Sei sulla pagina 1di 88

HSPA Optimization

Basic HSDPA parameters


queueSelectAlgorithm blerQualityTargetUl
featureStateHsdpaFlexibleScheduler eulMaxRotCoverage
cqiAdjustmentOn eulMaxOwnUuLoad
deltaCqi1 eulOptimalNoiseFloorEstimate
deltaCqi2 eulNoiseFloorLock
deltaAck1 hsMeasurementPowerOffset
deltaAck2 numHsScchCodes
deltaNack1 hsScchMaxCodePower
deltaNack2 maxNumHsPdschCodes
codeThresholdPdu656 numHsPdschCodes
sirMax sirMaxTti2
eulTargetRate eulDchBalancingLoad
eulDchBalancingOverload eulDchBalancingReportPeriod
eulDchBalancingTimerNg releaseAseUlNg
Basic HSDPA
queueSelectAlgorithm
SPI: 15
SPI: 14 Round Robin
SPI: ..
Max CQI
SPI: ..
SPI: 11 Proportional Fair
SPI: 10
HSDPA Equal Rate
SPI: 9
It is recommended to use the
SPI: 8 proportional fair and Max CQI
SPI: .. algorithms together with the CQI
adjustment feature, in order to
True or False SPI: .. achieve a throughput gain compared
Default= False to round
SPI: 1 robin.

SPI: 0

FeatureStateHsdpaFlexibleScheduler
cqiAdjustmentOn

› cqiAdjustmentOn
Since the accuracy of the CQI reporting depends on the UE
vendor and model, it is recommended to enable the CQI
adjustment feature so that both cell and user throughput
are maximized. This is particularly important when
proportional fair or Max CQI is selected as scheduling
algorithm.
› The method to evaluate cqiAdjustmentOn is to monitor
RBS throughput KPIs
CQI Adjustment
RBS throughput?
High number of NACKs
XXXXXX
X
High CQI √ = 10%
reporting UE
High retransmission rate cqiAdjustmentOn = TRUE
=> low throughput

High number of ACKs

√√√√√√ Since the accuracy of the CQI reporting


depends on the UE vendor and model, it
Low CQI is recommended to enable the CQI
reporting UE adjustment feature so that both cell and
Low utilization user throughput are maximized. This is
particularly important when proportional
=> low throughput
fair or Max CQI is selected as scheduling
algorithm.
Delta Parameters
› The quality of the HS-DPCCH can be adjusted with the
parameters deltaCqi1, deltaCqi2, deltaAck1, deltaAck2,
deltaNack1 and deltaNack2. There are two sets of
parameters for the information sent on HS-DPCCH, one
used in soft handover and the other in non-soft handover.

A-DCH
HS-DSCH

RBS B
RBS A
HS-DPCCH Delta
Parameters HS-DPCCH
ACK NACK CQI

deltaAck1 deltaNack1 deltaCqi1


UE Power
TCP ACKs
A-DCH

To maximize the performance for an HSDPA user, the


power allocated to these channels should be increased.
This can be done by reducing the power on the UL
control plane by adjusting the deltaAck1,deltaNack1
and deltaCQI1 leaving more power for the dedicated
channel. A 1 dB coverage improvement has been seen
in field by setting these values to deltaAck1=3,
deltaNack1=3 and deltaCQI1=3.
BLER Target

blerQualityTargetUl (-63 to 0)
blerQualityTargetUl = 10Log(BLER quality target)
E.g. BLER Quality Target = 1% (0.01) *default
=> blerQualityTargetUl = 10log(0.01) = -20

Changing blerQualityTarget to -30 (0.1% BLER):


• 10 to 15% Downlink Throughput improvement for 384/HS
• 5% Downlink Throughput improvement for 64/HS
HSDPA Throughput Key
Performance Indicators
PS Interactive HS Average Throughput [kbit/s]: HSDPA

8 X 1000 X (pmSentPacketDataHs3 + pmSentPacketDataHs4)


pmTotalPacketDurationHs3 + pmTotalPacketDurationHs4

Measured on an RNC level

PS Interactive HS User Throughput, Downlink [kbit/s]: High number


of users
pmSumHsDlRlcUserPacketThp brings
average down
pmSamplesHsDlRlcUserPacketThp

Measured on an Cell level


HSDPA Throughput KPI
counters
Managed Counter
Counter Name Description
Object Type

Accumulated amount of user data transmitted on the downlink


in bursts of size between 10 kilobytes and 100 kilobytes for all
pmSentPacketDataHs3 RncFunction Interactive RABs on HSDPA channels. PEG
Unit: kB (1000 bytes)

Accumulated amount of user data transmitted on the downlink


in bursts of size greater than 100 kilobytes for all Interactive
pmSentPacketDataHs4 RncFunction RABs on HSDPA channels. PEG
Unit: kB (1000 bytes)

Accumulated time that data has been transmitted in bursts of


size between 10 kilobyte and 100 kilobytes for all Interactive
pmTotalPacketDurationHs3 RncFunction RABs on HSDPA. PEG
Unit: msec

