WCDMA Radio Optimization

Contents
Radio planning optimization ................................................ 3-14 Quality of Service .............................................................. 15-22 Measurement and statistics collection ............................... 23-29 KPI .................................................................................... 30-32 Accessibility....................................................................... 33-38 Retainability....................................................................... 39-44 Integrity ............................................................................. 45-47 HSDPA-HSUPA ................................................................ 48-55

-2-

Radio planning optimization

-3-

Antenna height

Since WCDMA performance is interference limited the cell dominance areas should be kept as controlled as possible lf the antenna is located too high (no proper tilting) then The cell gathers more traffic and external interference and thus the effective capacity is decreased Produced interference decreases the capacity of the surrounding network Also surrounding networks service probability is negatively effected

Antenna azimuth

Natural obstacles and buildings should be used to create good dominance areas for WCDMA cells This improves the SHO performance and decrease interference Example of a UMTS cell, that is naturally bordered (wall effect) by buildings

Antenna height simulation

When re-using the GSM sites, analysis should be made whether the UMTS antennas should be positioned lower This analysis is done with simulations and visiting the site locations in practise
Part of network reused few +40meter GSM antenna heights High UMTS antenna positions lowered to 25-35m

Dominance areas become clear, so less interference is introduced and HO performance is better. Capacity is increased and performance enhanced!

-4-

Antenna tilt

ln addition to antenna height, downtilting is very important physical means for interference minimizing in WCDMA Basic rule of designing antenna tilt is that the height of the antenna should be selected with respect to the wanted amount of cell range If the cell range with respect to available antennas and their tilting with a feasible amount of tx-power becomes too large to suit the network plan, then the antenna must be lowered According to experience, the analysis should start with the optimum tilting and not by reducing the tx-powers of the cell, which can be optimized after the tiltings are done Horizontal plane h

Antenna tilt

According to experience even 15 degrees of downtilting is not impossible (lf the radiation pattern of the antenna supports it), although in practice not very often needed.

There has also been lot of discussion of a potential need to change the tilts often during the network lifecycle (even regularly) However practice have not shown such need if the tilts are design well from the start with help from simulations But once WCDMA gets congested this might be given another look (Remote tilts).

-5-

Sectorisation

According to simulations and analysis, sectorisation of WCDMA site helps to improve capacity of the network However, as permissions for additional antennas are quite hard to come by, e.g. 6-sector sites might be very rare
Antenna 3 Other to own dB cell Beam interference width ratio, i Served users Soft handover overhead UL coverage probability (outdoor to indoor) For 8/64/144 kbps 70/32/40% 85/50/59% 87/55/62% 86/59/62% 90/62/68% 92/67/72% 92/70/71% 88/65/64% 95/75/79% 96/80/82% 96/80/81% 93/76/76%

Sectorisation can increase the capacity if correct beamwidth antennas are selected and SHO properly controlled

OMNI CASE
Omni 1200 900 650 1200 900 650 330 120 900 650 330
0

0.79 1.33 1.19 0.88 1.72 1.49 1.09 0.92 2.18 1.97 1.43 1.15

240 441 461 575 489 510 604 691 593 627 758 880

28% 39% 35% 34% 54% 51% 41% 40% 64% 59% 55% 48%

THREE SECTORS CASE

FOUR SECTOR CASE

SIX SECTOR CASE

Master head amplifier

The MHA can be used in WCDMA in the uplink direction to compensate for the cable losses and thus reducing the required mobile stations transmit powers Using MHA the performance in uplink can be improved also in WCDMA systems. However in practice if the network turns to downlink limited then the MHA wont help
Other to Served own cell users in interference UL ratio, I Served users in DL UL coverage probability (outdoor to indoor) for 8/64/144 hbps 93/78/78% 95/82/82% 96/86/85% 98/89/89% 97/87/86% 98/90/90% 95/83/82% 98/90/90%

THREE SECTORED CASE, 65O antenna

No MHA With MHA No MHA With MHA no MHA with MHA No MHA 4dB cable losses WITH mha 4 DbN CABLE LOSSES 0.60 0.61 0.73 0.73 0.88 0.90 0.88 0.90 1038 1064 1089 1107 1124 1132 1109 1132 807 746 884 846 1052 1021 1057 1016

FOUR SECTTORED CASE, 65O antenna SIX SECTORED CASE, 33O antenna

-6-

Master head amplifier

Increases uplink coverage/capacity in low loaded network Compensates for feeder and combiner losses in the uplink direction, increasing coverage for suburban, rural and road sites where antennas are in very high positions and the feeder lines are long Allows UEs to reduce transmission power level With heavily loaded network (i.e. high interference) the benefit of the mast head amplifier is negligible Also in downlink limited 3G networks (DL oriented traffic, users in cell edge, DL tx-power per user low e.g. in for high bit rate indoor users) the usage of mast head amplifier is not justified Needs extra space in the masts and increase the wind load

