  apps.   mgmnt.  apps.  
TCP/UDP   OBEX  
AT  modem  
IP  
commands   TCS  BIN   SDP  
BNEP   PPP   Control  
RFCOMM  (serial  line  interface)  
Audio   Logical  Link  Control  and  Adapta7on  Protocol  (L2CAP)   Host  
Controller  
Link  Manager   Interface  
Baseband  
Radio  
AT:  aGen7on  sequence   SDP:  service  discovery  protocol  
OBEX:  object  exchange   RFCOMM:  radio  frequency  comm.  
TCS  BIN:  telephony  control  protocol  specifica7on  –  binary  
BNEP:  Bluetooth  network  encapsula7on  protocol  

21
 Bluetooth protocol stack

IEEE  802.15  version  of  the  Bluetooth  protocol  architecture  

 

Bluetooth#
frequency selection during data transmission
625  µs  
fk   fk+1   fk+2   fk+3   fk+4   fk+5   fk+6  
M   S   M   S   M   S   M  
t  

fk   fk+3   fk+4   fk+5   fk+6  

M   S   M   S   M  
t  
fk   fk+1   fk+6  
M   S   M  
t  

22
 Bluetooth: baseband layer
§  piconet/channel  defini7on  
§  low-­‐level  packet  defini7on  
§  access  code  
§  channel,  device  access,  e.g.  derived  from  master  
§  packet  header  
§  1/3-­‐FEC,  ac7ve  member  address  (broadcast  +  7  slaves),  
link  type,  alterna7ng  bit  ARQ/SEQ,  checksum    
68(72)   54   0-­‐2745   bits  
access  code   packet  header   payload  

4   64   (4)   3   4   1   1   1   8   bits  
preamble   sync.   (trailer)   AM  address   type   flow   ARQN   SEQN   HEC  

Bluetooth: SCO payload types

payload  (30)  

HV1   audio  (10)   FEC  (20)  

HV2   audio  (20)   FEC  (10)  

HV3   audio  (30)  

DV   audio  (10)   header  (1)  payload  (0-­‐9)   2/3  FEC   CRC  (2)  
(bytes)  

23
 Bluetooth: ACL payload types
payload  (0-­‐343)  

header  (1/2)   payload  (0-­‐339)   CRC  (2)  

DM1   header  (1)  payload  (0-­‐17)   2/3  FEC   CRC  (2)  

DH1   header  (1)  payload  (0-­‐27)   CRC  (2)   (bytes)  

DM3   header  (2)   payload  (0-­‐121)   2/3  FEC   CRC  (2)  

DH3   header  (2)   payload  (0-­‐183)   CRC  (2)  

DM5   header  (2)   payload  (0-­‐224)   2/3  FEC   CRC  (2)  

DH5   header  (2)   payload  (0-­‐339)   CRC  (2)  

AUX1   header  (1)  payload  (0-­‐29)  

Bluetooth: baseband data rates

 Payload  User      Symmetric  Asymmetric  
ACL    Header  Payload      max.  Rate    max.  Rate  [kbit/s]  
Type  [byte]  [byte]  FEC  CRC  [kbit/s]  Forward  Reverse    

1  slot   DM1  1  0-­‐17  2/3  yes  108.8  108.8  108.8    

DH1  1  0-­‐27  no  yes  172.8  172.8  172.8    

3  slot   DM3  2  0-­‐121  2/3  yes  258.1  387.2  54.4    

DH3  2  0-­‐183  no  yes  390.4  585.6  86.4    
DM5  2  0-­‐224  2/3  yes  286.7  477.8  36.3    
5  slot  
DH5  2  0-­‐339  no  yes  433.9  723.2  57.6  
AUX1  1  0-­‐29  no  no  185.6  185.6  185.6  
HV1  na  10  1/3  no  64.0    
SCO   HV2  na  20  2/3  no  64.0    
HV3  na  30  no  no  64.0    
DV  1  D  10+(0-­‐9)  D  2/3  D  yes  D  64.0+57.6  D      