Accumulated time that data has been transmitted in bursts of


size greater than 100 kilobytes for all Interactive RABs on
pmTotalPacketDurationHs4 RncFunction HSDPA. PEG
Unit: msec

Sum of all sample values recorded during a ROP for HS-DSCH


pmSumHsDlRlcUserPacketThp Hsdsch user RLC throughput measurements. TrigACC
Unit: kbps (1000 bits per second)

Number of samples recorded within the ROP for


pmSamplesHsDlRlcUserPacketThp Hsdsch TrigScan
pmSumHsDlRlcUserPacketThp.
RBS HSDPA Analysis
Counters
pmNoActiveSubFrames
Number of active 2 ms subframes.
pmCapAllocIubHsLimitingRatioSpiXX
pmRemainingResourceCheck[0] Where XX=[0.15] indicates in what degree the
HSDPA traffic in downlink is limited by the
Incremented for HS-SCCH code shortage
Iub/Iur interfaces, between SRNC and RBS.
pmRemainingResourceCheck[1]
Incremented for HS-PDSCH code shortage

pmRemainingResourceCheck[3]
Incremented for HS-PDSCH power shortage
RBS HSDPA Analysis
Counter Details
Counter
Counter Name Managed Object Description
Type

pmNoActiveSubFrames HsDschResources Number of active 2 ms subframes. PEG

The reason why it is not possible


to schedule another high-speed
user for immediate traffic.
pmRemainingResourceCheck HsDschResources PDF ranges: PDF
[0]: HS-SCCH code shortage
[1]: HS-PDSCH code shortage
[2]: HS-PDSCH power shortage

Indicates in what degree the


HSDPA traffic in downlink for
Scheduling Priority Indicator
pmCapAllocIubHsLimitingRatioSpiXX IubDataStreams PEG
(SPI) XX is limited by the Iub/Iur
interfaces, between SRNC and
RBS.
RBS Throughput Analysis
HS-SCCH Code Limited %: HS-PDSCH
Codes Power
pmRemainingResourceCheck [0]
X 100
pmNoActiveSubFrames
HS-PDSCH Code Limited %:
pmRemainingResourceCheck [1] HS-SCCH
X 100
pmNoActiveSubFrames
HS-SCCH
HS-PDSCH Power Limited %:
HS-SCCH
pmRemainingResourceCheck [2]
X 100 HS-SCCH
pmNoActiveSubFrames

Not Limited %:
pmRemainingResourceCheck [0] + [1] + [2]
1- X 100
pmNoActiveSubFrames
EulMaxRotCoverage
EulMaxRotCoverage can be tuned Total Interference
differently according to R99 traffic 0…550 in 0.1 dB steps
and number of carriers. In networks Default = 80 (8 dB)
with a second carrier mainly used for
HSPA, this parameter can be set to eUlMaxRotCoverage
high values. If only one carrier with
lot of R99 traffic is present, then the
default value is suggested.
Interference Headroom
RSSI and ROT related counters are for Coverage
useful to determine the increase in
UL interference, and together with Interference created by all
R99 accessibility/retainability other sources
performances and EUL throughput, Interference created in own
they can be used to evaluate the cell by DCH and non-
goodness of the parameter setting. scheduled EUL users
Thermal Noise
EulMaxOwnLoad
Own Cell Interference EulMaxOwnUuLoad is used to
0…550 in 0.1 dB steps prevent power rushes that can affect
Default = 30 (3 dB) R99 and EUL performances. It sets
the threshold for the own cell
eUlMaxOwnUuLoad interference determined by the SIR
for all the DCH users connected to
the cell. A low value for this
Own Cell parameter is dramatically affecting
Interference Headroom the EUL throughput performances.
The real load in the field can be
Interference created in own determined by the counter
cell by DCH and non- pmLEDchTot.
scheduled EUL users
EUL Throughput KPI

PS Interactive EUL User Throughput Uplink [kbit/s]:

pmSumEulRlcUserPacketThp
pmSamplesEulRlcUserPacketThp

Measured on an Cell level


EUL Throughput KPI
Counters
Managed Counter
Counter Name Description
Object Type

Sum of all sample values recorded during a


ROP for EUL user RLC throughput
measurements.
Incremented in the best cell in the active set.
pmSumEulRlcUserPacketThp Eul TrigACC
Values are read periodically from an
internal level counter and added to this
counter.
Unit: kbps (1000 bits per second)

Number of samples recorded within the


ROP for pmSumEulRlcUserPacketThp.
pmSamplesEulRlcUserPacketThp Eul Incremented by one when the value of the TrigScan
internal level counter is added to the
corresponding sum counter.
RBS EUL Analysis Counters
pmLOther
Estimated distribution of the Uplink load caused by inter-cell and external interferers

