Вы находитесь на странице: 1из 176

WRST 4

Module Three

W
RS
T

Call Performance
WRST 4
Jussi Reunanen
&
Benoist Guillard
&
Supranee Sarinkarnpoonperm
1

NOKIA

Call Performance.PPT / 16-09-2003 /Reunanen Jussi

Company Confidential

D
ss
el
d

or
f

Contents
Call Setup Performance

Idle Mode Performance

Paging

RRC Connection Establishment Performance

RAB Establishment Performance

Call Setup Failure & Time

UL Interference

Dropped Call Performance

SHO Performance (round-the-corner)

DL Power Allocation Parameters

UL Power Allocation Parameters

PS Call Performance
2

NOKIA

Call Performance.PPT / 16-09-2003 /Reunanen Jussi

Company Confidential

Call Setup Performance


-Idle Mode Cell Reselection-

Several UEs cannot perform cell reselection during RRC


Connection Setup phase (Kenny can sw >12.8x)
CPICH Ec/No
CPICH RSCP

-100dB

UE does not
start the cell
reselection
procedure if
default
parameter
settings are
used

UE Sends
First RRC
Connection
Request
Message
RF Conditions
OK and CPICH
Ec/No is
10dB

RNC responds with RRC


Connection Setup message
which is not heard by the UE
due to CPICH Ec/No falls
below 15dB

CPICH Ec/No
CPICH RSCP

-10dB
-15dB

Time

Cell A

Cell B

The UE sends all the rest of the RRC Connection Request


messages to the old cell, Cell A without changing the serving
cell to Cell B (Kenny can perform this reselection and send next
RRC Connection Request message to Cell B) and the call setup
fails

NOKIA

UE moving from Cell A towards Cell B

Call Performance.PPT / 16-09-2003 /Reunanen Jussi

Company Confidential

Call Setup Performance


-Idle Mode Cell Reselection-

Default parameter settings


Cell reselection triggering time (Treselection = 0s)
Reselection takes place immediately when the UE notices that
there is difference between the cells Ec/No values (in worst case
scenario there can be up to 3dB + Qhyst difference based on the
measurement accuracy requirement)
Cell reselection hysteresis 2 (Qhyst2 = 4dB)
This will add 4dB hysteresis to the neighboring cell evaluation
(target for the cell reselection)
Note that Qhyst1 is used only in case the cell selection and reselection quality measure is set to CPICH RSCP (default is CPICH
Ec/No so Qhyst 1 is not used in intra-FDD reselection)
Cell Re-selection Quality Offset 2 (AdjsQoffset2 = 0dB)
This parameter is used in the cell re-selection and ranking between
WCDMA cells. The value of this parameter is subtracted from the
measured CPICH Ec/No of the neighbor cell before the UE
compares the quality measure with the cell re-selection/ ranking
criteria
S intrasearch (Sintrasearch = 4dB)
This parameter is used by the UE to calculate the threshold (CPICH
Ec/No) to start intra frequency (SHO) measurements (Sintrasearch
above QqualMin value)
Minimum required quality level in the cell (QqualMin = -20dB)
4

NOKIA

Call Performance.PPT / 16-09-2003 /Reunanen Jussi

Company Confidential

Call Setup Performance


-Idle Mode Cell Reselection-

Default parameter settings


CPICH Ec/No
CPICH RSCP

UE starts the
neighboring cell
measurements here
and it takes some time
to perform the
reselection

UE hangs in CELL A
until
QqualMin+Sintrasearch =
-16dB threshold is
reached.
However at this level the
RRC Connection Setup is
probably not heard by the
UE any more or the BTS
does not hear the RRC
Connection Request

CPICH Ec/No
CPICH RSCP

Cell Reselection

Qhyst2= -16dB + 4dB =


-12dB
QqualMin+Sintrasearch
=-20dB + 4dB = -16dB
QqualMin = -20dB

Time

Cell A

Cell B

The time when the source cell CPICH Ec/No is too low as well as
the gap between new and old cell is too large i.e. the UEs that
cannot make (all except Kenny) make cell reselection during RRC
Connection Setup phase will have failed call attempts
5

NOKIA

Call Performance.PPT / 16-09-2003 /Reunanen Jussi

Company Confidential

Call Setup Performance


-Idle Mode Cell Reselection-

With default parameter settings:


There is big chance that the UEs with limited capability to
perform proper cell reselection will have failed call attempts
seen as consecutive RRC Connection Request messages sent
by the UE (seen from the UE log) but no response from RNC or
RRC Connection Setup received but then no actions after RRC
Connection Setup Complete message (i.e. RRC Connection
Setup Complete message is not heard by the BTS)
CPICH RSCP

CPICH Ec/No
Recommended area

Calls setups very successful


Calls setups ok
Calls setups difficult
Calls setups failures
QqualMin = -20dB

QqualMin+Sintrasearch
+
Qhyst2= -16dB + 4dB
=
-12dB
QqualMin+Sintrasearch
=-20dB + 4dB = -16dB

Time
6

NOKIA

Call Performance.PPT / 16-09-2003 /Reunanen Jussi

Company Confidential

Call Setup Performance


-Idle Mode Cell Reselection-

In order to optimise the call setup performance following changes


are suggested (to be in the recommended area all the time)
S intrasearch (Sintrasearch = 12dB)
The UE starts measuring the neighboring intra frequency
cells when the serving cell CPICH Ec/No reaches 20dB +
12 dB = -8dB
This setting should leave enough time for the UE to perform
reselection as well as enough room (~7dB depending on
the UE sensitivity) for the RRC Connection Setup to be
successful

NOKIA

Note: remember to tick on the tick box in front of the


Sintrasearch parameter (in PlanEditor: SIB3compmask2)
otherwise the system does not send this parameter and
according to spec the UE measures intra frequency cells all
the time -> battery consumption is high

Call Performance.PPT / 16-09-2003 /Reunanen Jussi

Company Confidential

Call Setup Performance


-Idle Mode Cell Reselection-

New set Case 2: Sintrasearch = 12 dB, Treselection = 0 sec and


Qhyst2 =4 dB

CPICH Ec/No
CPICH RSCP

More time to select


and monitor a better
cell

Cell reselection
happens here where
cell B is about 8 dB
and Cell A is about 12
dB

CPICH Ec/No
CPICH RSCP

Qhyst2 = -8dB + 4dB


-4dB

UE starts the
neighboring cell
measurements near
earlier than at cell
edge

Worst case scenario

QqualMin+Sintrasearch
=-20dB + 12dB = -8dB

QqualMin = -20dB

Cell A

Time

Cell B

2 second hysteresis give UE especially at LA, RA border to be more stable than ping
pong cell reselection. This also reduces signaling and call failure at LA/RA border
due to LA/RA update
8

Qqualmin shall not be changed as it will greatly affect cell coverage especailly
NOKIA
Performance.PPT / 16-09-2003 /Reunanen Jussi Company Confidential
when Call
traffic
is higher because CPICH Ec/No will lower and creating a coverage hole

Call Setup Performance


-Idle Mode Cell Reselection-

With optimised parameter settings:


CPICH RSCP

CPICH Ec/No
Recommended area

Calls setups very successful


Calls setups ok

Qhyst2 =
4dB

QqualMin+Sintrasearch
=-20dB + 12dB = -8dB

Calls setups difficult


Calls setups failures

QqualMin =
-20dB

Time

It should be noted that the so called optimized settings for cell


reselection are causing the UE to start the intra freq. neighbor
measurements earlier than the defaults and therefore UE battery
consumption is compromised
It is therefore recommended to find optimized value specific for
each environment separately and not to take the values
presented here as some defaults
9

NOKIA

Call Performance.PPT / 16-09-2003 /Reunanen Jussi

Company Confidential

Call Setup Performance

BLER, Call drop rate, MOC CSSR, MTC CSSR

-Idle Mode Cell Reselection : Effect-

10

Idle mode
param. set as
defaults

Sintrasearch = 10dB
Treselection = 0s

AMR performance

New CD

Time in Days

NOKIA

BLER rate

Call drop rate

MOC setup failure rate

MTC setup failure rate

Call Performance.PPT / 16-09-2003 /Reunanen Jussi

There is significant
improvement in performance
after due to changing of the
idle mode cell reselection
parameters
Call Setup Success rate
(MTC and MOC especially
MTC)
MTC improvement is due
to the fact that in case
there are problems in RRC
Connection Setup the
paging timer in MSC could
be expired and therefore
causing failed call
attempts

Company Confidential

Call Setup Performance


-Idle Mode Cell Reselection : Effect-

-2

All Call Attempts Test 1

CPICH Ec/No [dB]

-4
-6
-8
-10

From the charts on the left it can be seen


that from all the RRC Connection Request
messages only a few are done in the worse
than CPICH Ec/No = -15dB situation
CPICH Ec/No = -15dB can be
considered as some sort of UE
performance limit

-12
-14

In this case the Sintraseacrh setting seems


to be ok (10dB)

-16
-18

11

All Call Attempts Test 2

-6
-8
-10
-12
-14
-16
-18
-20
-22
-24

NOKIA

-2
-4
CPICH Ec/No [dB]

CPICH Ec/No [dB]

-2
-4

Call Performance.PPT / 16-09-2003 /Reunanen Jussi

-6
-8
-10
-12
-14
-16
-18
-20
-22
-24

Company Confidential

Call Setup Failures

Call Setup Performance


-Idle Mode Cell Reselection : Effect-

Nokia tested Common CH power sets and cell reselection


parameter set in site-basis to choose candidate parameter set,
which is applied to one area.
The parameter set for Cell reselection test:
Parameter
Sintrasearch
Qhyst2
Treselect

Default
12dB
4dB
0sec

Case4
12dB
4dB
2sec

Case5
12dB
0dB
0sec

Tests were done with Sanyo UE (3.23)

12

NOKIA

Call Performance.PPT / 16-09-2003 /Reunanen Jussi

Company Confidential

Call Setup Performance


-Idle Mode Cell Reselection : Effect-

Idle mode (Stationary test)


Parameter set
Default
Case4
Case5
Idle mode (Drive test)
Parameter set
Default
Case4
Case5

Low Coverage
CPICH RSCP
-81.2
-78.83
-77.47

CPICH Ec/No
-9.04
-7.51
-7.39

CPICH RSCP
-75.41
-76.11
-75.98

Med. Coverage
CPICH RSCP
-75.06
-76.42
-74.21

CPICH Ec/No
-5.16
-6.5
-4.03

CPICH Ec/No
-5.45
-5.89
-6.03

AMR MOC (Stationary test)


Parameter set
Default
Case4
Case5
AMR MOC (Drive
rivetest)
Parameter set
Default
Case4
Case5

13

NOKIA

Low Coverage
#of call attempts
100
100
100

Setup success
82
83
92

#of call attempts


78
100
100

Setup success
76 (97%)
92%
94%

Call Performance.PPT / 16-09-2003 /Reunanen Jussi

CPICH RSCP
-91.63
-92.18
-91.45

Company Confidential

CPICH Ec/No
-6.74
-6.18
-6.16

Med. Coverage
#of call attempts Setup success
100
93
100
97
100
89

Call Setup Performance


-Idle Mode Cell Reselection : Effect-

Call Setup Success Statistics (drive test in one area)


Call setup success rate has improved after parameter change
by 5~7%.
It seems that UE could find better serving cell after parameter
change
Call Setup Success Rate

AMR MTC
Parameter set # of call attempts Setup success Setup success rate
Default
484
412
85.1%
CR Case5
333
286
85.9%
UDI MOC
Parameter set # of call attempts Setup success Setup success rate
Default
320
262
81.9%
CR Case5
647
557
86.1%

Success rate

AMR MOC
Parameter set # of call attempts Setup success Setup success rate
Default
572
473
82.7%
CR Case5
854
765
89.6%

100.0%
95.0%
90.0%
85.0%
80.0%
75.0%
70.0%
65.0%
60.0%
55.0%
50.0%
45.0%
40.0%
35.0%
30.0%
25.0%
20.0%
15.0%
10.0%
5.0%
0.0%

Default
Case5

Services

14

NOKIA

Call Performance.PPT / 16-09-2003 /Reunanen Jussi

Company Confidential

Call Setup Performance

-Idle Mode Cell Reselection : Questions Why it is not very good idea to tune Qqualmin parameter too
much and what would be appropriate value?
What is the difference between Qhyst2 and Qhyst1?
What is the accuracy requirement of UE CPICH Ec/No and CPICH
RSCP measurements?

15

NOKIA

Call Performance.PPT / 16-09-2003 /Reunanen Jussi

Company Confidential

Contents
Call Setup Performance

Idle Mode Performance

Paging

RRC Connection Establishment Performance

RAB Establishment Performance

Call Setup Failure & Time

UL Interference

Dropped Call Performance

SHO Performance (round-the-corner)

DL Power Allocation Parameters

UL Power Allocation Parameters

PS Call Performance
16

NOKIA

Call Performance.PPT / 16-09-2003 /Reunanen Jussi

Company Confidential

Call Setup Performance


-Paging : PIs-

Some UEs do not support all DRX Cycle Length Coefficients,


resulting in lost paging
For instance, NEC VTX does not support DRX Cycle Length
<1280ms
If DRX Cycle Length=640ms, every second paging message is
lost
The MTC can still work thanks to CN Paging repetition (depends
on MSC/operator settings)
This phenomenon can be observed with MSC counters and KPIs
Nokia MSC:
Clear Code 0012 (No Paging Response)
! Clear Code 0012 is in the Normal Clearing Clear Code
Group ! (other groups are Internal, External and
Subscriber)
Paging Channel (PCH) is transmitted in the SCCPCH together with
FACH and power tuning of SCCPCH can help the performance of
paging (see chapter -RRC Connection Establishment: Power Offset
for SCCPCH Pilot & TFCI bits-)
17

NOKIA

Call Performance.PPT / 16-09-2003 /Reunanen Jussi

Company Confidential

Call Setup Performance


-Paging : PIs-

Paging for UEs in IDLE state


UE

BS

RNC

UE has no RRC connection

CN
RANAP: PAGING

PICH
(FP/AAL2/PCCH/PCH/S-CCPCH) : PAGING TYPE 1
RRC connection establishment
Paging response