Data  Medium/High  rate,  High-­‐quality  Voice,  Data  and  Voice  

24
 Bluetooth: baseband link types
§  polling-­‐based  TDD  packet  transmission  
§  625µs  slots,  master  polls  slaves  
§  SCO  (Synchronous  Connec7on  Oriented)  –  voice    
§  periodic  single  slot  packet  assignment,  64  kbit/s  full-­‐duplex,  point-­‐to-­‐point  
§  ACL  (Asynchronous  Connec7onLess)  –  data    
§  variable  packet  size  (1,3,5  slots),  asymmetric  bandwidth,  point-­‐to-­‐mul7point  
SCO   ACL   SCO   ACL   SCO   ACL   SCO   ACL  
MASTER   f0   f4   f6   f8   f12   f14   f18   f20  

SLAVE  1   f1   f7   f9   f13   f19  

SLAVE  2   f5   f17   f21  

Bluetooth: robustness
§  slow  frequency  hopping  with  paGerns  determined  by  master  
§  protec7on  from  interference  on  certain  frequencies  
§  separa7on  from  other  piconets  (FH-­‐CDMA)  
§  retransmission:  ACL  only,  very  fast  
§  Forward  Error  Correc7on   error  in  payload  
§  SCO  and  ACL   (not  header!)  
NAK   ACK  
A   C   C   F   H  
MASTER  

SLAVE  1   B   D   E  

SLAVE  2   G   G  

25
 Baseband states of a Bluetooth device
standby   unconnected  

detach   inquiry   page   connec1ng  

transmit   connected  
ac1ve  
AMA   AMA  

park   hold   sniff  

low  power  
PMA   AMA   AMA  

§  standby:  do  nothing   §  park:  release  AMA,  get  PMA    
§  inquire:  search  for  other  devices   §  sniff:  listen  periodically,  not  each  slot  
§  page:  connect  to  a  specific  device   §  hold:  stop  ACL,  SCO  s7ll  possible,  possibly    
§  connected:  par7cipate  in  a  piconet    par7cipate  in  another  piconet  

IEEE 802.15.4
§  IEEE  802.15.4  task  group  began  to  develop  a  standard  for  
LR-­‐WPAN  (low  rate  wireless  personal  area  network)  

§  goal  of  group  was  to  provide  a  standard  …  

§  with  ultra-­‐low  complexity,  cost,  and  power  
§  for  low-­‐data-­‐rate  wireless  connec7vity  
§  among  inexpensive  fixed,  portable,  moving  devices  

26
IEEE 802.15.4: general characteristics

IEEE 802.15.4: approaches for low power

to  achieve  the  low  power  and  low  cost  goals  established  by  
IEEE  802.15.4  ...  
§  reduce  the  amount  of  data  transmiGed    
§  reduce  the  transceiver  duty  cycle  and  frequency  of  data  
transmissions    
§  reduce  the  frame  overhead    
§  reduce  complexity    
§  reduce  range    
§  implement  strict  power  management  mechanisms  (power-­‐
down  and  sleep  modes)    

27
 IEEE 802.15.4 & ZigBee
§  IEEE  802.15.4    deals  with  only  PHY  layer  and  por1on  of  
data  link  layer  

§  higher-­‐layer  protocols  are  leg  to  industry  and  the  

individual  applica7ons  

§  the  ZigBee  Alliance  is  an  associa7on  of  companies  

involved  with  building  higher-­‐layer  standards  based  on  
IEEE  802.15.4  
§  includes  network,  security,  and  applica7on  protocols  

IEEE 802.15.4 in ISO-OSI layered network model

28
 IEEE 802.15.4: network layer
§  services  which  network  layer  provides  are  more  challenging  
to  implement  because  of  low  power  consump7on  
requirement  

§  network  layer  protocols  over  this  standard  are  expected  to  
be  self-­‐configuring  and  self-­‐maintaining  to  minimize  total  
cost  of  user  

§  IEEE  802.15.4  standard  supports  mul1ple  network  

topologies  including  star  and  peer-­‐to-­‐peer  topology  

§  topology  selec7on  is  applica7on  dependent  

§  PC  peripherals  may  require  low  latency  connec7on  of  star  topology  
§  perimeter  security  which  needs  large  coverage  area  may  require  
peer-­‐to-­‐peer  networking  

