You are on page 1of 34

Quality of Services for Wireless/Fixed

Communication Systems
ITU Centre of Excellence for Arab Region
Tunis – Hammamet
8-10 November 2010
Quality of Services for Wireless/Fixed
Communication Systems
ITU Centre of Excellence for Arab Region
Session 1
General aspects of Quality of Service and
Network Performance
Abdelkerim Amara
CERT
Amara.Abdelkerim@cert.mincom.tn
Workshop : QoS for wireless/fixed communication systems Session 1 : General aspects of Quality of Service and Network Performance
3
1. General Definition of QoS
2. Particular difinitions of QoS
3. Inter-relationships between various viewpoints of QoS
4. Network Performance (NP)
5. Service Performance
6. Relationship between QoS and NP
7. Methodology for handling aspects of NP related to QoS
8. User related QoS parameter definitions and measurements
9. QoS Parameters related to PLMN
Contents
Workshop : QoS for wireless/fixed communication systems Session 1 : General aspects of Quality of Service and Network Performance
4
Workshop : QoS for wireless/fixed communication systems Session 1 : General aspects of Quality of Service and Network Performance
5
Workshop : QoS for wireless/fixed communication systems Session 1 : General aspects of Quality of Service and Network Performance
6
Workshop : QoS for wireless/fixed communication systems Session 1 : General aspects of Quality of Service and Network Performance
7
Workshop : QoS for wireless/fixed communication systems Session 1 : General aspects of Quality of Service and Network Performance
8
“The collective effect of service performance which
determine the degree of satisfaction of a user of the
service “
CCITT Recommendation E.800
General definition of Quality of Service (QoS)
Workshop : QoS for wireless/fixed communication systems Session 1 : General aspects of Quality of Service and Network Performance
9
the statement of the level of quality of a particular service required or
preferred by the user/customer. The level of quality may be expressed by
the user/customer in technical or non-technical language
QoS Requirement of User/Customer
a statement of the level of quality expected to be offered to the
user/customer by the service provider. The level of quality is expressed by
values assigned to QoS parameters. These parameters are usually
designed to be understandable to the user/customer. Each service would
have its own set of QoS parameters.
QoS Offered by Service Provider
Particular definitions of QoS
Workshop : QoS for wireless/fixed communication systems Session 1 : General aspects of Quality of Service and Network Performance
10
A statement of the level of quality achieved by the service provider. This is
expressed by values assigned to parameters, which are, as far as possible,
same as those for the QoS offered. These performance figures are
summarized for specified periods of time
QoS achieved by service provider
A statement expressing the level of quality experienced by them. The QoS
perceived is expressed, usually in terms of degrees of satisfaction and not
in technical terms. Technical terms may be expressed where the
user/customer is able to understand and use these. QoS perceived is
assessed by customer surveys and from user's/customer's own comments
on levels of service.
QoS perceived by the user/customer
Particular definitions of QoS
Workshop : QoS for wireless/fixed communication systems Session 1 : General aspects of Quality of Service and Network Performance
11
Inter-relationships between various viewpoints of QoS
Workshop : QoS for wireless/fixed communication systems Session 1 : General aspects of Quality of Service and Network Performance
12
"The ability of a network or a network portion to provide the functions related to
communications between users"
CCITT Recommendation E.800
◊ A statement of the performance of connection element or concatenation of
connection elements employed to provide a service.
◊ It is defined and measured in terms of parameters which are meaningful to the
network and service provider and are used for the purposes of system design,
configuration, operation and maintenance.
◊ NP is defined independently of terminal performance and user/customer actions.
◊ It is also service independent in that it must be able to support all the services the
particular network level is required to transport.
◊ The performance of the network is generally aimed to provide the QoS offered to the
user/customer
Network Performance (NP)
Workshop : QoS for wireless/fixed communication systems Session 1 : General aspects of Quality of Service and Network Performance
13
A statement of performance of a telecommunications service expressed in
parameters, applicable to that service, together with values for those parameters.
These parameters will apply to QoS, technical and non-technical features of the
service
◊ Each service would have its own set of performance parameters and values to constitute
the service performance.
◊ Service performance is expressed in a more formal language, yet understandable and
meaningful to the user/customer.
◊ The QoS parameters included in the service performance is the QoS offered.
◊ It is up to the service provider to state if any guarantee of the QoS is implied over a
period of time.