MHA is sometimes called as Tower Mounted Amplifier (TMA)

Transmit power increase

-7-

Transmission powers

Default transmission powers are determined by the equipment vendors. In initial phase of the planning Transmission powers of TCHs and CCHs needs to be set Maximum UE transmission power is to be defined In DL the power tuning between TCHs and CCHs has effect on network performance More power to CCHs > better channel estimation, which improves the Eb/No performance and thus improves coverage More power to TCHs > better capacity Rule of thumb: 15-20% of DL total power is used for CCHs Maximum UE transmission power should be set to 21-24 dBm (network operation and battery life) Most important control channel is the common pilot channel (CPICH)

Transmission powers

Also other control channels beside CPICH need power (for example BCH) to enable correct functioning of the system All the other common control channels are powered in relation to the P-CPICH The goal of allocating power to the common channels is to find a minimum power level needed for each channel to secure the network operation and to provide the same cell coverage area as with CPICH, but not to waste any capacity left for the traffic channels.

Typical DL power recommendations

Channel Max power of the Node B CPICH PCH SCH FACH BCH Allocated power 43 dBm Max power-10dB Max power-11 -13 dB Max power-11 12dB Max power-12-13dB Max power-11 13dB

-8-

Recall: Some control channels

PCH: Paging channel initiates the communication from network side SCH: Synchronization channel FACH: Forward access channel carries control information to terminals that are known to be located in the given cell. ls used to answer to the UL RACH message. BCH: Broadcast channel carries network specific information to the given cell (random access slots for UL, antenna configuration etc) PICH: Paging indicator channel is used to provide sleep mode operation for UE AICH: Acquisition indicator channel is used to indicate the reception of RACH CCPCH: Primary and secondary common control physical channels (P-CCPCH and S-CCPCH) are physical channels that carry BCH, FACH and PCH.

Transmission powers

P-CCPCH transmitted with activity factor 0,9 S-CCPCH transmitted with activity factor 0,25 SCHs transmitted with activity factor 0,1 AICH, PICH and CPICH are transmitted continuously

The BCH is transmitted on the P-CCPCH and FACH and PCH on the S-CCPCH The BCH is transmitted on the P-CCPCH continuously expect during the 256 first chips, when the P-SCH and S-SCh are transmitted we can assume 0,1 activity factor for the SCHs and 0,9 for the P-CCPCH
Channel P-SCH S-SCH PICH AICH P-CCPCH S-CCPCH CPICH All common ch. Allocated power 0,331W 0,224W 0,1W 0,126W 0,245W 1,165W 1W 3,191W 31% 100% 5% 16% Power out of the total common channel powers Power out of the maximum Node B transmission power (20W)

-9-

Carrier addition

Adding a carrier to less transmit power per carrier, if no additional PA is installed Additional carrier can also be used for e.g. optimisation of indoor coverage with clever network planning and parametrisation (not with power reduction) Even with less transmit power, there is a capacity gain possible especially for high traffic areas (low cell range) Actual gain produced is heavily dependent on the traffic mix Carrier configuration 1C>2C 2C>3C Dense Urban 350m 92% 41% DL Capacity gain Urban Suburban 550m 1700m 87% 37% 77% 27% Rural 7km 60% 15%

Indoor coverage aspects

Most of the UMTS users are indoors. Therefore good indoor coverage is vital for UMTS success In GSM indoor coverage is pretty straightforward to plan. However this is not the case with WCDMA Indoor coverage provided from outdoor base stations is highly sensitive to cell load increase in WCDMA If outdoor users is given a high-data rate bearer this can result in loss of coverage to users indoors INDOOR COVERAGE ANALYSIS Consider different RAB / coverage scenarios Carefully estimate the effect of cell loading to the coverage Use repeaters if possible Assess the need for indoor sites Carry out real-life verification of the planning
- 10 -

Pilot pollution

Pilot pollution is faced on a certain area when there is no clearly dominant CPICHs over the others. The pilot pollution creates an abnormally high level of interference, which is likely to result in the performance problems Increased interference level Poor service quality, decreased throughput or increased delay Decreased service access Frequent changes in Active Set and potential risk for unnecessary handovers. Higher non-controllable load

Pilot pollution

The yellow dots represent points where 4-5 CPICHs were received within 6dB window As Active Set size is typically 3, in this situation the rest of the Pilots produce unnecessary interference

Pilot pollution

Pilot pollution can be (at least partly) avoided by planning the CPICH powers and SHO parameters so that throughout the network there is only 2-3 CPICHs available for the UEs, strong enough to be included in the Active Set. All CPICH outside Active Set should be clearly weaker Antenna design, height and tilt are selected carefully Balanced UL & DL SCH/DCH power adjustments