IEEE 802.15.4:#
star and peer-to-peer topologies

29
 IEEE 802.15.4: data link layer
§  IEEE  802  splits  DLL  into  MAC  &  LLC  sublayers.  

§  LLC  is  standardized  and  is  common  in  802.3,  802.11,  802.15.1  

§  features  of  the  IEEE  802.15.4  MAC:  

§  associa7on  and  disassocia7on  
§  acknowledged  frame  delivery  
§  channel  access  mechanism  
§  frame  valida7on  
§  guaranteed  7me  slot  management  
§  beacon  management  

IEEE 802.15.4: MAC

§  provides  data  and  management  services  to  upper  layers  

§  MAC  management  service  has  26  primi7ves  

§  cf.  802.15.1  (Bluetooth)  about  131  primi7ves  
 
§  very  low  complexity  MAC  
§  very  suitable  for  its  intended  low-­‐end  applica1ons  
§  at  the  cost  of  a  smaller  feature  set  than  Bluetooth  
§  e.g.  802.15.4  does  not  support  synchronous  voice  links  

30
 IEEE 802.15.4: MAC frame format

IEEE 802.15.4: MAC frame format

§  frame  control  field  
§  indicates  the  type  of  MAC  frame  being  transmiGed  
§  specifies  the  format  of  the  address  field  
§  controls  the  acknowledgment

§  mul1ple  address  types:  

§  64  bit  physical  address  
§  short  16  bit  network  assigned  address  

§  address  field  size  may  vary  from  0  to  20  bytes  

§  payload  field  is  variable  

§  with  condi7on  size  of  MAC  frame  <=  127  bytes  
 

§  FCS  is  used  for  integrity  check  using  16  bit  CRC  

31
 IEEE 802.15.4: superframe
§  certain  applica7ons  require  dedicated  bandwidth  to  achieve  
low  latency  
§  can  operate  in  op7onal  superframe  mode  

§  PAN  coordinator  transmits  superframe  beacons  in  

predetermined  intervals  which  are  divided  into  16  1me  slots  

§  channel  access  in  the  1me  slots  is  conten1on-­‐based  

§  but  PAN  coordinator  may  assign  7me  slots  to  a  single  
device  requiring  dedicated  bandwidth  or  low-­‐latency  
transmissions  
§  assigned  7me  slots  are  called  guaranteed  1me  slots  (GTS)  
and  together  form  a  conten7on-­‐free  period  

IEEE 802.15.4: superframe structure

32
 IEEE 802.15.4: other MAC features
§  in  a  beacon-­‐enabled  network  with  superframes,  
sloied  CSMA-­‐CA  is  used  

§  in  others,  standard  CSMA-­‐CA  is  used  

§  i.e  it  first  checks  if  another  device  is  transmivng  in  the  same  channel,  if  
so  backs  off  for  certain  7me  
§  MAC  confirms  successful  recep7on  of  data  with  an  
acknowledgement  

§  provides  for  three  levels  of  security:  

§  no  security  of  any  type  
§  access  control  lists  (non  cryptographic  security)  
§  symmetric  key  security,  employing  AES-­‐128  

IEEE 802.15.4: PHY layer

§  2.4  GHz  ISM  band  has  worldwide  availability  and    provides  
a  transmission  rate  of  250  kbps  
 

§  868  MHz  band  in  Europe  -­‐  data  rates  of  20  kbps  
§  915  MHz  ISM  band  in  the  USA  -­‐  data  rate  of  40  kbps  

33
IEEE 802.15.4: PHY layer packet structure

IEEE 802.15.4: channelization

34
 IEEE 802.15.4: channelization
§  27  frequency  channels  are  available  across  all  the  3  bands  
 

§  standard  includes  the  necessary  techniques  to  implement  

dynamic  channel  selec7on  to  avoid  interference  

§  PHY  layers  contain  several  lower-­‐level  func1ons  which  

enable  channel  assessment,  e.g.  
§  receiver  energy  detec7on  
§  link  quality  indica7on  
§  channel  switching  
 

§  func7ons  are  used  by  the  network  to  establish  its  ini7al  
opera7ng  channel  and  to  change  channels  in  response  to  
a  prolonged  outage  

IEEE 802.15.4: interference

§  interference  is  common  in  unlicensed  2.4  GHz  ISM  band  
because  of  other  services  opera7ng  in  that  band  