◊ It is often convenient to translate the user's/customer's QoS requirements into service
performance before these are translated into network performance parameters.
Service Performance
Workshop : QoS for wireless/fixed communication systems Session 1 : General aspects of Quality of Service and Network Performance
14
QoS is the starting point for the
development of NP parameters and
targets.
NP parameters specified are chosen
for the most effective operation of the
network
Relationship between QoS and NP
Workshop : QoS for wireless/fixed communication systems Session 1 : General aspects of Quality of Service and Network Performance
15
Methodology for handling aspects of NP related to QoS
1
2
3
4
5 6
Workshop : QoS for wireless/fixed communication systems Session 1 : General aspects of Quality of Service and Network Performance
16
◊ Step 1 : Define service specific framework
◊ Step 2 : Ascertain customer/user QoS requirement (questionnaire)
◊ Sources influencing QoS requirement : Users; Service providers, regulators.
Capture of customer QoS requirement
Mapping QoS and specifying performance parameters
◊ SP studies the implications of the level of quality required
◊ SP may wish to review the cost, strategic and other relevant implications before deciding
the level of performance to be offered.
◊ The QoS offered is expressed, where possible, in user/customer understandable terms.
◊ The SP decides on the number of parameters with which to specify the QoS offered
Methodology for handling aspects of NP related to QoS
Workshop : QoS for wireless/fixed communication systems Session 1 : General aspects of Quality of Service and Network Performance
17
◊ The parameters of QoS offered are separated into non-network and network related.
◊ The network related QoS criteria are mapped into network performance parameters
◊ The network performance targets are aimed to become planning documents
Mapping and specifying NP objectives
Measurement of network Performance
◊ In a carefully designed network monitoring system, maximum benefit will accrue from
the minimum set of measurements.
◊ measurement systems should reflect directly the network performance parameters
◊ End-to-end performance should be estimated in case of network performance is braked
down to elemental performance.
◊ measurements are recommended only for those parameters which will be of concern to
the user/customer, the regulator, the service provider or the NP
◊ measurements may be taken on a sample basis.
Methodology for handling aspects of NP related to QoS
Workshop : QoS for wireless/fixed communication systems Session 1 : General aspects of Quality of Service and Network Performance
18
QoS Acheived
QoS perceived
◊ Assessment of QoS perceived is usually made by customer
◊ The QoS perceived and the QoS achieved are compared for correlation
◊ The variation in correlation could be studied for accountable reasons. If no reasons are
found this becomes a matter for further investigation
◊ The aim will be to reach a level of QoS achieved which correlates satisfactorily with the
QoS perceived
◊ The QoS achieved is obtained by combining results achieved from the non-network and
network related monitoring systems
◊ End-to-end performance is to be estimated where the elemental performance is
measured
◊ The confidence limit of the resulting end-to-end performance is to be stated whenever
these are computed from samples
◊ where possible mapping is to be carried out to express the QoS achieved with the same
parameters as the QoS offered.
Methodology for handling aspects of NP related to QoS
Workshop : QoS for wireless/fixed communication systems Session 1 : General aspects of Quality of Service and Network Performance
19
User related QoS parameter definitions and measurements
Parameter
Use
Specifying
the level of
QoS
Comparing
QoS aspects
of different
service
offers
Preparing
long term
studies on
QoS aspects
of a specific
service
Comparing
the QoS of
different
service
providers
Use of parameters
Workshop : QoS for wireless/fixed communication systems Session 1 : General aspects of Quality of Service and Network Performance
20
User related QoS parameter definitions and measurements
Considerations
Networks may not be
designed with the
same priorities for
coverage
The absolute average
performance level of
networks may be of
little interest to users
The capacity and
coverage of mobile
networks changes
frequently
The performance
measured on test
drives applies only to
the routes chosen
 One network may provide better coverage in one area and
another in a different area
 Test drives conducted at one point in time during a reporting period may not be indicative of the
average performance over the whole period
 The choice of routes may not represent the average of user's locations over the whole network
 For test drives carried out independently of the operators, it is advisable that the operators do not have knowledge of the routes in advance
 Networks may focus on
different areas of a country
Workshop : QoS for wireless/fixed communication systems Session 1 : General aspects of Quality of Service and Network Performance
21
 In some cases measured QoS figures may be distorted (disasters, freak
weather, etc.)