- 11 -

Neighbour cell relations

The Monitored Set is also called as a Neighbour List. This list can be defined in network planning and it can be later changed in network optimization. The list of neighbours play an important role since WCDMA is interference limited. Insufficient planning of neighbour relations will lead to unnecessary high interference E.g. if suitable SHO candidate is not in the monitore set and thus it is not selected to active set then its turning to a pilot poIIuter On the other hand, unnecessary neighbours increase signalling and effects the SHO selection negatively Accurate neighbour relations planning is much more important than in GSM In GSM it is possible to hide cell planning mistakes by frequency planning, in CDMA the such inaccuracies will effect the system capacity The effort saved in frequency planning is spent in more detailed cell planning

Neighbour cell relations

The parameters to control the neighbour relations and the algorithms how system evaluates neighbours for cell lists, depend on vendor minimum CPICH RSCP or Ec/lo Ec/lo margin maximum number of neighbours A neighboring set (or monitored set) is defined for each cell Utilise planning tools automatised functions and check with drive tests Optimise according to CPICI-l coverage and SHO parameters UE monitors the neighboring set that may contain Intra-frequency monitored list: Cells on the same WCDMA carrier (Soft HO) Inter-frequency neighbor list: Cells on another WCDMA carrier (hard HO) Inter-system neighbor list: For each neighboring PLMN Missing neighbour can be detected during drive tests If the best cell shown in the 3G scanner does not enter the active set missing neighbour
- 12 -

Incilude the missing cell to neighbour list if its wanted to active set or change cell plan if FIO

SHO optimisation

Soft/Softer HO planning and correct operation is one of the most important means of optimizing WCDMA networks The importance is high because of the high biterate (pathloss sensitive) and RT (delay sensitive) RABs SHO is measured in terms of probability, the percentage of all connections that are in SHO state The probability is effected by network planning and parameter settings

SHOs have effect to the network performance Advantages Required to avoid near-far effects Coverage increases when more distant users can connect Capacity can be increased if more users can be connected Alongside with PC, SHO is the main interference migitation means in WCDMA Inconvenient Requires more connections, thus eats DL transmission power and decreases capacity Introduces more interference to DL Increases the traffic in lub 40% SHO probability 1.4 times the traffic!

SHO optimisation

- 13 -

SHO optimisation

Probability for soft HO should be set to 30-50% and for softer HO to 5-15%, depending on the area Too high SHO% results in excess overlapping between cells > other-cell interference increases > capacity decreases Too high SHO% also leads to poorly utilised network capacity (unnecessary links) With too low SHO% the full potential of network is not utilised and transmission powers cannot be minimized > trouble with interference SHO performance is planned with a planning tool and optimised by measurements in live network. In early stage SHO% can be planned high, since the traffic density is smaller. With increasing traffic coverage decreases and SHO areas become smaller. SHO% can be tuned with related parameters and dominance areas SHO most important in urban areas due to serious shadowing

Summary KPI
Indicator Coverage Interference Cell overlap Qualitative KPI KPI target example Measured RSCP > -88 dBm over 97% of area (value should be adapted based on required margins) Measured Ec/No > -9 dB over 95% of area Cell overlay < 3 cells over 95% of area Cell Overshoot No cell detected above -111 dBm (CPICH RSCP) Integrity of cell No cell fragmentation detected coverage Best server plot Clean boundary without unnecessary change of best server

- 14 -

Quality of Service

- 15 -

QoS (ITU-T): << The collective effect of service performance which determines the degree of satisfaction of a user of the service>>. Network Performance, NP (ITU-T): << The ability of a network portion to provide the functions related to communication between users>>.

Quality of Service definitions (1)

User domain: throughput, accuracy, dependability (reliability, availability), Provider domain: delay, loss, utilisation,
User QoS Requirements QoS offered by Provider

Quality of Service definitions (2)

QoS experienced By Users

QoS achieved by User

QoS and NP, Performance network (ITU Rec. E800)

Quality of Servicedefinitions (3)

- 16 -

Quality of Service and user satisfaction

Commercial offer Competition User expectations in terms of QoS Users satisfaction Technical QoS

Trends

Non-technical QoS

Network performance

Terminal performance

Sales points

Customer care

Radio Access Bearer QoS

- 17 -

Radio Access Bearer

Main task of the UTRAN is to create and maintain RAB for communication between UE and CN. RAB is build up in order to give for CN elements an illusion about fixed communication path to UE. The network builds up the end-to-end QoS connection from small pieces, which compose a complete chain without bottlenecks These pieces are called Bearers When the connection is set up, the network elements negotiate the QoS requirements of the bearers set up between them The result is a compromise, in which the QoS requirements and networks capacity is taken into account.