§  IEEE  802.15.4  applica7ons  have  low  QoS  requirements  

and  may  need  to  perform  mul7ple  retries  for  packet  
transmissions  on  interference  

§  since  IEEE  802.15.4  devices  may  be  sleeping  as  much  as  
99.9%  of  the  7me  they  are  opera7onal  &  employ  low-­‐
power  spread  spectrum  transmissions,  they  should  be  
among  the  best  of  neighbors  in  the  2.4  GHz  band  

35
 Bluetooth vs. IEEE 802.15.4
Bluetooth-­‐based  WPAN   IEEE  802.15.4  LR-­‐WPAN  
§  few  devices     §  many  devices  
§  data  range  is  10m  to  100m   §  data  range  is  nearly  10m  
§  data  rate  is  nearly  1  Mbps   §  data  rate  is  20/40/250  kbps  
§  low  power  consump7on   §  ultra-­‐low  power  consump7on  
§  low  baGery  life  is  low   §  baGery  lasts  years  
§  star  topology  only   §  peer-­‐to-­‐peer  &  star  topology  

ZigBee alliance
§  associa7on  of  companies  involved  with  building  higher-­‐
layer  standards  based  on  IEEE  802.15.4  
§  includes  network,  security,  and  applica7on  protocols  
 
§  rapidly  growing,  worldwide,  non-­‐profit  industry  
consor7um  consis7ng  of  leading  semiconductor  
manufacturers,  technology  providers,  end-­‐users  
 
§  organiza7on  with  a  mission  to  define  reliable,  cost-­‐
effec7ve,  low-­‐power,  wirelessly  networked,  monitoring  
and  control  products  based  on  an  open  global  standard  

36
 ZigBee alliance
§  targeted  at  home  and  building  automa7on  and  controls,  
consumer  electronics,  PC  peripherals,  medical  monitoring,  
and  toys  
 

§  primary  drivers  are  simplicity,  long  baGery  life,  

networking  capabili7es,  reliability,  and  cost  
 

§   alliance  provides  interoperability,  cer7fica7on  tes7ng,  

and  branding  
§  6  original  promoter  companies  –  Honeywell,  Invensys,  Mitsubishi,  
Motorola,  Samsung  and  Philips  
§  rapidly  growing  list  of  industry  leaders  worldwide  commiGed  to  
providing  ZigBee-­‐compliant  products  &  solu7ons  

ZigBee advantages
over proprietary solutions?
§  product  interoperability  

§   vendor  independence  

§  no  more  having  to  invest  resources  to  create  a  new  

proprietary  solu7on  from  scratch  every  7me  

§  companies  now  can  leverage  these  industry  standards  to  

instead  focus  their  energies  on  finding  and  serving  
customers  

37
 Wireless standards comparison

Power consumption
§  NXP  Bluetooth  2.0  
§  HV3  Mode:      14mA  (op7mis7c)  
§  HV1  master:  27  mA  
§  HV1  salve:        27  mA  
§  HV3  master:  26  mA  
§  Sniff  mode:  127  uA  
§  System  off:          5  uA  

§  TI  CC2420  chipset  

§  RX/TX:      27mA  
§  Power-­‐down:    0.3-­‐0.5uA  

§  Wi-­‐Fi  
§  pushing  towards  130  mA  (+18  dBm)  and  53  mA  in  Rx  
 

§  possible  to  push  power  envelope  to  even  lower  

§  special  solu7ons,  e.g.  EnOcean  

38
 – Exam#

Exam

§  typical  arrangement  of  an  oral  examina1on  

§  no  material  allowed  …  whatsoever  
§  bring  your  ID  card  

§   grading  of  the  exam/course  

§  homework  will  be  taken  in  account  
§  immediate  grade  feedback  

39
 Pedagogic points

§  aim  of  the  exam  is  to:  

§  test  comprehensive  understanding  of  all  aspects  
 
§  gauge  understanding  of  some  “nice”  details  
 
§  check  that  terminology  is  memorized  &  also  understood  
 
 

