UK

Radio Datasheet 0239 Nokia GPRS Radio KPIs
This datasheet is compiled and owned by Radio Engineering. Date of last update : 23 June 2004.

SUMMARY: This datasheet shows how to use the Nokia GPRS Radio Key
Performance Indicators (KPIs). Each KPI is defined, the target or
optimum range of values is shown, and the corrective action to be taken if
the values are outside the desired range. The datasheet provides the
recommended action for optimum performance of the GPRS radio
network and the ability to support video streaming.

Target Audience Regional Planners, Capacity Planners, Optimisers

Introduction There are 9 main key performance indicators presented in this document. For a full
explanation of these and the associated Metrica counter names, please refer to [1].
All of these statistics can be found in the Performance Engineering Groups Weekly
performance reports.

KPI 1 - Available Timeslots
There are three types of GPRS timeslots : dedicated, default and additional.
Dedicated timeslots are reserved exclusively for GPRS and cannot be used for
voice. Default timeslots are pre-switched for GPRS use but can be used for voice
when timeslots are required. Additional timeslots are pre-switched to voice but
converted to GPRS if the voice load is low enough and GPRS demand is high
enough. Note that for default and additional timeslots, voice traffic has priority
over GPRS if there is contention for timeslots.
Target : The most important measure for the performance of GPRS is how many
timeslots (a sum of dedicated, default and additional) are available for GPRS. The
target for this depends of the site configuration : for a 1 carrier site an average of at
least 4.5 timeslots must be maintained during the circuit switched busy hour, while
all other configurations of site must maintain an average of at least 5.4 timeslots.
Action to be taken : The optimum solution is to add an extra carrier to a cell to
increase the average number of timeslots available. The timeslots on this extra
carrier do not need to be dedicated to GPRS.

KPI 2 - TBFs Per Timeslot
GPRS differs from circuit switched operation by virtue of multiple users being able
to share timeslots. A timeslot can be simultaneously shared by up to nine users on
the downlink and seven on the uplink and so an important measure of both cell and
user performance is how much sharing is taking place. The more users that share
the same resource, the slower each users’ throughput will be. The Average TBFs per
Timeslot KPI measures the amount of resource sharing taking place; i.e. the
average number of users
It is important to note that this KPI is in the range 0-9 on the downlink and 0-7 on
the uplink. It is normal for values of this KPI to be less than 1. For example, if
there is a single user is on a cell using three timeslots to transfer data, then the cell
will report the KPI as 3/6=0.5 as there are 6 timeslots pre-switched to GPRS.
Target : The target for this KPI is less than or equal to 0.7 during the circuit
switched busy hour. Values greater than 1 imply significant sharing of resource and
a poor user throughput experience. This statistic can be found in the Performance
Engineering Group’s Weekly Report titled “GPRS Performance Summary”
Action to be taken: The optimum solution is to add an extra carrier to a cell to
increase the average timeslots available.

Issue 2 01 June 2004 In Confidence Page 1 of 4

KPI 4 . This statistic can be found in the Performance Engineering Group’s Weekly Report titled “GPRS Traffic and Blocks by Cell”. This statistic can be found in the Performance Engineering Group’s Weekly Report titled “GPRS Traffic and Blocks by Cell” Action to be taken : As it is not currently possible for the network to force the handset to hand GPRS sessions from cell to cell. but at the expense of increased voice blocking. Cells with low voice blocking will not benefit from dedication of GPRS timeslots – option 1 is recommended in this case. Target : Normal operation of a site will operate with a territory utilisation between 0% and 20%. averaged over a one hour period.Cell Traffic A measure of GPRS activity on a cell is the Cell Traffic intensity. As such. this is determined by the handset’s Multislot Class. This statistic can be found in the Performance Engineering Group’s Weekly Report titled “GPRS Performance Summary” KPI 5 . either for uplink or downlink data transfer. packet switched blocking can be alleviated by two methods: 1. Action to be taken : An extra carrier is required in order to reduce the territory utilisation. dedication of timeslots to GPRS may help to alleviate customer performance issues. greater than 50% occupancy indicates that the users will be experiencing low throughput and requires action. Most handsets allow a maximum of either 3 or 4 timeslots to be used in the Issue 2 01 June 2004 In Confidence Page 2 of 4 . measured in kilobits per second.Multislot Allocation Percentage Each GPRS handset has a capability to transmit and receive using a specified maximum number of timeslots. The extra carrier need not be dedicated to GPRS as voice traffic will not need to use much of the additional timeslots – essentially these timeslots are acting as dedicated resource.Territory Utilisation The Territory Utilisation KPI shows the occupancy of the GPRS timeslots by GPRS traffic. However blocking levels in excess of this indicate that the customer experience is being degraded. 2.O2 UK Radio Datasheet 0239 Nokia GPRS Radio KPIs KPI 3 . There is no specific target for this KPI as it is highly sensitive on the traffic mix of the site. Target : A cell is typically likely to require attention when the level of blocking exceeds 5% in either the uplink or the downlink.GPRS Blocking The most fundamental requirement for a GPRS subscriber is to be able to access the resource. In circumstances where there is circuit switched blocking of more than 1%. Dedication of timeslots to GPRS on cells with high voice blocking The addition of dedicated timeslots will reduce blocking on the GPRS network. It must be borne in mind that the GPRS traffic carried may be limited by the amount of voice traffic: a cell may appear to carry low traffic volumes but this may be due to high voice traffic and blocking. If this is not possible then the resultant effect is called blocking. The carrier will also allow alleviation of voice blocking. Addition of an extra carrier This will increase the availability of the network and user throughput. but at the expense of voice revenue and voice blocking. KPI 6 . (in the same manner as handover in Circuit Switched operation) the only solution to increasing GPRS availability is to increase the GPRS timeslots on the blocking cell. Blocking levels of less than 5% offer an acceptable quality of service to the users.