CN (RANAP: PAGING Message received


from CN = UE in idle mode) originated
paging amount: When RNC sends
Paging Type 1
M1006C25 PAGING TYPE 1 ATT CN
ORIG
A number of RNC originated paging
type 1 attempts (UE in PCH/URA
substate): When RNC sends Paging
Type 1
M1006C26 PAGING TYPE 1 ATT RNC
ORIG

Paging for UEs in CELL_PCH or URA_PCH


states
MM
2
MM Idle

UE

BS

RNC

Connected
CN 1

CN 2

UE has signalling connection to CN1


UE is in URA_PCH or CELL_PCH state

RANAP:PAGING

PICH

UE

RNC

The
The MAC-d
MAC-d (RLC)
(RLC) in
in
RNC
RNC indicates
indicates that
that
1.PDP PDU
there
there is
is downlink
downlink
user
data
in
RLC
user data in RLC
buffers
buffers then
then the
the RRC
RRC
signaling
signaling entity
initiates
initiates the paging
2. Paging Request (P-TMSI)
procedure
procedure
RRC: PAGING TYPE 1 with PICH BTS -> UE
MM Cell
Update

(FP/AAL2/PCCH/PCH/S-CCPCH) : PAGING TYPE 1

Paging response to CN 2
A number of paging messages
received from the CN
M1003C36 REC PAG MSG
18

NOKIA

Call Performance.PPT / 16-09-2003 /Reunanen Jussi

Company Confidential

SGSN

Call Setup Performance


-Paging : PIs-

Paging for UEs in CELL_FACH or CELL_DCH


MM Connected
MM Idle
states
UE

BS

RNC

CN 1

CN 2

A number of paging messages


received from the CN
M1003C36 REC PAG MSG

UE has signalling connection to CN1


UE is in CELL_FACH or CELL_DCH state

RANAP:PAGING

RRC:PAGING TYPE 2

Paging response to CN 2

19

NOKIA

Call Performance.PPT / 16-09-2003 /Reunanen Jussi

Company Confidential

CN (RANAP: PAGING Message received


from CN = UE in CELL_FACH or
CELL_DCH states) originated paging
amount: When RNC sends Paging Type
2
M1006C27 PAGING TYPE 2 ATT

Call Setup Performance


-Paging : PIs-

Paging Success Rate

1001C32 MTC Conv Att 1001C34 MTC Stream Att 1001C36 MTC Inter Att 1001C38 MTC Backg Att 1006C37 Cell Update Att Due to Paing Resp
1003C36 Received Paging Messages

The counters M1001C32, M1001C34, M1001C36 and M1001C38


are updated according to the IE: Establishment cause in RRC
Connection Request message therefore this PI might not be very
accurate
This formulas denominator M1003C36 REC PAG MSG does NOT
include the case where the paging originates from RNC (case 3)
-> therefore the formula might give too optimistic results as the
M1006C37 CELL UPDATE ATT DUE TO PAGING RESP counter
includes also case 3

22

NOKIA

Call Performance.PPT / 16-09-2003 /Reunanen Jussi

Company Confidential

Call Setup Performance


-Paging : Questions-

What is the paging procedure for the UE ?


Is in CELL_PCH state and has made GPRS Attach and PDP
Context is active
There are some PDUs are sent to that UE from the SGSN
What is the response to the paging message?
What is DRX and how does it relate to paging?

23

NOKIA

Call Performance.PPT / 16-09-2003 /Reunanen Jussi

Company Confidential

Contents
Call Setup Performance

Idle Mode Performance

Paging

RRC Connection Establishment Performance

RAB Establishment Performance

Call Setup Failure & Time

UL Interference

Dropped Call Performance

SHO Performance (round-the-corner)

DL Power Allocation Parameters

UL Power Allocation Parameters

PS Call Performance
24

NOKIA

Call Performance.PPT / 16-09-2003 /Reunanen Jussi

Company Confidential

Call Setup Performance


-RRC Connection Establishment-

UE

RNC

Node B
[RACH] RRC:RRC Connection Request
NBAP: RL Setup Request
NBAP: RL Setup Response
ALCAP:ERQ
ALCAP:ECF
[FACH] RRC: RRC Connection Setup
L1 Synchronisation

NBAP: Synchronisation Indication


[DCH] RRC: RRC Connection Setup Complete

The power level for sending of the RRC Connection


Request message is set/calculated by the UE
Uplink open loop PC take place at the random
access procedure:
UE measures the received power of the CPICH
which reported back in "measured results on
RACH" IE in: initial direct transfer (UTRAN must
extract this IE from the NAS message), cell
update / ura update, measurement report (UL
traffic volume measurement report), RRC
connection request, uplink direct transfer
(UTRAN must extract this IE from the NAS
message)

Transmission power of CPICH (SIB5&6), total received uplink power at BS


(SIB7) and a parameter value PRACHRequiredReceivedCI are sent from BS in
system information broadcast messages (SIB5&6)
Based on this information UE calculates the transmission power of the
first pre-amble of PRACH ptx

ptx = CPICHtransmissionPower-RSCP(CPICH)+RSSI(BS)+"constant
value
25

NOKIA

Call Performance.PPT
/ 16-09-2003 value"
/Reunanen Jussi
Company
where
the "constant
is the
valueConfidential
of the RNC parameter PRACHRequiredReceivedCI

Call Setup Performance


-RRC Connection Establishment-

UE

RNC

Node B
[RACH] RRC:RRC Connection Request
NBAP: RL Setup Request
NBAP: RL Setup Response
ALCAP:ERQ
ALCAP:ECF
[FACH] RRC: RRC Connection Setup

If no acquisition indicator is received UE


raises the transmission power by
PowerRampStepPRACHpreamble which is a
broadcast parameter
This procedure is repeated until
acquisition indicator is received, after
which UE sends the RACH message at
power which offset by
PowerOffsetLastPreamblePRACHmessage
from the last preamble
L1 ACK / AICH

PRACH_preamble_retrans parameter
determines how many times PRACH
preamble can be sent (without AICH
response) within one preamble
ramping cycle (SIB5&6)

Not detected
Downlink
BS

RACH_tx_Max defines how many


P2
P1
Uplink
times the PRACH pre-amble ramping
MS PreamblePreamble
cycle procedure can be repeated
1
2
before UE MAC reports a failure on
RACH transmission to higher layers
26
NOKIA
Call Performance.PPT / 16-09-2003 /Reunanen Jussi Company Confidential
(SIB
5&6)

Message part

Call Setup Performance


-RRC Connection Establishment-

Some of the UEs (other than Kenny) are not capable of performing
Open Loop PC correctly i.e. they have problems in RRC Connection
Establishment
The UE cannot detect the SIB7 UL interference level correctly
This can be seen from the UE logs as consecutive failed RRC
Connection Requests => This could be seen in calls without any
RRC connection request by analyzing sent PRACH preambles
The situation can be corrected by increasing the parameter value
PRACH_preamble_retrans or RACH_tx_Max
Increasing the parameter values should be made with extra care
as increasing the maximum UE TX power on RACH increases the
UL load is as well
In case of problematic site which does not answer to PRACH at
all due to BTS problems; increasing the power on RACH
increases the UL interference on the surrounding sites