*pmGrantReductionReasonEul
[0]: TN Congestion

[1]: Uu Congestion Serving

[2]: Uu Congestion Non-serving

[3]: HW Congestion

[4]: Low usage happy

[5]: Low usage unhappy

pmLDch [6]: Inactivity at low grant


Measured load of release 99
DCH in the scheduler [7]: Inactivity at high grant
RBS EUL Analysis Counter
Details
Counter
Counter Name Description
Type

The measured load of release 99 DCH in the scheduler. The load is sampled every 100 msec
and stored in the unitless range below:
[0]: Number of samples in range [0..0.02[
pmLDch [1]: Number of samples in range [0.02..0.04[ PDF

[49]: Number of samples in range [0.98..1]
Where 0 means no load and 1 is maximum possible load

The estimated distribution of the Uplink load caused by inter-cell and external interferers.
The load is sampled every 100 msec and stored in the unitless range below:
[0]: Number of samples in range [0..0.02[
pmLOther [1]: Number of samples in range [0.02..0.04[ PDF

[49]: Number of samples in range [0.98..1]
Where 0 means no load and 1 is maximum possible load

Counter for recording the number of grant reductions made by EUL Scheduler due to
different reasons. The count for each reason is given by a separate bin as below:
[0]: TN Congestion -- Incremented every time a grant reduction is triggered due to a TCI
(TNL Congestion Indication) received by the scheduler.
[1]: Uu Congestion Serving -- Incremented every time an AG overload action is taken due
insufficient scheduling headroom.
[2]: Uu Congestion Non-serving -- Incremented every time a RG overload action is taken
due insufficient scheduling headroom.
[3]: HW Congestion -- Incremented every time an overload action is taken due insufficient
pmGrantReductionReasonEul UL HW. PDF
[4]: Low usage happy -- Incremented every time a happy user's grant is reduced due to
consistently low usage of the granted rate.
[5]: Low usage unhappy -- Incremented every time an unhappy user's grant is reduced due
to consistently low usage of the granted rate.
[6]: Inactivity at low grant -- Incremented when a user's grant is reduced from the minimum
UL HW allocation to zero due to complete data inactivity on the UL.
[7]: Inactivity at high grant Incremented when a user with a high grant is reduced to the
minimum UL HW allocation due to data inactivity on the UL.
[8]: Rescheduling -- Incremented every time a user's rate is reduced due rescheduling in the
scheduler.
Optimization Areas
› Basic HSPA parameters
› HSPA mobility and traffic management
› Channel switching
› Downlink power allocation
› Capacity management
› Advanced HSPA
› Scheduler Efficiency
› HSPA Quality of Service
› Multicarrier HSPA
Throughput of PS Interactive
Users

› The throughput of PS interactive data users can be


monitored by means of the following GPEH event:
› 441 INTERNAL_PACKET_DEDICATED_THROUGHPUT
› This enables the throughput of specific users to be
followed. This information can be also correlated with other
GPEH events in order to get the throughput while the user
is in a specific RAB state (e.g. using HSDPA).
Monitoring events for HSPA
mobility
› The following GPEH events are useful to monitor and optimize HS
mobility:
› 387 INTERNAL_CHANNEL_SWITCHING
› 398 INTERNAL_IRAT_HO_CC_EVALUATION
› 399 INTERNAL_IRAT_HO_CC_EXECUTION
› 422 INTERNAL_IFHO_EXECUTION
› 423 INTERNAL_IFHO_EXECUTION_ACTIVE
› 432 INTERNAL_SUCCESSFUL_HSDSCH_CELL_CHANGE
› 433 INTERNAL_FAILED_HSDSCH_CELL_CHANGE
› 442 INTERNAL_SUCCESSFUL_TRANSITION_TO_DCH
› 443 INTERNAL_FAILED_TRANSITION_TO_DCH
› 438 INTERNAL_SYSTEM_RELEASE
GPEH Event 438 Example

Cells in active set


Interact. PS 16/HS

RRC Connection Release

RCS lost
Monitoring channel switching
performance
› The following GPEH events are useful to monitor and optimize channel
switching performance:
› 19 RRC_RRC_CONNECTION_RELEASE
› 34 RRC_RRC_CONNECTION_REQUEST
› 415 INTERNAL_RAB_ESTABLISHMENT
› 387 INTERNAL_CHANNEL_SWITCHING
› 392
INTERNAL_CONGESTION_CONTROL_CHANNEL_SWITCH_AND
_TERMINATE_RC
Monitoring downlink power
congestion

› 393 INTERNAL_START_CONGESTION
› 394 INTERNAL_STOP_CONGESTION
Monitoring capacity

› The following GPEH events enable the precise calculation of who is


blocked (e.g. guaranteed/non-guaranteed; handover/non-handover)
and what is the limiting resource:
› 391 INTERNAL_ADMISSION_CONTROL_RESPONSE
› 392
INTERNAL_CONGESTION_CONTROL_CHANNEL_SWITCH_
AND _TERMINATE_RC
› 393 INTERNAL_START_CONGESTION
› 394 INTERNAL_STOP_CONGESTION
› 426 INTERNAL_RBS_HW_MONITOR_UPDATE
› 431 INTERNAL_SYSTEM_BLOCK
› 438 INTERNAL_SYSTEM_RELEASE
› 451 INTERNAL_SYSTEM_UTILIZATION
GPEH Event 451 Example