Coding Scheme 2 is faster but is less resilient to errors. In order to maintain a high user throughput experience on both the uplink and the downlink. A value of 100% means that all resource requests were fully honoured. close to a motorway or railway). and these are used to produce the “Multislot Allocation Percentage” KPI. is required to allow acceptable performance for the users. Both the handset and network are able to alter which coding schemes are in use according to the number of errors incurred during data transmission. This key performance indicator is a measure of the network’s success in fulfilling the resource that is requested. however this action causes a poor user experience of the service. Action to be taken : If the TBF drops are taking place due to capacity shortages. However. If only two are granted then the calculation is (2*1)/(4*1)=50%. the network may not be able to fulfil the resource request completely and is capable of providing the handset with as many timeslots as possible. coverage issues. interference. If the cell is known to carry high mobility users (e. etc) 2. and is then released to allow other users access to the resources. KPI 8 . A combination of a TBF drop rate greater than 1% with blocking greater than 5% will indicate this. Coding Scheme 2 will be used whenever possible. dropped) for two reasons: 1.g. if 4 timeslots are requested and 4 are granted then the KPI calculation is (4*1)/(4*1)=100%. For example. This limits the requirement to repeat lengthy streams of data to that which cannot be corrected by error correction. It is held only for as long as is required. Action to be taken : A value less than 90% suggests a stress on the existing resources and indicates that an additional carrier. This statistic can be found in the Performance Engineering Group’s Weekly Report titled “GPRS Performance Summary” Target : The multislot allocation percentage should be greater than 90%. The TBF may be released abnormally (i. There are counters available to peg allocations of timeslots for both the uplink and downlink. It is also possible that the resource can be released abnormally before all the data has been transferred. KPI 7 .e.g. then re-alignment of the sectors may help. then the site should be considered for a capacity upgrade.Block Error Rate The O2 GPRS network operates by transferring data across the air interface using one of two transmission encoding types known as Coding Schemes. The network may lose radio contact with the handset (e. The network will try to supply the maximum amount of resource that the device can handle when required. or an increase in dedicated timeslots if carrier expansion is not possible. Values higher than this should be dealt with on a cell-by-cell basis. The network may need to use the GPRS timeslot for a voice call. So it can be seen that a low value means that devices were allocated a lot less resource than was requested. Issue 2 01 June 2004 In Confidence Page 3 of 4 . If the drops are due to radio coverage.O2 UK Radio Datasheet 0239 Nokia GPRS Radio KPIs downlink. This statistic can be found in the Performance Engineering Group’s Weekly Report titled “GPRS Performance Summary” Threshold : Acceptable performance is indicated by a drop rate of less than 1%. and either 1 or 2 timeslots in the uplink. then up to 2% dropped TBFs is acceptable.. and as such the user experience will be poor. An increase of dedicated timeslots will also help to alleviate the condition. Coding Scheme 1 offers slow but robust transmission.TBF Drop Rate A temporary block flow is the name given to the resource used to transfer data across the air interface.

A re-orientation of the sector or increasing the transmit power of the carrier may improve this KPI. This can be confirmed by looking at the Minute per Drop characteristic of the voice traffic . either through limited coverage of adjacent or co-channel interference. coding scheme 2 should be used 100% of the time for data transfer. there are two coding schemes used to transfer data across the GPRS air interface.com. Action to be taken : A block error rate of greater than 10% tends to indicate a poor radio environment. if the radio conditions are poor then the call drops due to radio coverage will be high. A re-orientation of the sector or increasing the transmit power of the carrier may improve this KPI. In order to achieve maximum user throughput.Coding Scheme Utilisation As described in KPI 8. Issue 2 01 June 2004 In Confidence Page 4 of 4 .engineering. KPI 9 . if the radio conditions are poor then the call drops due to radio coverage will be high. either through limited coverage of adjacent or co-channel interference. Action to be taken : A coding scheme 2 utilisation of less than 90% tends to indicate a poor radio environment. please contact the Radio Engineering helpdesk on 01753 565882 or e-mail radio. Radio Engineering. For any questions or issues. This can be confirmed by looking at the Minute per Drop characteristic of the voice traffic . This statistic can be found in the Performance Engineering Group’s Weekly Report titled “GPRS Performance Summary” Threshold : The utilisation of coding scheme 2 should be greater than 90% for both the uplink and downlink.uk@O2.O2 UK Radio Datasheet 0239 Nokia GPRS Radio KPIs This statistic can be found in the Performance Engineering Group’s Weekly Report titled “GPRS Performance Summary” Target : The block error rate should be less than 10% for both the uplink and downlink and coding schemes 1 and 2. References [1] UKRE_04_005 GPRS Radio Network Key Performance Indicators END OF DOCUMENT The author of this document is Neil McDonald. The measure of coding scheme 2 utilisation is given in this key performance indicator.