27

NOKIA

Call Performance.PPT / 16-09-2003 /Reunanen Jussi

Company Confidential

Call Setup Performance

-RRC Connection Establishment Logging CPICH Ec/NoCPICH Ec/No as in RRC Connection Setup Request

-20 -18 -16 -14 -12 -10

-8

-6

-4

-2

Results missing

-2
-4
-6
-8
-10
-12
-14
-16
-18

CPICH Ec/No as reported by fingers

For Filtering purposes it is


good to have Ec/No
recorded when the RRC
Connection Setup is sent
Whether the reading is
done from the Finger data
or from the RRC
Connection Setup
message needs to
decided

As it can be seen from the picture, the difference between RRC


Connection Setup message reported CPICH Ec/No and finger data CPICH
Ec/No increases when the level of CPICH Ec/No is decreasing
Recommendation is to use the finger data as it is updated basically on at
least every 200ms intervals (specified by the 3GPP spec 25.302)
28

NOKIA

Call Performance.PPT / 16-09-2003 /Reunanen Jussi

Company Confidential

Call Setup Performance

-RRC Connection Establishment CPICH Ec/No / RSCP vs. UE Tx


PowerUE Tx Power [dBm]
-40 -35 -30 -25 -20 -15 -10 -5

10 15 20 25
-40
-60
-70
-80
-90

RSCP [dBm]

-50

-100
-110
-120
30

10
0
-30

-25

-20

-15

-10

-5

-10
-20
-30

UE Tx Power [dBm]

20

For failed calls the UE Tx vs. CPICH


Ec/No and CPICH RSCP should be
investigated in order to see possible
DL coverage or open loop PC problems
In case CPICH RSCP is low but UE Tx
Power is low
There is possibility that UE open
loop PC is not working correctly
Or the CPICH Tx power is pretty
low and there is DL coverage
problem
Or the SCCPCH power offset to
CPICH power is set wrongly too
big offset
In case CPICH Ec/No is low but the
UE Tx power is low then there could
be some problem in DL quality provided that the CPICH RSCP is ok
Some high DL loading in the
cell
Pilot pollution on that area

-40
29

NOKIA

CPICH Ec/No [dBm]


[dB]

Call Performance.PPT / 16-09-2003 /Reunanen Jussi

Company Confidential

Call Setup Performance

-RRC Connection Establishment CPICH Ec/No / RSCP vs. UE Tx


Power-30

CPICH Ec/No [dB]


-20
-15
-10

-25

-5

From CPICH Ec/No vs. CPICH RSCP it


can be seen that only a few call failures
are having relatively good CPICH RSCP
but still pretty poor CPICH Ec/No:
indicating possibility of Pilot Pollution
Indicating possibility of relatively
high DL load
Those cells in case they constantly do
have this kind of phenomena should be
investigated more

0
-50
-60
-70
-80
-90
-100

CPICH RSCP [dBm]

-40

-110
-120
Plot of RSCP and Ec/No vs Data No. in problem Area
Pilot pollution area/poor
dominance area

-20

For possible pilot pollution investigation the


amount of SCs and their strength can be
-5
plotted from the drive test data and possible
-10
the cells causing the pilot pollution can be
-15
identified and dominance optimisation can be
-20
-25
done
0

Ec/No (dB)

RSCP and RSSI (dBm)

-40
-60
-80

-30
-35

-100

-40
-120

-45

-140

ANT1_RSSI
SC174 RSCP
SC69 Ec/No
30

-50

Data No.

NOKIA

SC97 RSCP
SC180 RSCP
SC174 Ec/No

SC106 RSCP
SC97 Ec/No
SC180 Ec/No

SC103 RSCP
SC106 Ec/No

SC69 RSCP
SC103 Ec/No

Call Performance.PPT / 16-09-2003 /Reunanen Jussi

Company Confidential

Call Setup Performance

-RRC Connection Establishment CPICH Ec/No / RSCP vs. UE Tx


Power By investigating the Call Setup
Success Probability distribution
among all the call setup related
messages it can be seen:
RRC Connection Setup Complete
plays a significant role as a
starting point of what can be
achieved in terms of CSSR
Therefore it is utmost important to
understand how the RRC
Connection Establishment
procedure works

This kind of Call Setup Performance analysis can also help for the
whole call set-up troubleshooting and possible trouble spot finding and
tuning

31

NOKIA

Call Performance.PPT / 16-09-2003 /Reunanen Jussi

Company Confidential

Call Setup Performance

Before 1.5.2 CD15

-RRC Connection Establishment-

UE

RNC

Node B
[RACH] RRC:RRC Connection Request

NBAP: RL Setup Request


NBAP: RL Setup Response
ALCAP:ERQ
ALCAP:ECF
[FACH] RRC: RRC Connection Setup

UE

RNC

Node B
[RACH] RRC:RRC Connection Request

There is no
retransmission of RRC
Connection Setup
-message.
This is why the UE just
sends another RRC
Connection Request
message.

UE
UE does
does not
not hear RRC Connection Setup from RNC

NBAP: RL Setup Request


NBAP: RL Setup Response
ALCAP:ERQ
ALCAP:ECF
[FACH] RRC: RRC Connection Setup

UE
UE does
does not
not hear
hear RRC
RRC Connection
Connection Setup
Setup from
from RNC
RNC
[RACH] RRC:RRC Connection Request

32

NOKIA

The more RRC


Connection Request
retransmission the
more (in case RNC
hears the message) RLs
are setup and more Iub
Capacity used.
The RNC internal timer
(T_RRC_Resp_CCH) will
release the unused RLs
on Iub after 6seconds
(from
the Confidential
Call Performance.PPT / 16-09-2003 /Reunanen
Jussi Company

NBAP: RL Setup Request


NBAP: RL Setup Response
ALCAP:ERQ
ALCAP:ECF
[FACH] RRC: RRC Connection Setup

Before 1.5.2 CD15

Call Setup Performance


-RRC Connection Establishment-

UE

RNC

Node B
[RACH] RRC:RRC Connection Request

NBAP: RL Setup Request


NBAP: RL Setup Response
ALCAP:ERQ

[FACH] RRC: RRC Connection Setup

T300

ALCAP:ECF

RRC
RRC Connection
Connection Setup
Setup message
message
is
is transmitted
transmitted in
in several
several
successive
successive transport
transport blocks

[RACH] RRC:RRC Connection Request

Current default values of the parameters are:


T300=3000ms and N300=3
UE will send the RRC Connection Request
message four times in the interval of three
seconds if there is no RRC Connection
Setup message received from the RNC
The RRC Connection Setup message is sent
only once (UM RLC SAP is used for this RRC
procedure)
RRC Connection Setup is quite a long
message that must be sent in several
successive transmission blocks (TB) via
the FACH transport channel
Only one erroneous TB is sufficient to
cause the MS to reject the whole
Setup message

Additionally there is a possibility for loss of the RRC Connection


Request message since this message is using TM RLC SAP
But since this message can be fit into only one RACH message, the
probability of loss can be considered much smaller than in DL
direction

33

NOKIA

Call Performance.PPT / 16-09-2003 /Reunanen Jussi

Company Confidential

Before 1.5.2 CD15

Call Setup Performance


-RRC Connection Establishment-

Repetition procedures defined in RAN1.5 for RRC messages using the UM-RLC SAP
Repetition counters presented in the following table are defined for unacknowledged
mode RRC messages in RAN1.5.