% of baseband capacity used


Events monitoring advanced
HSPA

› 426 INTERNAL_RBS_HW_MONITOR_UPDATE
› 451 INTERNAL_SYSTEM_UTILIZATION
› 438 INTERNAL_SYSTEM_RELEASE
› 431 INTERNAL_SYSTEM_BLOCK
Events for monitoring
scheduler efficiency

› 19 RRC_RRC_CONNECTION_RELEASE
› 34 RRC_RRC_CONNECTION_REQUEST
› 148 NBAP_RADIO_LINK_FAILURE_INDICATION
› 149 NBAP_RADIO_LINK_RESTORE_INDICATION
› 414 INTERNAL_RC_SUPERVISION
› 438 INTERNAL_SYSTEM_RELEASE
Event monitoring HSPA QoS

› 441 INTERNAL_PACKET_DEDICATED_THROUGHPUT
Multi Carrier HSDPA

› New connection properties related to Multi Carrier HSDPA


are now available in GPEH. Information regarding the
secondary serving HS-DSCH cell have been included for
some GPEH events related to Multi Carrier connections
(modified events).
› EVENT_PARAM_SOURCE_C_ID_1_SECONDARY_SER
V_HSDSCH_CELL
› EVENT_PARAM_TARGET_C_ID_1_SECONDARY_SERV
_HSDSCH_CELL
› Example of modified event
› INTERNAL_PACKET_DEDICATED_THROUGHPUT
hsMeasurementPowerOffset
1/2
› hsMeasurementPowerOffset
The channel quality indicator is used by the UE to
signal to the RBS it’s current radio conditions. In
fact it sends the RBS a number relating to the
biggest transport block that the UE thinks it can
successfully decode under the current conditions.
› The UE reports a number between 1 and 30 and
this indicates a maximum transport block size
depending on the UE category.
hsMeasurementPowerOffset
Optimization Hint

If the RBS receives CQI equal to zero from a UE, no data is sent to that
particular UE in the coming TTI and the HSDPA throughput is impacted.
Therefore, it is important that the CQI distribution is not clipped on the lower
end

The throughput performance is also degraded when the CQI distribution is


clipped on the higher end (CQI = 30). The best performance is obtained
when all reported CQI values are distributed between 0 and 30.

The UE calculates the CQI by measuring the received power of the pilot
channel and estimating how large a transport block it could successfully
decode under the existing radio conditions if the same power that was put
into the pilot channel was put into the HSDPA channel
hsMeasurementPowerOffset
2/2

Based on the received power of the CPICH the UE


estimates how large a transport block it could
correctly decode in the prevailing radio conditions.
If the same power as the RBS put into the CPICH
is put into the HSDPA channel(s)

However since the CPICH is only one channel


with a power ranging from about .5 watt to 1 watt
this might mean that the UE will always only
indicate relatively small transport block sizes.
HSDPA channel only using at
most same power as CPICH

Measure
CPICH_RSCP

CPICH 0.5 watt to 1 watt

HSDPA 0.5 watt to 1 watt

Relatively small transport block size


Channel Quality Indicator

Obviously as we can have multiple HSDPA


channels with 0% to 80% power in them the RBS
could at times have much more power available
than just the power of the CPICH to put into the
HSDPA channels.

This would mean the channel quality indicator


could report much higher block sizes
HSDPA channel using whatever
power is available in the RBS

Measure
CPICH_RSCP

CPICH

HSDPA .1 watts to 60 watts

Potentially much larger


transport block
UE Calculation of Transport
Block
So to take advantage of the possibility that the
RBS might have more power than that used in the
CPICH available for the HSDPA channels a power
offset is added to the measured power by the UE
when estimating the power available in the RBS
for the next HSDPA transmission.

PHS_assumed = PCPICH_RX +  +  [dB]

This results in a maximum transport block size


which is indicated to the RBS in the CQI.
Reference Power Adjustment

 is the reference power adjustment used when


the radio conditions are so good that the maximum
block size can be sent with less power than
PCPICH_TX + 

The aim of hsMeasurementPowerOffset or  is to get


the full range of CQI values reported by the UE.
Channel Quality Indicator

› If the RBS has the CPICH power plus the


hsMeasurementPowerOffset available then it can
send the maximum transport block size to the UE.
› If not it scales back the block size it sends based
on the power that it has available.
› A value of 8dB works well for
hsMeasurementPowerOffset.
› The setting of this parameter can be checked by
pmReportedCqi[0].
numHsScchCodes and
hsScchMaxCodePower
› numHsScchCodes
It is recommended to multiplex up to 3 users
› hsScchMaxCodePower
When using HS-SCCH power control, the power
save compared to fixed HS-SCCH is significant.
The average power can be reduced more than 3
dB when using power controlled HS-SCCH. Field
trials have shown that HSDPA throughput
performance is improved, when increasing the
maximum allowed code power for HS-SCCH by 1
dB from -2 dB to -1 dB relative CPICH.
Evaluate power consumption
versus cell throughput