 This could degrade QoS by inducing exceptional traffic levels etc.
 Service providers should provide the measured QoS and may additionally
provide a second figure which excludes the effects of the exceptional
circumstances.
 A note clearly explaining the difference should also be provided.
 The effect on the reported QoS of a service provider covering a small area
is likely to be more severe.
Data processing issues
User related QoS parameter definitions and measurements
Workshop : QoS for wireless/fixed communication systems Session 1 : General aspects of Quality of Service and Network Performance
22
 Where the measurements are to be used for long term comparisons, it is
recommended that QoS data should be collected and calculated on a
quarterly basis
 Stakeholders may also decide to use longer or shorter data collection periods.
 There may be cases were a longer period is more practicable, e.g. extensive
customer surveys.
 Shorter periods are advisable for QoS aspects where frequent and fast
changes in quality are likely to occur
Data collection period
User related QoS parameter definitions and measurements
Workshop : QoS for wireless/fixed communication systems Session 1 : General aspects of Quality of Service and Network Performance
23
 Test drives should be planned to ensure as far as practicable that the results
adequately reflect the QoS perceived by customers for the period under review.
 Test calls should be made to a test number in an interconnected fixed network
(PSTN/ISDN).
 The drive round surveys shall be designed to be representative of the population to
measure for the results to have the needed accuracy.
 Measurements should be scheduled so as to reflect accurately traffic variations over
the hours of a day, the days of the week and the months of the year, and user's
behaviour
Sampling and test calls
General considerations
Workshop : QoS for wireless/fixed communication systems Session 1 : General aspects of Quality of Service and Network Performance
24
Results should be compared only where the same measurement methods are
used
For example:
Measurements made with drive round tests on one network may be compared with
measurements made with drive round tests on another network but not with
measurements made with network element counters.
Comparability of the results
User related QoS parameter definitions and measurements
Workshop : QoS for wireless/fixed communication systems Session 1 : General aspects of Quality of Service and Network Performance
25
QoS Parameters related to PLMN
1. The ability to provide the services where they
are required by the user
2. It is perceived by the user as a lack of service
where it was expected. Depending on the kind
of network failure, it can be seen as no
coverage in the mobile, as an impossibility to
make calls having coverage, or as an
impossibility to access a concrete service
3. Service accessibility is the accessibility to a
service when there is network access. Where
there is network access, service access may
not be possible because:
- There are no available radio channels to
support the service.
- There are no available transmission links
between the base station and the mobile
switching centre
Workshop : QoS for wireless/fixed communication systems Session 1 : General aspects of Quality of Service and Network Performance
26
QoS Parameters related to PLMN
Main QoS Parameters
Unsuccessful Call Ratio
 Unsuccessful call ratio is defined as the ratio of unsuccessful calls to the total
number of call attempts in a specified time period.
 An unsuccessful call is a call attempt to a valid number, while in a coverage area,
where neither the call is answered nor called party busy tone nor ringing tone, is
recognized at the access of the calling user within 40 seconds from the instant when
the last digit of the destination subscriber number is received by the network
Dropped Call Ratio
 The proportion of incoming and outgoing calls which, once they have been correctly
established and therefore have an assigned traffic channel, are dropped or
interrupted prior to their normal completion by the user, the cause of the early
termination being within the operator's network.