34

NOKIA

Repetition Counter

Default

Usage

L3Rep_UnAckRLC

2 times

Repeat counter for RRC messages when using UM RLC. Used


for the following DL RRC messages:
1. DCH to FACH state transition command (interval 400ms)
2. Cell Update Confirm (interval 300ms)
3. URA Update Confirm (interval 300ms)

L3Rep_RRC_Conn_Rel

2 times
(interval
300ms)

Repeat counter for RRC Connection Release message sent on:


DCCH mapped to DCH and DCCH/CCCH mapped to FACH.

Call Performance.PPT / 16-09-2003 /Reunanen Jussi

Note. Repetitions shall not be used when the message is sent


over Iur (possible repetitions must be executed by the DRNC).

Company Confidential

Before 1.5.2 CD15

Call Setup Performance


-RRC Connection Establishment-

UE

RNC

Node B
[RACH] RRC:RRC Connection Request

NBAP: RL Setup Request


NBAP: RL Setup Response

Based on the experiments made in the


network with a sufficiently low load one
can get improvement in call setup delay
when reducing the value of T300

ALCAP:ERQ

By reducing the value of T300


the UE will repeat the request sooner
[FACH] RRC: RRC Connection Setup
and the delay in RRC connection
RRC
RRC Connection
Connection Setup message
establishment
caused
by
lost
is
is transmitted
transmitted in
in several
several
successive
successive transport
transport blocks
blocks
Request/Setup messages can be
minimized
UE
UE does
does not
not hear RRC Connection Setup from RNC
However This is wrong way to
correct
the "problem
When load of the network
increases,
too short repetition
interval of RRC Connection Request message will
contaminate the normal operation of the pre-emption
procedure in RAN and generate even more load to already
congested cells
Each RRC Connection Request message might trigger a
packet call release = pre-emption in order to release
capacity for in AAL2
or in BTS or in radio interface
Company Confidential
ALCAP:ECF

35

NOKIA

Call Performance.PPT / 16-09-2003 /Reunanen Jussi

Call Setup Performance

Before 1.5.2 CD15

-RRC Connection Establishment-

UE

RNC

Node B
[RACH] RRC:RRC Connection Request
NBAP: RL Setup Request
NBAP: RL Setup Response
ALCAP:ERQ
ALCAP:ECF
[FACH] RRC: RRC Connection Setup

UE
UE does
does not
not hear
hear RRC
RRC Connection
Connection Setup
Setup from
from RNC
RNC
[RACH] RRC:RRC Connection Request
NBAP: RL Setup Request
NBAP: RL Setup Response
ALCAP:ERQ
ALCAP:ECF
[FACH] RRC: RRC Connection Setup

If the repeated RRC Connection Request


messages are received before the RNC
has been able to complete the ongoing
RRC Connection Setup procedures (i.e.
release the established RLs for the
unsuccessful RRC Connection Setup
attempts) this will cause new RRC
connection establishment procedures to
be triggered (requiring maybe more calls
to be pre-empted in vain)
Initiation of the pre-emption procedure
may need Radio Link Setup Request to
be rejected by the BTS or AAL2-setup
failure to happen
Then the DCH-allocation of packet data
call (to be pre-empted) must be released
(state transition CELL_DCH -> CELL FACH
is triggered)

36

NOKIA

Call Performance.PPT / 16-09-2003 /Reunanen Jussi

Company Confidential

Call Setup Performance

Before 1.5.2 CD15

-RRC Connection Establishment-

UE

RNC

Node B
[RACH] RRC:RRC Connection Request
NBAP: RL Setup Request

State transition from Cell_DCH to Cell_FACH


is in the optimum case fast

NBAP: RL Setup Response


ALCAP:ERQ
ALCAP:ECF
[FACH] RRC: RRC Connection Setup

UE
UE does
does not
not hear
hear RRC
RRC Connection
Connection Setup
Setup from
from RNC
RNC

In case there is a cell reselection procedure


executed before the MS acknowledges the
state transition command, then the RL(s)
cannot be released until the Cell Update is
received via the RACH transport channel or
Radio Link Failure indication(s) is/are received
from the BTS(s)
There may also be radio link(s) over Iur due to
inter-RNC SHO

[RACH] RRC:RRC Connection Request


NBAP: RL Setup Request
NBAP: RL Setup Response
ALCAP:ERQ
ALCAP:ECF
[FACH] RRC: RRC Connection Setup

Additional RRC Connection Requests will cause


additional resource reservations and due to
that
might
cause
some
pre-emption
procedures to be triggered

More complex situation is more time pre-emption procedure takes -> more
new RRC Connection Requests are sent by the UE -> more pre-emption
37

NOKIA

Call Performance.PPT / 16-09-2003 /Reunanen Jussi

Company Confidential

Call Setup Performance

Before 1.5.2 CD15

-RRC Connection Establishment-

UE

RNC

Node B
[RACH] RRC:RRC Connection Request
NBAP: RL Setup Request
NBAP: RL Setup Response
ALCAP:ERQ
ALCAP:ECF
[FACH] RRC: RRC Connection Setup

UE
UE does
does not
not hear
hear RRC
RRC Connection
Connection Setup
Setup from
from RNC
RNC
[RACH] RRC:RRC Connection Request
NBAP: RL Setup Request
NBAP: RL Setup Response
ALCAP:ERQ
ALCAP:ECF

Resources reserved for these unnecessary


"double reservation" cannot be released
immediately; radio links must be kept
allocated for a certain time
If there is no RRC Connection Setup
Complete message received from the UE,
the RL is released after Layer3 time
supervision for the MS acknowledge has
expired

[FACH] RRC: RRC Connection Setup

38

NOKIA

Call Performance.PPT / 16-09-2003 /Reunanen Jussi

Company Confidential

Call Setup Performance

Before 1.5.2 CD15

-RRC Connection Establishment-

CPICH Ec/No
CPICH RSCP

-100dB

First RRC
Connection
Request sent
by the UE
RNC Answers
with RRC
Connection
Setup UE
does not hear
it

Cell A

UE sends new RRC


Connection Request to Cell B
which is successful and call is
setup successfully

CPICH Ec/No
CPICH RSCP

-10dB
-15dB

UE waits T300 seconds until it sends next RRC


Connection Request. During the T300 time the UE
makes cell reselection to Cell B

Time

Cell B

Additionally mobility i.e. cell reselections of the UE during the RRC


connection establishment procedure will cause certain amount of
these "double resource reservations
These mobiles repeat the RRC Connection Request via the new cell if
the cell reselection occurs before the MS has succeeded to receive the
RRC Connection Setup message via the previous serving cell
39