› numHsScchCodes and hsScchMaxCodePower


The method to evaluate power consumption versus cell
throughput is to monitor pmNoOfHsUsersPerTti,
pmSumOfHsScchUsedPwr and the RBS throughput KPIs
Dynamic code allocation
featureStateHsdpaDynamicCodeAllocation
maxNumHsPdschCodes
HS-DSCH
DCH

Mainly R99 use HS TX15 numHsPdschCodes = 1 or 2 and maxNumHsPdschCodes = 10.


Dedicated carrier used HS TX45 numHsPdschCodes = 5 and maxNumHsPdschCodes = 15.

Common
Channels DCH
DCH Load Min - HS-PDSCH
dependent Flexible numHsPdschCodes
increase

Check code tree usage using pmUsedHsPdschCodes


Code ThresholdPdu656

› codeThresholdPdu656 When using bit rates above 6


Mbps it is necessary to increase the RLC PDU size from
336 bits to 656 bits in order to not be throughput limited by
the RLC PDU size.
› If codeThresholdPdu656 is smaller than
numHsPdschCodes the 656 bit RLC PDU size will be
used for Category 7-10 UEs, otherwise the 336 bit RLC
PDU size is used. If Category 10 UEs are used in the
system, and very high downlink throughput is important, it
is recommended to set codeThresholdPdu656 <
numHsPdschCodes.
SirMax and SirMaxTti2

› SirMax and sirMaxTti2 Optimal settings for these


parameters are important to assure that the Outer Loop
Power Control can maintain the wanted retransmission rate
for EUL. Higher setting is needed for Eul 2ms TTI (17.3 dB
suggested). If no EUL enabled then sirMax can be set
lower (recommended is 10dB), while with EUL 10 ms the
recommended value is 12 dB.
eulTargetRate
Scheduled Grant
Left over
User rate
User Data rate

Max step
Recommend < 384kbps

Max step
eulTargetRate

Step up ‘TargetRate’

Minimum hardware allocation

Zero kbps

Time
UE rate increase request Absolute Grant order to UE
DCH/EUL Load Balancing
EUL Scheduled Rate Cell Admission Policies Based on EUL Cell Congestion State

Admission Request Type EUL Cell Congestion State

Uncongested DetectedCongestion HighCongestion

Requests for DCH on lowest Yes Yes Yes


DCH rate admitted
Requests for DCH up switch Yes No No
admitted
Requests for DCH on higher Yes Yes No
DCH rate/EUL admitted

Requests for handover Yes Yes Yes


admitted
HSPA cell change admitted Yes Yes Yes

Request for guaranteed UL Yes Yes Yes


service admitted
DCH/EUL Load Balancing

A higher eulDchBalancingLoad
initiates congestion resolution
actions earlier. A higher
eulDchBalancingOverload
agressively blocks admission
for EUL and DCH users.
DCH/Eul Load Balancing
Parameters
› It is recommended that the license for “Interact. PS (16/HS) + SRB
(3.4/3.4)” (FAJ 121 905) is installed and activated because the UL
resource used by DCH needs to be low in order to make enough room
for EUL users.
› Other parameters that should be considered when deploying this
feature are:
› releaseAseUlNg, amount of ASE in the uplink to be released with each
periodic congestion-resolution action targeting the non-guaranteed
traffic class connections in this cell.
› eulDchBalancingTimerNg, interval between periodic EUL congestion-
resolution actions targeting the non-guaranteed traffic class
connections in this cell.
DCH/Eul Load Balancing
Parameters
releaseAseUlNg

…… Time

eulDchBalancingTimerNg
Improve Cell Edge Coverage
› This feature provides significant improvement of cell edge
coverage for EUL connections with 3GPP Rel 8 UEs. EUL
Cell Edge Coverage will allow the UE to stay on another 2
dB for 10ms TTI users and 4 dB for 2 ms TTI users before
complete loss of coverage. This is done by increasing the
minimum ratio between the E-DPDCH and the DPCCH
power.
Improve data power at cell
edge
› Different gains are obtained
depending on whether event 6d is data
E-DPDCH
enabled (event 6d occurs when
the UE Tx power reaches
maximum during a time at least
equal to timeTrigg6d). With Control
DPCCH
event6d off
HSPA
(txPowerConnQualMonEnabled on UL
= FALSE) full advantage of the
feature is achieved. The relative
coverage gain is larger for 2 ms
TTI than for 10 ms TTI EUL
connections.
HSPA Optimization

Downlink Power
Downlink Power Parameters
› primaryCpichPower
› usedFreqThresh2dRcsp
› usedFreqThresh2dEcNo
› usedFreqRelThresh4_2bEcNo
› hsPowermargin
› maxUserEhichErgchPowerDl
› maxEagchPowerDl
› individualOffset
› maximumTransmissionPower
› extraPowerForSrbOnHsdpa
› extraHsScchPowerForSrbOnHsdpa
Increasing Cell Downlink Power
An optimum balance between coverage and capacity should be a main
goal for R99.