UMTS QoS Classes

Traffic Class Conversation class Streaming class Interactive class Background class

Example application Speech and video calls Real-time streaming video Web surfing File downloading, e-mails

UMTS QoS Classes

Traffic Class Properties Conversation class Streaming class Interactive class Background class Minimum fixed delay, no buffering, symmetric traffic, guaranteed bit rate Minimum variable delay, buffering allowed, asymmetric, guaranteed bit rate Moderate variable delay, buffering allowed, asymmetric traffic, no guaranteed bit rate Big variable delay, buffering allowed, asymmetric traffic, no guaranteed bit rate

- 18 -

UMTS QoS Parameters

Parameter Maximum bit rate Guaranteed bit rate Explanation Defines the maximum bit rate when delivering information between end points of UMTS bearer (<2Mbps) Defines the bit rate that the UMTS bearer must carry between its end points

Allowed transfer delay Set the limits for delay (>80ms) QoS negotiable QoS of some services are not negotiable (speech), packet data services admit various QoS classes

Some values of QoS UMTS parameters classes

Traffic class Conversation Streaming Interactive <2048 Yes/No <_ 1500 or 1502 Yes/No 5* 10-2, 10-2, 4* 10-3, 10- 4* 10-3, 105, 5, 5* 10-3, 10-3, 10-4, 10-5, 10-6 6* 10-8 6* 10-8 10-1, 10-2, 10-3, 10-4, 10-3, 10-4, 7* 10-3, 10-3, 10-6 10-6 10-4, 10-5 250 maximum value Background Maximum throughput (kb/s) Scheduling Max. SDU size (octers) Corrupted SDU delivery Residual BER 5* 10-2, 10-2, 5* 10-3, 10-4, 10-6 SDU error rate 10-2, 7*10-3, 10-3 10-4, 10-5 Transfer delay 100 maximum (ms) value

- 19 -

QoS Negotiation
UE
E2E service request Maximum bit rate Guaranteed bit rate Transfer delay QoS negotiable (y/n)

UTRA (NB, RNC)

CN

UMTS bearer service: Request for UMTS QoS Class Maximum bit rate Guaranteed bit rate Transfer delay QoS negotiable (y/n) RAB assignment request RRM: Admission control

Radio bearer and radio link establishment

QoS negotiation RAB assignment response

UMTS Bearer service with negotiated QoS

QoS in UMTS
In early UMTS Release 99 all conversational and streaming class traffic were offered over the CS bearer Voice RT multimedia (e.g. videotelephony) In early Release 99 only Interactive and background class traffic utilisises the PS bearer Release 4 capable networks introduce some streaming class traffic on PS bearer as well Release 5 brings along a full portofolio of PS bearers also utilised for conversation traffic

- 20 -

QoS in UMTS

The QoS over the air interface is implemented by matching each radio bearer with a transport channel whose format set defines the QoS parameters The mapping is performed during the establishment of the RAB RNC performs the mapping of RAB characteristics to actual resource requirements (vendor dependent) Example of mapping for web service, which belongs to the interactive class
Interactive Class
128 kbps 1500 10 -6 NA 64 kbps yes 1% NO
^

Parameters
Maximum bit rate Maximum SDU size Residual BER Transfers Delay Guaranteed bit rate Delivery order SDU Error Ratio Delivery of errorneous

Radio Resource mapping

SF=16 Map to Transport formats 1/3 turbo encoder Interleaver=40 or 80 msec SF=16 Use Acknowledged RLC Set appropriate threshold for Use Acknowledged RLC

QoS in UMTS

Operators can define the wanted QoS profile (in HLR) per subscriber Users can be categorised (QoS differentiation) for various tariffing schemes Traffic handling priorities can be set (THP)

Business Remote office Basic free time Traffic class All four allowed All four allowed Only converational (voice calls) and background Max bit rate 400 kbps 800 kbps 64 kbps Guaranteed bit 384 kbps 64 kbps 12 kbps rate Allowed THPs THP 1 (e.g. for THP 2 (e.g. for THP 3 e-mail file tranfer) download)

- 21 -

QoS in UMTS

- 22 -

Measurement and statistics collection

- 23 -

Measurement tools typology

Field Measurements Generic Measurement tools System Measurements OMC counters Specific Passive capture tools Calls generators

A. Field measurements Drive test equipment

measures and softwares

GPS

External antennas

Controler

Energy
Processing

Mobile QoS test equipment

Man to machine interface

Poor Coverage example

- 24 -

Poor Cell Dominance example

Pilot Pollution example

- 25 -

Example of neighbor missing (Ec/Io)

Corner effect (Ec/Io)

- 26 -

QVOICE
PSTN / ISDN Cellular Network

QVS

QVP-Server

Data collection Post processing QVP-Client

3 parts: QVM (QV Mobile), QVS (QV Stationary) et QVP (QV Post processing).