 The objective of this parameter is to obtain a measurement of the ability of the
mobile network used by the service provider to maintain a call once it has been
correctly established. This parameter measures failures in coverage, problems with
the quality of the signal, network congestion and network failures
Workshop : QoS for wireless/fixed communication systems Session 1 : General aspects of Quality of Service and Network Performance
27
QoS Parameters related to PLMN
Considerations on Drive Round tests
Objective 1
simplest objective
provide an external
indicative snapshot of the
performance of a network
on the limited routes
chosen for the survey
Objective 2
most complex objective
obtain an accurate
measure of the traffic
weighted average
performance of the whole
network as would be
perceived by a real user
 may be useful to obtain an
impression of the
performance
 will not, however, produce
results that are free from bias
and necessarily representative
of the whole network
 comparability of the results
will be necessary
 samples with a sufficient number
of calls
 samples must be representative
of the whole network
 Handling of the user's behaviour
of the terminal, the location and
the timing of the calls
 area to be surveyed may be
made up of many different
sample areas
 Areas size should be inversely
proportional to the expected
traffic density,
 sample areas should cover all
types of geography in the whole
area or country in a
representative proportion
Workshop : QoS for wireless/fixed communication systems Session 1 : General aspects of Quality of Service and Network Performance
28
QoS Parameters related to PLMN
Considerations on Drive Round tests
Objective 1
simplest objective
provide an external
indicative snapshot of the
performance of a network
on the limited routes
chosen for the survey
Objective 2
most complex objective
obtain an accurate
measure of the traffic
weighted average
performance of the whole
network as would be
perceived by a real user
 may be useful to obtain an
impression of the
performance
 will not, however, produce
results that are free from bias
and necessarily representative
of the whole network
 comparability of the results
will be necessary
 Survey has to cover the different
circumstances in which mobiles
are used e.g. in car, in building,
walking, in homes, in offices, in
trains and the different ways in
which mobiles are held by their
users
 test calls to be made at
representative times of day in
proportion to the time
distribution of real traffic
 test calls to be made for
durations representative of the
durations of real calls (drop)
 Survey to be repeated at
intervals to take account of
changes with time of the
demand for calls and the
capacity of the network
Workshop : QoS for wireless/fixed communication systems Session 1 : General aspects of Quality of Service and Network Performance
29
QoS Parameters related to PLMN
Considerations on Network elements counters
 QoS experienced by users will be taken into account as the calls are made with the real
terminals by the users, so results will reflect accurately the QoS.
 Any measurements intended to compare the quality of different networks should use a
common measurement system agreed by the network operators. As the real traffic
provides measurements from the entire network, all the different locations and different
times are measured => the representativity is assured.
 The network operators should agree on the set of network element counters that are
equivalent for different network manufacturers in order to get comparable results.
Workshop : QoS for wireless/fixed communication systems Session 1 : General aspects of Quality of Service and Network Performance
30
QoS Parameters related to PLMN
Measurements
 Measurement is made using an automatic data collection system, based on network
counters, which register the real traffic of the network.
 Network counters collect information for 24 hours a day, every day of the year to reflect
the variations in traffic which occur during the different days, weeks and months of the
year.
Considerations on Network elements counters
Workshop : QoS for wireless/fixed communication systems Session 1 : General aspects of Quality of Service and Network Performance
31
QoS Parameters related to PLMN
Service accessibility: Unsuccessful call ratio
Considerations on Network elements counters
( )
100
call a for TCH at taking Attempts
call a for TCH of taking Successful - call a for TCH at taking Attempts
Ratio Call ul Unsuccessf × =
formula includes the attempts to seize a TCH for an originated or terminated call and the success at assigning
a TCH for an originated or terminated call.
100
calls Successful
calls d Interrupte
D ×
|
.
|

\
|
= Ratio roppedCall
formula includes the interrupted calls which consist of failures which cause the dropping of the channel once
the TCH has been successfully established, and the successful seizure of TCH for an originated or terminated call
Service retainability: Dropped calls ratio
Workshop : QoS for wireless/fixed communication systems Session 1 : General aspects of Quality of Service and Network Performance
32
Drive tests vs Network elements counters
Drive Round Tests
measure the network from an external
point as would be seen by the user
The same test system can be used to
compare results for different networks
Locations where there is no coverage are
taken fully into account
not depend on any correct functioning of
the network to enable a measurement
to be made
Comparability of results at the same
point and time is high
Network Elements Counters
include the effects of all calls, and so
provide better comparability of congestion
and network failures
take account of changes in terminals and
the actual performance achieved by real
terminals used by real users
indicators may be produced for different
regions and periods out of the same data
base
Drive tests vs Network elements counters
Workshop : QoS for wireless/fixed communication systems Session 1 : General aspects of Quality of Service and Network Performance
33
Drive tests vs Network elements counters
Drive Round Tests
tests are not indicative of how users
actually handle their terminals
To obtain adequate accuracy for
comparison purposes, a very large
number of samples is needed on the
chosen test drives
results are not necessarily highly
representative of the performance
of the whole network
Network Elements Counters
Call attempts made out of coverage are
not taken into account
They depend on software algorithms in
the switches and base station
controllers that implement the
counters
Drive tests vs Network elements counters
Workshop : QoS for wireless/fixed communication systems Session 1 : General aspects of Quality of Service and Network Performance
34