P-CPICH power is set by the parameter primaryCpichPower. The power


level of all other R99 downlink common channels is expressed as an offset
relative to it.

PrimaryCpichPower tuning related activities should be considered in


conjunction with other activities such as Capacity Optimization or Traffic
Management.

Example: primaryCpichPower is increased, then the coverage area of the


cells is increased, which also implies a possible increase of the R99 traffic
carried by the cells.
Increasing Cell Downlink Power

More R99
Increasing Cell Downlink Power
From the HSDPA perspective, it is important to have sufficient P-CPICH
coverage to ensure required CQI measurements.

HINTS
primaryCpichPower is increased  overall amount of unused power
available for HSDPA will be reduced and the average interference level
from R99 might increase, and hence the HS throughput could suffer.

Reducing the cell’s primaryCpichPower has been shown to improve


HSDPA throughput however there are negative implications
Reducing Cell Downlink Power
Reducing the cell’s primaryCpichPower
has been shown to improve HSDPA
throughput, due to less non-HS power
being allocated, therefore increasing the
available power for HS users. However, Lower CQIs

this will cause a degradation in cell’s


CPICH Ec/No, hence degradation in CQI
reported by the UE.
Ec/No Value
Maximum Power
41dBm Ec/No = 27 –
41 = -14 dB

HSDPA This could


trigger
compressed
mode
Common Channels
27dBm CPICH power
Parameters for Compressed Mode

The parameters controlling entry into compressed


mode, IFHO, and GSM HO (such as
usedFreqThresh2dRscp, usedFreqThresh2dEcNo,
usedFreqRelThresh4_2bEcNo, etc) can be
considered for optimization as well if high HS utilisation
is causing RF performance degradations.
Downlink Power
measurements from GPEH
Setting primaryCpichPower
In scenarios where HSDPA and R99 traffic
are mixed, the primaryCpichPower
should be high enough to guarantee that
the maximum Ec/No value in the cell is
well above (e.g. 1-2 dB) the thresholds for
(bad) connection quality monitoring and
IFHO/GSM HO in order to minimize the
impact of HSDPA transmission on R99
traffic.
Downlink Power stability
hsPowerMargin
hsPowerMargin
Total available cell power Pmax

HSPA PHS

Dedicated channels (power controlled)

Pnon-HS

Common channels (not power controlled)

PHS = Pmax – hsPowerMargin – Pnon-HS


PPDSCH = PHS – PHS-SCCH - PEUL
Downlink Power for HSUPA
E-AGCH
maxUserEHichERgchPowerDl
E-RGCH
and maxEAgchPowerDl
E-HICH
These parameters set the
E-DPDCH
maximum downlink power for an
E-DPDCH
individual E-HICH/E-RGCH and
E-DPCCH AGCH in the cell. In order to
improve the reception of E-
DPCCH
AGCH, EHICH/E-RGCH in areas
HS-PDSCH with poor coverage, the power
HS-PDSCH
HS-DSCH levels at which they are
HS-PDSCH
: transmitted should be increased.
HS-SCCH

HS-DPCCH

DPDCH/DPCCH
Downlink Power for HSUPA
E-AGCH

E-RGCH
Increasing
E-HICH
maxUserEHichERgchPowerDl
has a greater impact on
E-DPDCH
interference in DL, and also on
E-DPDCH
power available for HSDPA,
E-DPCCH
reducing the HSDPA
performances. On the other
DPCCH
hand a value too low might lead
HS-PDSCH to a high percentage of UL
HS-DSCH HS-PDSCH retransmissions, which will
HS-PDSCH
: affect not only UL, but also DL
HS-SCCH performances.
HS-DPCCH

DPDCH/DPCCH
Using individualOffset
Ec/No

P_CPICH best cell


P_CPICH best cell plus
individualOffset
reportingRange1b
reportingRange1a
Hysteresis1a/2

Hysteresis1a/2

Hysteresis1b/2

Hysteresis1b/2

P_CPICH 2

time
In case different types of MCPA are deployed in the network the cell
individualOffset parameter can be used to restore the previous SHO
region by adding to the higher MCPA cell an offset equal to the difference
of the maximumTransmissionPower.
Signalling Radio Bearer on
HSDPA
SRB on HSDPA + F-DPCH

HS-PDSCH HS-DSCH

Downlink Signalling Radio Bearer (DL SRB)

Uplink Signalling Radio Bearer (UL SRB)