NOKIA

Call Performance.PPT / 16-09-2003 /Reunanen Jussi

Company Confidential

Before 1.5.2 CD15

Call Setup Performance


-RRC Connection Establishment-

Resource allocations (time between RRC Connection Request


and RRC Connection Setup) within "fragment of a second" can
be assumed only for networks operating in low load and in
good radio conditions (no RRC messages lost on RACH/FACH)
RRC connection establishment procedure requiring a preemption in the congested cell may take several seconds to be
completed
Too short repetition interval for RRC Connection Request
message will cause double resource reservations and
unnecessary pre-emption of the ongoing calls (in our current
implementation)

40

NOKIA

Call Performance.PPT / 16-09-2003 /Reunanen Jussi

Company Confidential

Before 1.5.2 CD15

Call Setup Performance


-RRC Connection Establishment-

There is possibility for collecting "success-rate" statistics only


RRC Connection Setup/Complete phase, so there is no "RNC
level" statistical distribution available for actual cause for RRC
connection establishment delay (amount of lost RRC
Connection Requests is not known)
Values less than 3000ms cannot be recommended for T300,
and thus there is no possibility to speed up the RRC
connection establishment in the cases where the RRC
Connection Request message is lost
Cell reselections during the RRC connection setup phase will
cause "double resource reservations where a new RRC
connection establishment is started in the new cell. Ongoing
RRC connection establishment in the old cell cannot be
interrupted and the reserved resources cannot be released
until the Layer3 time supervision (default=6 seconds) expires
41

NOKIA

Call Performance.PPT / 16-09-2003 /Reunanen Jussi

Company Confidential

Before 1.5.2 CD15

Call Setup Performance


-RRC Connection Establishment-

UE

Node B

RNC

[RACH] RRC:RRC Connection Request


NBAP: RL Setup Request

Start
Start
TX/RX
TX/RX
NBAP: RL Setup Response
ALCAP:ERQ
ALCAP:ECF
[FACH] RRC: RRC Connection Setup

Start
Start
TX/RX
L1 Synchronisation

Timer
T300
is
starte
d

UE sends
RRC
Connectio
n Request
-message

NBAP: Synchronisation Indication

NBAP:R
L Setup
Reques
t

RNC starts
timer
T_RRC_Resp_C
CH
RNC sends RRC
Connection Setup
-message

BTS Starts
sending DPCCH:
Pilot and TPC &
Starts UL
synchronization
procedure

[DCH] RRC: RRC Connection Setup Complete

BTS sends: NBAP


RADIO LINK
SETUP RESPONSE
-message

2
UE start
sending
DPCCH for
UL sync
purposes
(pilot) after
DL frame
sync is
established

The UL synchronization is based on the 1024 chips timing offset


between the uplink and downlink DPCHs, the channelization code
and the scrambling code of the uplink DPCH specified in the radio
link setup parameters
When the UE has established the frame synchronization to the
downlink DPCH, it starts the transmission of the uplink DPCH
The DPDCH is transmitted only when there are Transport Blocks
received from the UE L2
42

NOKIA

Call Performance.PPT / 16-09-2003 /Reunanen Jussi

Company Confidential

Call Setup Performance


-RRC Connection Establishment-

Before the synchronization of an uplink DPCH is established, the


radio link is in initial synchronization state
The synchronization state of a radio link is determined based on
the physical channel BER of the DPCCH
The physical channel BER is the relation of the incorrectly
detected pilot bits to the total number of pilot bits in a radio
frame
L1 estimates the synchronization status (In-sync or Out-of-sync) of
each radio frame
The criteria for estimating the synchronisation status are defined in
two different phases

43

NOKIA

Call Performance.PPT / 16-09-2003 /Reunanen Jussi

Company Confidential

Call Setup Performance


-RRC Connection Establishment-

The first phase


starts when the searching of the synchronization is started and
lasts 160 ms after the first frame with In-synch status has been
found
During this time the Out-of-sync status shall not exist
During the first phase the radio frame has the In-sync status if the
following criterion is fulfilled:
The physical channel BER of the DPCCH over the previous 40
ms period has been better than a threshold Qin (def. 20% BER)
This criterion shall be assumed not to be fulfilled before 40 ms
of DPCCH BER measurements have been collected
Qin (A radio frame has an In-sync status, if, in addition to
other criteria, the physical channel BER of a DPCCH over
the last 160 ms has been better than the threshold Q_IN.
Reported valueof the Q_IN
Measured
quantityto
valuethe actual physical
The mapping
values
PhCh_ BER_LOG_ 000
Physical channel BER = 0
- < Log10(Physical
channel
BER)25.133
< -2.06375
PhCh_ BER_LOG_
001 given
channel
BER
is
in
3GPP
TS

-2.06375 Log10(Physical channel BER) < -2.055625


PhCh_ BER_LOG_ 002

44

NOKIA

PhCh_ BER_LOG_ 003

-2.055625 Log10(Physical channel BER) < -2.0475

PhCh_ BER_LOG_ 253


PhCh_ BER_LOG_ 254
PhCh_ BER_LOG_ 255

-0.024375 Log10(Physical channel BER) < -0.01625


-0.01625 Log10(Physical channel BER) < -0.008125
-0.008125 Log10(Physical channel BER) 0

Call Performance.PPT / 16-09-2003 /Reunanen Jussi

Company Confidential

Call Setup Performance

-RRC Connection Establishment-

The second phase


starts 160 ms after the first frame with In-sync status has been found
During this phase the criteria for the Out-of-sync and In-sync status are as
follows
A radio frame has an Out-of-sync status if the following criterion is
fulfilled
The physical channel BER of the DPCCH over the previous 160
ms period has been worse than the threshold Q out (def. 15% BER)

Qout A radio frame has an Out-of-sync status, if the physical


channel BER of a DPCCH over the last 160 ms has been worse
than the threshold Q_OUT. The mapping of the Q_OUT values to
the actual physical channel BER is given in 3GPP TS 25.133.
A radio frame has an In-sync status if the following criterion is fulfilled
the physical channel BER of the DPCCH over the previous 160
ms period has been better than a threshold Q in

NOTE: It is possible that a radio frame has neither In-sync status nor Out-ofsync status
These frames are ignored in determining the sync state of a radio link

45

NOKIA

Call Performance.PPT / 16-09-2003 /Reunanen Jussi

Company Confidential

Call Setup Performance


-RRC Connection Establishment-

When the BTS L1 has detected N_INSYNC_IND consecutive


indications with In-sync status, the radio link is moved from the
initial state to an In-sync state
L1 informs BTS L3 about the established synchronization and BTS
L3 sends the NBAP:SYNCHRONIZATION INDICATION message to
the RNC
The parameters Qin and Qout and N_INSYNC_IND are given by the RNC
to the BTS in the NBAP: CONFIGURATION DATA message
The BTS L1 searches the synchronization in the initial state as long
as the synchronization has been established or the radio link is
released by the RNC with the NBAP:RADIO LINK DELETION message
After the BTS has established the frame synchronization to the uplink
DPCH, the transmission power of the downlink DPCH is controlled
based on the TPC bits transmitted by the UE
Also, the TPC bits transmitted in the downlink dedicated physical
channel are based on the SIR measurements from the uplink
DPCH
46