B. System measurements OMC measurements

Specific Alcatel: RNO Siemens: SPOTS Ericsson: TEMS Analyser Generic APIC from Metrica MyCom from MYCom AirCom NetAct SQM: Nokia OVPI: HP (for IP equipments)

- 27 -

KPI processing tools

Commercial tools:

Analyzer (Tektronix)

BiVision ADC/Metrica, NetAct (Nokia, for 3G) UTRAN Network and service Actix

Analysis tools using these counters (generally they are specific). Example: RNO or NPA of Alcatel, SPOTS from Siemens, etc.

Analysis based on OMC-R counters

- 28 -

 HP: Ovis (data services tests, producers KPIs). RamCom: Network Consultant (A, Gb, Gi, Gn, Iub, Iur, Gi and Gn interfaces) Trafica (NetAct from Nokia) Ipanema: Ipanema (2,5 G and 3G data traffic). Cigale (Astellia): 2 and 3G traffic.

Passive tools examples

GIS display

- 29 -

K PI

- 30 -

Optimization process
Performance measurements Update of parameters, site configuration Key Performance Indicators (KPI)

Network tuning

Reasons that lead to otimisation: Improve the performance Business reasons (cost-effective) Troubleshooting

Performance analysis

Network statistics

Network statistics are collected from different network elements with counters Different types of counters are used KPIs are needed to provide information of the network performance Raw counter data too detailed to be used in monitoring and optimisation (Some counters can be used as KPIs)

KPI definition

KPIs are composed from several counters KPI categories Accessibility Retainability Integrity Documentation of KPIs is important Same KPI can be defined from different counters or formula can be incorrect Measurement period must be reasonable Too much averaging if too long Not enough statistical information if too short
- 31 -

KPI Example

Optimisation based on KIPs: Optimisation is performed for each category Find the worst performing cells Find the reasons behind the poor performance Make the changes in the network Monitor the performance after the changes

- 32 -

Accessibility

- 33 -

Accessibility : call set-up

MS Originating Call Setup Random Access RRC Connection Setup Service Request Authentication Security

RAB Assignment

Accessibility workflow
Other Modules Performance Measurements
Alarms Cell Availability Counters

Performance Analysis
Idle mode RRC Connection Random Access NAS RAB Assignment

Recommendation & Implementation

Verification of changes

Squal, Srxlev, qQualmin, qRxLevMin, maxTxPowerUl, t3212, t3312, aichPower, powerOffsetP0, preambleRetransMax, constantValueCprach

pmTotNoRrcConnectCsSucc pmTotNoRrcConnectPsSucc pmNoRabEstablishAttempt<RAB> pmNoRabEstablishSuccess<RAB> pmNoPageDiscardCmpLoadC pmNoPagingAttemptUtranRejected

Worst performing cell for CS and PS

100
pmTotNoRrc ConnectReq CsSuccess pmTotNoRrc ConnectReq Cs

pmTotNoRab EstablishS uccess < RAB >

pmmTotNoRa bEstablish Attempt < RAB >

100

pmTotNoRrc ConnectReq PsSuccess pmTotNoRrc ConnectReq Ps

pmTotNoRab EstablishS uccess < RAB > pmmTotNoRa bEstablish Attempt < RAB >

- 34 -

Service success set up rate (CS)

Speech
100x pmTotNoRrcConnectReqCsSucc x pmTotNoRrcConnectReqCs pmTotNoRabEstablishSuccessSpeech pmmTotNoRabEstablishAttemptSpeech

Circuit-Switched 64
100x pmTotNoRrcConnectReqCsSucc x pmTotNoRrcConnectReqCs pmTotNoRabEstablishSuccessCS64 pmmTotNoRabEstablishAttemptCS64

Circuit-Switched 57
100x pmTotNoRrcConnectReqCsSucc x pmTotNoRrcConnectReqCs pmTotNoRabEstablishSuccessCS57 pmmTotNoRabEstablishAttemptCS57

Service success set up rate (PS)

Packet-Switched Data Streaming
100 x (Y) x pmTotNoRabEstablishSuccessPacketStream + pmRabEstablishSuccessPacketStream128 pmTotNoRabEstablishAttemptPacketStream + pmRabEstablishAttemptPacketStream128

Where Y =

pmTotNoRrcConnectReqPsSuccess PmTotNoRrcConnectReqPs

Packet-Switched Data Interactive

pmTotNoRabEstablishSuccessInteractive 100 x (Y) x pmTotNoRabEstablishAttemptPacketInteractive + HS1_HardHO_Flow

Where
Yx pmTotNoRrcConnectReqPsSuccess pmTotNoRrcConnectReqPs

HS1_HardHO_Flow= pmNoOutgoingHsHardHoAttempt pmNoHsHardHoReturnOidSource -pmNoIncomingHsHardHoAttempt - pmNoHsHardHoReturnOldChTarget

- 35 -

Idle mode paging

Successful First and Repeated Page attempts of total number of first attempts, Paging success rate in aMSC
100 x NPAAG1RESUCC + NPAG2RESUCC NPAG1GLTOT + NPAG1LOTOT