E-DPDCH or DPDCH E-DCH or DCH


Faster DL Signalling
=> Decreased time for RAB establishment and other transitions
Parameter Settings for SRB
on HSDPA
ManagedElement
+-NodeBFunction
+-Sector *recommended
+-Carrier
extraPowerForSrbOnHsdpa = 20*
(0..100, unit: 0.1 dB, Default = 0)

SRB

HS-SCCH

ManagedElement
+-NodeBFunction
+-Sector *recommended
+-Carrier
extraHsScchPowerForSrbOnHsdpa = 20*
(0.. 100, unit: 0.1 dB, Default = 0)
HSPA Optimization

Channel Switching
Channel Elements Required for
Some Commonly Used RABS
UL RAXB DL TXB
RAB
SF CE SF CE
AMR Speech (4.75 & 5.9 kbps) 128 1 256 1
AMR Speech (7.9 & 12.2 & 12.2+0 PS kbps) 64 1 128 1
PS Streaming (PS 16/64 + PS 8 kbps) 32 2 32 2
CS Streaming (57.6 kbps) 16 4 32 2
CS Data (64 kbps) 16 4 32 2
CS Data + Interactive (CS 64 + PS 8 kbps) 16 4 32 2
AMR Speech + Interactive (12.2 + PS 64 kbps) 16 4 32 2
PS Interactive (64 kbps) 16 4 32 2
PS Interactive (128 kbps) 8 8 16 4
PS Streaming (PS 16/128 + PS 8 kbps) 32 2 16 4
PS Interactive (384 kbps) 4 16 8 8
Channel Switching
Parameters
› hsdschInactivityTimer downswitchTimerThreshold
› downswitchThreshold inactivityTimer
› featureStateHsdpaMcInactCtrl hsdpaMcInactivityTimer
› licenceStateHsdpaMcInactCtrl hsdpaMcActivityBufferThreshold
› t323 fastDormancyMethod
› hsFachEnabled hsFachSupport
› hsFachQosLevel hsFachQueueSpi
› dlRlcBufUpswitchHsFach featureStateHsFach
› sib5RepPeriod sib11StartPos
› ulRlcBufUpswitch dlRlcBufUpswitch
HSDPA Channel Switching
EUL/HS

DCH/HS

Trade off between resource


FACH
usage and user access time

URA_PCH

Idle Settings depend on network maturity


HSDPA Channel Switching
hsdschInactivityTimer
Dedicated Channel

downswitchTimerThreshold

downswitchThreshold

Decrease in upswitches can


hsdschInactivityTimer
be seen using
Range: 1….. 255 Unit: 1 s Default 2 s
pmUpswitchFachHsAttempt
pmUpswitchFachHsSuccess

Common Channel
Increasing hsdschInactivityTimer to about 10 seconds increases 64/HS traffic at
RNC by 20%-50% Downswitch decrease can be monitored by pmPSIntHsToFachSucc.
HSDPA Channel Switching
hsdschInactivityTimer
Increasing hsdschInactivityTimer
decreases switches and gives
better user experience

A low value saves HW on uplink


especially for smartphone users
HSDPA Channel Switching
inactivityTimer
Increasing inactivityTimer reduces FACH
channel switches and access time.
Increasing to 60 s reduces
pmChSwitchFachIdle and Idle
pmTotNoRrcConnectReqPs by 50%.
60 s or even 90 s is recommended however an
increase in PS drop call rate may be seen as
FACH is more prone to dropped calls
Multi Carrier
E-DCH

F1
Serving HS-DSCH cell

F1

Secondary Serving HS-DSCH cell

MultiCarrier HSDPA MC-HSDPA will be able to F2


deliver peak downlink throughput of 42Mbps,
i.e.double the throughput achievable without MC.
The capacity gained by using MC instead of
single carrier (SC) improves the efficiency of the
network resource being utilized.
HSDPA Channel Switching
MultiCarrier Inactivity
ManagedElement
+-NodeBFunction
+-RbslocalCell
featureStateHsdpaMcInactCtrl = ACTIVATED
All MC PQs buffers empty for:
(Operator parameter to activate feature on cell)
F1 ManagedElement 1..2550
Unit 0.1sec
Serving cell +-NodeBFunction Default = 2
+-Sector (0.2 sec)
F1 +-Carrier
hsdpaMcInactivityTimer
=> Secondary Serving cell deactivated
Secondary Serving cell

Lower F2
battery
consumption Amount of data in any MC PQ buffer:
ManagedElement 0..200
ManagedElement +-NodeBFunction Unit 250 Bytes
+-NodeBFunction +-Sector Default = 2
licenseStateHsdpaMcInactCtrl = ENABLED +-Carrier
(500 Bytes)
(Read only showing the license is activated) hsdpaMcActivityBufferThreshold
=> Secondary Serving cell Activated
HSDPA Channel Switching
MultiCarrier Inactivity