NOKIA

Call Performance.PPT / 16-09-2003 /Reunanen Jussi

Company Confidential

Before 1.5.2 CD15

Call Setup Performance


-RRC Connection Establishment-

UE

Node B

RNC When a physical dedicated channel establishment is initiated by

[RACH] RRC:RRC Connection Request


NBAP: RL Setup Request

Start
Start
TX/RX
TX/RX
NBAP: RL Setup Response
ALCAP:ERQ
ALCAP:ECF
[FACH] RRC: RRC Connection Setup

the UE, the UE starts a timer T312 and wait for layer 1 to
indicate N312 "in sync" indications
On receiving N312 "in sync" indications, the physical channel is
considered established and the timer T312 is stopped and reset
On the BTS side after receiving N_INSYNC_IND synchronisation
indicators the BTS sends NBAP: SYNCHRONIZATION INDICATION
message to RNC after which the closed loop and outer loop PC
start to control the powers

Start
Start
TX/RX
L1 Synchronisation
NBAP: Synchronisation Indication
[DCH] RRC: RRC Connection Setup Complete

Timer T312
started

UE initiates physical
dedicated channel
establishment before
sending e.g. RRC
Connection Setup Complete
message on DPDCH
47

NOKIA

N_INSYNC_IND
indicators on
L1

in sync
indicators on
L1
N312 L1 in
sync
indicators

L1 Synchronization
established
BTS sends NBAP:
SYNCHRONIZATION
IDICATION
-message
Timer
T312
stopped
L1
Synchronizati
on
established

Other cases HHO,


FACH=>DCH-state
Call
Performance.PPT / 16-09-2003 /Reunanen Jussi Company Confidential

RNC receives RRC


Connection Setup
Complete message
->
RNC stops timer
T_RRC_Resp_CCH

RRC Connection Setup


Complete message
SENT

UE
Node B
Before 1.5.2 CD15

RNC

[RACH] RRC:RRC Connection Request


NBAP: RL Setup Request

Start
Start
TX/RX
TX/RX
NBAP: RL Setup Response
ALCAP:ERQ
ALCAP:ECF
[FACH] RRC: RRC Connection Setup

Start
Start
TX/RX
L1 Synchronisation

Timer
T312
starte
d

-RRC Connection Establishment-

In case UE is not able to establish synchronization


within timer T312 it stops TX on the DCH
In case BTS is not able to establish synchronization
it does not send NBAP:Synchronization Indication
message to RNC
The BTS tries to establish synchronization until
T_RRC_Resp_CCH timer in RNC expires and RNC
sends NBAP:Radio Link Deletion -message

Less than
N_INSYNC_IND
indicators on
L1

2
UE initiates physical
dedicated channel
establishment before
sending e.g. RRC
Connection Setup Complete
message on DPDCH

Call Setup Performance

in sync
indicators on
L1
Less
thanN312 L1
in sync
indicators

RNC does not


receive RRC
Connection Setup
Complete message
within
T_RRC_Resp_CCH
timer
Counter
1001C9 is
RNC sends
incremented
NBAP:RADIO LINK
DELETION
message
RRC Connection
establishment
procedure failed -> wait
until T300 expires and if
number of sent RRC
Connection Requests <
N300

L1 Synchronization not
established BTS does NOT
send NBAP:
SYNCHRONIZATION
IDICATION -message

Timer T312
expired => L1
Synchronization
not established

Physical channel
Other cases HHO,
establishment
FACH=>DCH-state
failed => UE stops
48
NOKIA
Call Performance.PPT / 16-09-2003 /Reunanen Jussi Company Confidential
transitions
TX/RX

UE
Node B
Before[RACH]
1.5.2
CD15
RRC:RRC
Connection Request

Call Setup Performance

RNC

NBAP: RL Setup Request

Start
Start TX/RX
TX/RX
NBAP: RL Setup Response
ALCAP:ERQ
ALCAP:ECF

-RRC Connection Establishment-

In case BTS is able to establish synchronization it sends


NBAP:Synchronization Indication message to RNC

In case UE is not able to establish synchronization within timer


T312 it stops TX on the DCH
As the UE TX is off the BTS looses the L1 synchronization and
sends NPAB: Radio Link Failure message to RNC (and
continues to send the message in 5 sec. intervals until RNC
sends NBAP: Radio Link Deletion message back)

[FACH] RRC: RRC Connection Setup

Start
Start TX/RX
TX/RX
L1 Synchronisation
NBAP: Synchronisation Indication
NBAP: Radio Link Failure

After T_RRC_Resp_CCH expires in RNC the RNC sends NPAB:


Radio Link Deletion to BTS which then stops searching for the
synchronization
Counter
1001C10 is
incremented

L1 Synchronization established
BTS sends NBAP: SYNCHRONIZATION
IDICATION -message
Timer T312
started

N_INSYNC_IND
indicators on
L1

3
UE initiates physical
dedicated channel
establishment before
sending e.g. RRC
Connection Setup
Complete message
on DPDCH
49

NOKIA

in sync
indicators on
L1
Less than N312 L1
in sync indicators

Timer
T312
expired

Timer
T_RLFAILURE
expires
N_OUTSYNC_IN
timer
D indicators on
T_RRC_Resp_
L1
CCH expired

Physical channel
establishment
failed => UE stops
TX/RX

Timer
T_RLFAILU
RE started

RRC Connection establishment procedure


failed -> wait until T300 expires and if
Call Performance.PPT
/ 16-09-2003
/Reunanen
Jussi Company
Confidential
number
of sent
RRC Connection
Requests

RNC receives
NBAP:RADIO LINK
FAILURE message form
BTS with a cause value
"Synchronization
Failure"

RNC sends
NBAP:RADIO
LINK
DELETION
message

UE
Node B
Before[RACH]
1.5.2
CD15
RRC:RRC
Connection Request

Call Setup Performance

RNC

-RRC Connection Establishment-

NBAP: RL Setup Request

Start
Start TX/RX
TX/RX
NBAP: RL Setup Response
ALCAP:ERQ
ALCAP:ECF
[FACH] RRC: RRC Connection Setup

Start
Start TX/RX
TX/RX
L1 Synchronisation
NBAP: Synchronisation Indication

In case the BTS does not send NBAP: RL Failure


message to RNC but the RNC does not
receive RRC Connection Setup Complete
message from UE within timer
T_RRC_Resp_CCH, the RNC initiates RRC
Connection Release procedure
Counter
1001C10 is
incremented

L1 Synchronization established
BTS sends NBAP: SYNCHRONIZATION
IDICATION -message
Timer T312
started

N_INSYNC_IND
indicators on
L1

3
UE initiates physical
dedicated channel
establishment before
sending e.g. RRC
Connection Setup
Complete message
on DPDCH
50

NOKIA

in sync
indicators on
L1

Timer
T312
expired

timer
T_RRC_Resp_
CCH expired

Physical channel
establishment
failed => UE stops
TX/RX

Less than N312 L1


in sync indicators

RRC Connection establishment procedure


failed -> wait until T300 expires and if
Call Performance.PPT
/ 16-09-2003
/Reunanen
number
of sent
RRC Jussi
Connection
CompanyRequests
Confidential