Paging intensity per cell in a RNC (if RNC, LA and RA consist of exact same cells):
pmCnlnitPagingToldleUeLa + pmCninitPagingToldieUeRa + pmCnlnitPagingToldleUe Measurement period x total number of cells in that RNC

Number/percentage of false detections, which is the case that preamble is detected but there is no enough energy in message part, due to noise on the random access channel for a carrier (it could be due to loss of AICH, wrong recognition of preamble or loss of RACH message part after the UE sends message out):

Random access: preamble detection

pmNoPreambleFalseDetection or pmNoPreambleFalseDetection x100% pmPositiveMessages

Percentage of getting AICH but no RRC connection setup, excluding cell (re)selection:

Random access: AICH detection

No of AICH_ACK-No of RRC connection setup-No of cell (re)selection during RRC establishment No of AICH_ACK

x100%

- 36 -

Admission control: DL transmission carrier power

DL transmission carrier power
Average DL TX power for a cell-carrier:
102 =0 102 =0

pmTransmittedCarrierPoweri x

i 2

pmTransmittedCarrierPoweri

Admission control: UL RSSI

Average UL RSSI for a cell-carrier:
62 =0 62 =0

[pmAverageRssii x (0.5xi 110.5)]

pmAverageRssii

Admission control: Air Interface Speech Equivalent (ASE)

Average UL ASE for a cell:

pmSumOfSampleAseUI pmNoOfSampleAseUI

Average DL ASE for a cell:

pmSumOfSampleAseDI pmNoOfSampleAseDI

- 37 -

Admission control: code allocation

pmNoDIChCodeAllocFailureSF128 pmNoDIChCodeAllocAttemptSF128 x100%

Code allocation failure for SFn, where n is the spreading factor for a cell could be found in the following formula (as an example the SF 128 was used):

Admission control: compressed mode

pmSumCompMode pmSampesCompMode

How many users are in compressed mode? Well the average number of users in compressed mode for a cell:

Admission control: load sharing

Ratio between RRc connection returning and redirection due to load sharing for a cell:

pmNoOfReturingRrcConn pmNoLoadSharingRrcConn

The failures can be observed by the successful rate of directed retry to GSM for a cell:

pmNoDirRetrySuccess pmNoDirectionRetryAtt

x100%

- 38 -

Retainability

- 39 -

Service retainability workflow

Other Modules Performance Measurements Performance Analysis Recommendation & Implementation
ReleaseConnOffset maxTxPowerUl,SirMax, MinPwrRl, treselection, timetotrigger1, reportingrange1

Verification of changes

pmSystemRabRelease<RAB> pmNormalRabRelease<RAB> pmNoSysRelSpeechULSynch pmNoOfTermSpeechCong pmNoSysRelSpeechSoHo UL out of Synch Congestion control, SHO functions IFHO functions IRAT Handovers

Dropped call rate CS

Speech
100x
pmNoSystemRab ReleaseSpeech (pmNoNormalrAB ReleaseSpeech + pmNoSystemRab ReleaseSpeech)

Circuit-switched 64
100x
pmNoSystemRabReleaseCs64 (pmNoNormalRab ReleaseCs64 + pmNoSystemRab ReleaseCs64)

Circuit-switched Streaming
100x
pmNoSystemRabReleaseCsStream (pmNoNormalrRab ReleaseCsStream + pmNoSystemRab ReleaseCsStream)

Dropped call rate PS

Packet Switched data Streaming
100x pmNoSystemRabReleasePacketStream + pmNoSystemRabReleasePacketStream128 (pmNoNormalRabReleasePacketStream + pmNoSystemRabReleasePacketStream + pmNosystemRabReleasePacketStream128)

Packet Switched data Interactive

100x pmNoSystemRabReleasePacket (pmNoNormalRabReleasePacket + pmNoSystemRabReleasePacket)

- 40 -

Minutes per drop CS

Speech
100x Sp_U_User pmNoSystemRabReleaseSpeech x number of minutes

Circuit-switched 64
100x Cs64_U_User pmNoSystemRabReleaseCs64 x number of minutes

Circuit-switched Streaming
100x

x number of minutes pmNoSystemRabReleaseCsStream

Cs57_U_User

Minutes per drop PS

Packet Switched data Streaming
100 Pstr_P8_U_User pmNoSyatemRab ReleasePacketStream + pmNoSystemRab ReleasePacketStream128 x No of minutes

Packet Switched data Interactive

100 Plntdch_U_User+PlntHs_U_User+PlntFach_U_User pmNoSystemRab ReleasePacket x No of minutes

Packet Switched data Interactive HS

100 PlntHs_U_User pmNoSystemRab ReleaseHs x No of minutes

Handover failure rate

100x

The following formula shows the failure rate for RL addition/replacement to active set
pmNoTimesCellFailAddToActSet (pmNoTimesCellFailAddToAct + pmNoTimesRlAddToActSet

- 41 -

Handover failure rate: HS cell change

100x pmHsCcSuccess pmHsCcAttempt

The following metric measures the success rate for HS Cell Change in target cell

Handover failure rate: out of synchronization

100x (pmNoSysRelSpeechULSynch)

Shows fraction of drop due to uplink Out of Sync reason.