Reducing the value of


hsdpaMcInactivityTimer
conserves UE battery however it
increases the number of
activations/deactivations and
increases the time taken for UE to
reach maximum throughput.
Fast Dormancy Handling
CELL_DCH
› Down switch to URA HS
– When receiving signaling
connection release NW
Controlled
indication Fast CELL_FACH
– With 3GPP standardized Dormancy Fast
Dormancy
cause value Today
– Reduce radio and core URA_PCH

network load
IDLE

Improve signaling load and battery life for chatty applications


Fast Dormancy Parameters
Rrc::t323
is a UE timer that is used by the network to
indicate that the Fast Dormancy Handling feature
is supported and used by the UE to limit the
frequency of UE sending the Signalling
Connection Release Indication message.
RncFunction::fastDormancyMethod
indicates if there will be an attempted direct
downswitch to URA_PCH after reception of
RRC:Signalling Connection Release Indication
message in the case that a cause IE is missing.
Fast Dormancy
Phone starts t323 timer T323 expires

UE can not send SCRI


Wake up Sleep mode

Use phone 15 seconds


Use phone 15 seconds
UE sends RRC SCRI Phone will not send SCRI
DCH

FACH
Phone switches to
URA_PCH CELL_FACH after 1
second
hsdschInactivityTimer Phone switches to URA_PCH by Phone switches
IDLE =1 fast dormancy to URA_PCH by
inactivityTimer
High Speed Downlink for
FACH
UEs in CELL_FACH

‘Classic’ FACH (32 kbps)

HS-DSCH (>10X32 kbps)

UE supporting 3GPP Rel 7 Enhanced CELL_FACH

More efficient Cell_FACH state


‘snappier’ response for applications specifically
in conjunction with the URA PCH feature.
High Speed FACH Parameters
RncFunction::hsFachEnabled
enables or disables the HS-FACH feature on RNC level. The parameter
steers the inclusion of HS-FACH related parameters in those RRC messages
that forces the UE to enter the CELL_FACH state.
Hsdsch::hsFachSupport
indicating whether or not the cell supports HS-FACH. There is always a one-
to-one mapping between this parameter and the existence of SIB5 IE” HS-
DSCH common system information” in System Broadcast (read only).
SpiQosClass::hsFachQosLevel
sets the QoS-level used by HS-FACH for this CELL_DCH SPI value.
RnlQosClassProfile::hsFachQueueSpi
defines the SPI-values used for the common queues used for DTCH.
ChannelSwitching::dlRlcBufUpswitchHsFach
downlink RLC buffer threshold for triggering an upswitch from CELL_FACH
to CELL_DCH.
HS FACH Queue and SPIs

PQ1 PQ2 PQ3

SPIs for the cell_FACH


Queues 2(3) 4(NA) NA(NA)
(hsFachQueueSpi)
This table defines the number of HS-FACH queues used for DTCH
and the SPI to use for those queues. In this example,
hsFachQueueSpi would be set to [2,4]. The default settings can be
found within brackets.
Mapping between CELL DCH SPIs
and CELL FACH Qos Level
PS Interactive CELL DCH SPI PS Interactive CELL FACH
Qos Level (hsFachQosLevel)
0 1(1)
1 1(1)
2 1(1)
3 1(1)
4 2(1)
5 2(1)
….. 2(1)
13 2(1)
14 2(1)

CELL DCH SPIs 0 to 3 are mapped to first Qos Level while 4 to 14


are mapped to the higher Qos Level
High Speed Downlink for
FACH Activation
SIB 5
HS-DSCH common system information * upswitch from CELL_FACH to
CELL_DCH is triggered at arrival of
RRC Connection Setup Request data in the RNC buffer

‘HS-PDSCH in CELL_FACH=TRUE’

ManagedElement ManagedElement
+-NodeBFunction +-RncFunction
+-RbsLocalCell hsFachEnabled= TRUE*
featureStateHsFach= ACTIVATED
sib5RepPeriod: 32 -> 64
sib11StartPos: 20 -> 28
HSDPA Channel Switching
Additional Parameters
Range: 0 .. 64
Dedicated Channel Unit: kbps
Default 0
(16 illustrated)

downswitchTimerThreshold

downswitchThreshold

PING! Range: 0 .. 32
downswitchtimer Unit: kbps
Default 0
(8 illustrated)
Range: 0 ….. 255
Unit: 1 s
Default 2

Common Channel
Default values of downswitchTimerThreshold and downswitchThreshold
mean that UEs stay on DCH all the time even with only background traffic.
HSDPA Channel Switching
Additional Parameters
MEASUREMENT REPORT

ulRlcBufUpswitch dlRlcBufUpswitch
OR

Range = 8, 16, 32, 64, 128, Will Range = 0 …20000


256, 512, 1024, 2048, 3072, request a Unit = Bytes
4096, 6144, 8192. switch to Resolution = 100
Unit = Bytes DCH Default = 500
Default = 256
(or HSPA)
Default values seem to keep most background traffic on FACH.
Decreasing these parameters has too much impact on resource
usage and number of switches measured by pmChSwitchFachDch
and pmChSwitchDch64Fach.

Potrebbero piacerti anche