Overview of topics
§  overall  evolu1on  &  state  of  the  field  
§  who  did  what,  where  did  it  came  from;  vision/challenges  of  field  
§  trends:  you  should  remember  the  big  picture  
(but  we  will  not  ask  years  etc.)  
hello
 
§  graph  theory  &  rou1ng  fundamentals  
§  spanning  trees  
§  Dijkstra,  Bellman-­‐Ford,  Floyd-­‐Warshall  
§  count-­‐to-­‐infinity  &  split  horizon  
§  graph  theory  applica7ons,  
incl.  stable  marriage  for  resource  sharing  

40
 Overview of topics
§  rou1ng  for  ad  hoc  networks  
§  design  issues  (why  new  protocols  for  ad  hoc?)  
§  classifica7on:  proac7ve  vs.  reac7ve  
which way
§  DSDV,  DSR,  AODV,  ZRP,  …     do I go?

§  link-­‐reversal  algorithm  &  TORA  

§  underlying  principles  such  as  flooding,  etc.  
§  pros  &  cons;  and  protocols  themselves  

§  MAC  for  ad  hoc  networks  

§  hidden  terminal  problem  
§  MACA/MACAW,  CSMA/CA,  IEEE  802.11  MAC  
§  fairness,  energy  conserva7on,  adap7ve  modula7on,  …  
§  impact  of  direc7onal  antennas  on  MAC  &  rou7ng  
 

Overview of topics
§  radio  for  ad  hoc  networks  
§  basic  characteris7cs  of  building  blocks  of  
wireless  communica7ons  (radios)  
§  what  is  available?  what  are  the  characteris7cs?  
§  IEEE  802.11  (Wi-­‐Fi)  
§  Bluetooth  
§  IEEE  802.15.4  (&  ZigBee)  

§  wireless  sensor  networks  

§  mo7va7on,  classifica7on,  design  challenges  
§  WSN  =  MANET?  
§  sensor  nodes  (“motes”)  
§  WSN  applica7ons  
§  LEACH  clustering  protocol  

41
 Overview of topics
§  Ar1cles,  ar1cles,  ar1cles!  
§  remember  &  understand  what  you  have  read  
§  note  that  homework  material  is  also  a  part  of  the  exam  curriculum  

“should we go deep into research papers mentioned in lectures or

is the lecture material enough for the exam?”
§  lecture  notes  include  references  to  original  research  papers  where  
various  protocols  were  proposed,  so  you  could  obtain  further  details  of  
protocols  we  studied,  if  and  when  needed  
§  it  is  not  expected  that  you  have  read  each  paper  in  detail,  
except  for  those  papers  assigned  for  homework  exercises!  
 

§  good  rule  of  thumb:  amount  of  7me  spent  in  lectures  &  exercises  on  a  
given  protocol/concept  gives  you  a  good  indica7on  of  its  rela7ve  
importance  in  overall  course  

Exam strategy tips

§  be  well  prepared  &  come  on  1me  

§  at  least  15  minutes  before  your  assigned  7me-­‐slot  is  good  
prac7ce  

§  1me  is  1cking  &  logic  is  needed  

§  the  exam  7me  is  limited  (esp.  as  we  have  an  oral  exam)  
§  we  try  to  give  you  the  best  help  and  7me,  but  you  should  aim  
at  logical  and  well-­‐though  out  answers  …  this  saves  7me  and  
gives  you  the  opportunity  to  achieve  a  higher  grade  
§  you  can  (i.e.  should)  use  the  whiteboard  to  extend  your  
explana7ons  

42
 Exam strategy tips
§  remember  that  you  should  be  able  to  not  only  explain  what  
something  is,  but  preferably  should  be  able  to  also  discuss  
“why”  something  is  done  like  it  is  

§  note  that  apart  from  direct  ‘raw  knowledge’  ques7ons,  

we  might  also  ask  ques7ons  like:  
§  “normally  we  do  X  with  assumpXon  Y;  what  would  happen  if  Y  is  not  
true  but  it  is  Z  -­‐  does  X  sXll  work?”  

“how do you send a text message?”

Exam strategy tips

Don’t panic!


43
 Exam strategy tips

Study!

§  if  you  achieve  a  grade  beier  than  1.3  you  are  on  track  for  
doing  your  M.Sc.  Thesis  at  iNETS  …  
 

44
45