(pmNoNormalRabReleaseSpeech + pmNoSystemRabReleaseSpeech)

Handover failure rate: missing neighbor

100x pmNoSysRelSpeechSoHo (pmNoSystemRabReleaseSpeech + pmNoNormalRabReleaseSpeech)

Shows fraction of speech drop due to HO action when a valid or non-valid cell cannot be added to active set. This includes also drop due to missing neighbour.
100x pmNoSysRelSpeechNeighbr (pmNoSystemRabReleaseSpeech + pmNoNormalRabReleaseSpeech)

Shows fraction of speech drop due to missing neighbour reason when a non-valid cell cannot ne added to active set.

- 42 -

Drop due to IHO failure for speech: Outgoing IFHO failure when UE failed to return to present active set.

Inter-frequency handover failure rate (CS)

100x

pmFailNonBlindInterFreqHoFailRevertCsSpeech12 pmAttNonBlindInterFreqHoCsSpeech12

Inter-frequency handover failure rate (PS)

Drop due to IFHO failure for PS less or equal to 64 kbps: Outgoing IFHO failure when UE failed to return to present active set.
100x pmFailNonBlindInterFreqHoFailRevertPsInteractiveLess64 pmAttNonBlindInterFreqHoPsInteractiveLess64

Drop due to IFHO failure for PS greater than 64 kbps: Outgoing IFHO failure when UE failed to return to present active set.
100x pmFailNonBlindInterFreqHoFailRevertPsInteractiveGreater64 pmAttNonBlindInterFreqHoPsInteractiveGreater64

Drop due to IFHO failure for PS streaming and others: Outgoing IFHO failure when UE failed to return to present active set.
100x pmFailNonBlindInterFreqHoFailRevertStreamingOther pmAttNonBlindInterFreqHoStreamingOther

- 43 -

IRAT handover

The following metric measures hard handover success rate between UtranCell and target GSM cell for speech calls. The formula is considering the GsmRelation.
100x pmNoSuccessOutIratHoSpeech pmNoAttOutIratHoSpeech

The following metric measures hard handover success rate between UtranCell and target GSM cell for streaming calls. The formula is considering the GsmRelation.
100x pmNoSuccessOutIratHoCs57 pmNoAttOutIratHoCs57

IRAT handover

The following metric measures hard handover success rate between UtranCell and target GSM cell for Multi-RAB calls. The formula is considering the GsmRelation.
100x pmNoSuccessOutIratHoMulti pmNoAttOutIratHoMulti

The following metric measures cell change failure rate between UtranCell and target GSM cell for PS calls when the UE successfully returns to UtranCell. The formula is considering the GsmRelation.
100x pmNoOutIratCcReturnOldCh pmNoOutIratCcAtt

Congestion
100x

Shows fraction of speech drop due to cogestion action

(pmNoOfTermSpeechCong) (pmNoNormalRabReleaseSpeech + pmNoSystemRabReleaseSpeech)

Shows fraction of video call drop due to cogestion action

100x (pmNoOfTermSpeechCong) (pmNoNormalRabReleaseCs64 + pmNoSystemRabReleaseCs64)

- 44 -

Integrity

- 45 -

Service integrity workflow

Performance Measurements Performance Analysis Recommendation & Implementation Verification of changes

BLER counters and Down Switching counters

BLER, power, SIR parameters Throughput

Test the settings

Check statistics If not OK, roll back

pmFaultyTransportBlocksBcUl pmTransportBlocksBcUl pmNoOfSwDownNgCong PmNoOfSwDownNgAdm PmDl Traffic volume counters

The method for finding worst performing cells is based on top to down analysis. Initial worst 10-15 performing cells can be identified based on the Uplink Block Error rate before combining. 100x pmFaultyTranspoertBlocksBcUL pmTransportBlocksBcUI

BLER

Throughput
Throughput =

Average throughput per cell and RAB in the DL, excluding HSDPA:
pmDlTrafficVolume<RAB> pmSum<RAB>RabEstablish pmSamples<RAB>RabEstablish *ROPsec

RAB efficiency excluding HSDPA

The RAB efficiency can also be checked RABEfficiency =
UeRc stands for different RABs UeRc=2, Speech UeRc=3, Video Call UeRc=4, Packet Common Channel UeRc=5, PS 64/64 UeRc=6, PS64/128 UeRc=7, PS 64/384 UeRc=10 multirab Speech+PS 0 or PS 64/64).

Actual Bitrate per RAB Nominal Bitrate per UeRc=x

- 46 -

Payload counters
Radio UL Payload counter DL Payload counter Connection type Speech pmUITrafficVolumeCs12 pmUITrafficVolumeCs12 PS64/64 pmUITrafficVolumePs64 pmUITrafficVolumePs64 PS64/128 pmUITrafficVolumePs128 pmUITrafficVolumePs128 PS64/384 pmUITrafficVolumePs384 pmUITrafficVolumePs384 CS 57.6 pmUITrafficVolumeCs57 pmUITrafficVolumeCs57 (streaming) CS 64 (UDI) pmUITrafficVolumeCs57 pmUITrafficVolumeCs57 Speech/PS pmUITrafficVolumeCs12Ps64 pmUITrafficVolumeCs12Ps64 64 multirab PS pmUITrafficVolumePsCommon pmUITrafficVolumePsCommon Common

- 47 -

HSDPA-HSUPA

- 48 -

HSPA Key features-AMC

Cood CQI Bad CQI

High Code Effective Rate

Low Code Effective Rate

Channel Quality Feedback (CQI) UE measures channel quality (SNR or Ec/No) and reports to Node B every 2ms or longer time. Node B chooses modulation scheme, transport block size and code effective rate based on CQI

Good coverage

Bad coverage

AMC could improve radio bandwidth and fit for high speed radio transmission.

HSPA Key feature System resource

Resource allocation Reasonable resource allocation can improve throughput performance
Minimum Code available for HSPA but not for R99, so this resource cant allocated too much to avoid no code for HSPA access channel.

OVSF Code resource

DCH Code HSPA Maximum For R99 and HSPA access and traffic channel Code

HSPA Minimum Code

Power resource

Threshold for R99 load control, which should not be allocated too much to avoid no power for HSPA user

- 49 -

HSPA Key feature-Scheduling

Transmission slot 2 ms UE2 UE1

UE3

Data transmission slot

Mobility management

- 50 -

Sample message flow to begin HSDPA operation

UE
1

Bode B

RNC

UTRAN decides to start HSDPA for the UE

Measurement Control message (Setup Event 1d) 2 Measurement Report message (Report Event 1d) Radio link reconfiguration prepare 3 Radio link reconfiguration ready Radio link reconfiguration commit Radio Bearer Reconfiguration message 4

User plan data can flow on the HS-DSCH

Radio Bearer Reconfiguration Complete message

Sample message flow to stop HSDPA operation

UE Bode B RNC
1

UTRAN decides to stop HSDPA but keep DCH for the UE due to: Low Downlink data activity High UE mobility

Radio link reconfiguration prepare Radio link reconfiguration ready Radio link reconfiguration commit Radio Bearer Reconfiguration message

User plane data only flows on the DCH

Radio Bearer Reconfiguration message

- 51 -

Cell change triggering with event 1d

Ec/No Cell 1
Hysteresis

Cell 2

Time-to-

Cell change

HS-DSCH on Cell 1

HS-DSCH on

Time

Differences about information collection between R99 and HSPA

- 52 -

HSPA dimensioning optimization

HSPA Optimization target Improve CQI

HSPA RF optimization target R99 RF optimization target

CQI

RSCP & Ec/Io

Make sure that CQI is distributed as appropriate proportion. Cell edge throughput requirement could be fulfilled in the door coverage area.

Make sure that cell gets target coverage probability

CDI User experience Equipment Ec/Io

9>CQI Poor -15dB

15>CQI 9 Fair -15 dB - -9dB

CQI 15 Good -9dB

- 53 -

Performance Optimization Procedure

HSPA CALL SUCCESS RATIO HSPA HANDOVER SUCCESS RATIO HSPA THROUGHPUT PERFORMANCE KPI Collection DT/CQT/Statistics

The three kind of KPI of should be paid more attention

Performance optimization

Modify the load balance policy Carry out smart admission Power and lub bandwidth
algorithm such as DRD or Downsize Access

Problem Analyze

Code congestion Power overload HSPA subscriber lub bandwidth Unsupported

congestion number overload

The higher bit rate such as

congestion means the capacity should be expanded

13.6K DCH channel can help to improve access performance

configuration by UE

Inter-freq Handover is the most difficult for HSPA mobility optimization

- 54 -

HSPA Throughput Optimization

The Power and Codes available for HS-PDSCH, the lub bandwidth and RF environment of UE position will all impact throughput user got.
Available Codes for HSPA Available Power for HSPA

Available lub backhaul for HSPA The Data Power to be transmitted Node B RNC

The data should be delivered

DATA DATA DATA

CQI CQI CQI

Check point

The air environment condition CQI

GSN Process ability

DATA

The web sever performance. App Sever Firewall

SGSN/GGSN

- 55 -

- 56 -