RNC sends
RRC:RRC
CONNECTI
ON
RELEASE
message
on DCH to
UE

RNC sends
NBAP:RADIO LINK
DELETION
message
No RRC: RRC CONNECTION
RELEASE COMPLETE message
is not received => RRC entity
of the UE is deleted at the
same time with the radio link
deletion procedure
No RRC: RRC CONNECTION
RELEASE COMPLETE message
from UE within within the
message repetition timer
(100ms)
the RNC retransmits the
message
L3Rep_RRC_Conn_Rel times
with interval of 100 ms

After 1.5.2 CD15


UE

Call Setup Performance

RNC starts timers


-RRC Connection EstablishmentT_RRC_Resp_CCH,

Node B

RNC

[RACH] RRC:RRC Connection Request


NBAP: RL Setup Request

Start
Start
TX/RX
TX/RX
NBAP: RL Setup Response
ALCAP:ERQ
ALCAP:ECF
[FACH] RRC: RRC Connection Setup

Start
Start
TX/RX
L1 Synchronisation

RNC receives RRC


RRCconnRepTimer1
Connection Request and
and
starts buffering timer
RRCconnRepTimer2
N300*T300
Timer
NBAP:R
T300
L Setup
RNC sends RRC
is
Reques
Connection Setup
starte
t
-message
d

UE sends
RRC
Connectio
n Request
-message

NBAP: Synchronisation Indication

BTS Starts
sending DPCCH:
Pilot and TPC &
Starts UL
synchronization
procedure

[DCH] RRC: RRC Connection Setup Complete

BTS sends: NBAP


RADIO LINK
SETUP RESPONSE
-message

UE start
sending
DPCCH for
UL sync
purposes
(pilot) after
DL frame
sync is
established

The UL synchronization is based on the 1024 chips timing offset between


the uplink and downlink DPCHs, the channelization code and the scrambling
code of the uplink DPCH specified in the radio link setup parameters
When the UE has established the frame synchronization to the downlink
DPCH, it starts the transmission of the uplink DPCH
The DPDCH is transmitted only when there are Transport Blocks received
from the UE L2
RRCconnRepTimer1 and RRCconnRepTimer2 are having static values of
400ms and 1400ms respectively
51

NOKIA

Call Performance.PPT / 16-09-2003 /Reunanen Jussi

Company Confidential

Call Setup Performance

After 1.5.2 CD15

-RRC Connection Establishment-

UE

Node B

RNC When a physical dedicated channel establishment is initiated by

[RACH] RRC:RRC Connection Request


NBAP: RL Setup Request

Start
Start
TX/RX
TX/RX
NBAP: RL Setup Response
ALCAP:ERQ
ALCAP:ECF
[FACH] RRC: RRC Connection Setup

the UE, the UE starts a timer T312 and wait for layer 1 to
indicate N312 "in sync" indications
On receiving N312 "in sync" indications, the physical channel is
considered established and the timer T312 is stopped and reset
On the BTS side after receiving N_INSYNC_IND synchronisation
indicators the BTS sends NBAP: SYNCHRONIZATION INDICATION
message to RNC after which the closed loop and outer loop PC
start to control the powers

Start
Start
TX/RX
L1 Synchronisation
NBAP: Synchronisation Indication
[DCH] RRC: RRC Connection Setup Complete

Timer T312
started

UE initiates physical
dedicated channel
establishment before
sending e.g. RRC
Connection Setup Complete
message on DPDCH
52

NOKIA

N_INSYNC_IND
indicators on
L1

in sync
indicators on
L1
N312 L1 in
sync
indicators

L1 Synchronization
established
BTS sends NBAP:
SYNCHRONIZATION
IDICATION
-message

RNC receives RRC


Connection Setup
Complete message ->
RNC stops timers
T_RRC_Resp_CCH,
RRCconnRepTimer1
and
RRCconnRepTimer2

Timer
T312
stopped
L1
Synchronizati
on
established

RRC Connection Setup


Complete message
SENT

Other cases HHO,


FACH=>DCH-state
Call
Performance.PPT / 16-09-2003 /Reunanen Jussi Company Confidential

UE

Node B

RNC

[RACH] RRC:RRC Connection Request


NBAP: RL Setup Request

Start
Start
TX/RX
TX/RX
NBAP: RL Setup Response
ALCAP:ERQ
ALCAP:ECF
[FACH] RRC: RRC Connection Setup

Start
Start
TX/RX
L1 Synchronisation

Timer
T312
starte
d

Call Setup Performance


After 1.5.2 CD17
-RRC Connection Establishment-

In case UE is not able to establish synchronization


within timer T312 it stops TX on the DCH
In case BTS is not able to establish synchronization
it does not send NBAP:Synchronization Indication
message to RNC
The BTS tries to establish synchronization until
T_RRC_Resp_CCH timer in RNC expires and RNC
sends NBAP:Radio Link Deletion -message

Less than N_INSYNC_IND indicators


on L1
L1 Synchronization not
RRC Connection Setup
established BTS does NOT
Complete not received
send NBAP:
RRCconnRepTimer1 expires
SYNCHRONIZATION
-> RRC Connection Setup is
IDICATION
-message
resent to the UE
RRC Connection Setup
Complete not received
RRCconnRepTimer2 expires
-> RRC Connection Setup is
resent to the UE and
RRCconnRepTimer2 reset
RRCconnRepTimer1 stopped

Timer T312
expired => L1
Synchronization
not established

UE initiates physical
dedicated channel
establishment before
sending e.g. RRC
Connection Setup Complete
message on DPDCH
Physical channel
establishment
Less
thanN312
L1
in
sync
Other cases HHO,
failed => UE stops
indicators
FACH=>DCH-state
53
NOKIA
Call Performance.PPT / 16-09-2003 /Reunanen Jussi Company Confidential TX/RX
transitions

RNC does not


receive RRC
Connection Setup
Complete message
within
T_RRC_Resp_CCH
timer
RNC sends
NBAP:RADIO LINK
DELETION
message
Counter 1001C9 is
incremented
-no Synchronisation
indication received
within supervision time
-no rrc connection
setup complete reeived
within supervision time

UE

Node B

RNC

[RACH] RRC:RRC Connection Request


NBAP: RL Setup Request

Start
Start TX/RX
TX/RX
NBAP: RL Setup Response
ALCAP:ERQ
ALCAP:ECF
[FACH] RRC: RRC Connection Setup

Call Setup Performance


After 1.5.2 CD17
-RRC Connection Establishment-

In case BTS is able to establish synchronization it sends NBAP:Synchronization Indication


message to RNC
In case UE is not able to establish synchronization within timer T312 it stops TX on the DCH
As the UE TX is off the BTS looses the L1 synchronization and sends NPAB: Radio Link
Failure message to RNC
After T_RRC_Resp_CCH expires in RNC the RNC sends NPAB: Radio Link Deletion to BTS
which then stops searching for the synchronization

Start
Start TX/RX
TX/RX
L1 Synchronisation
NBAP: Synchronisation Indication
NBAP: Radio Link Failure

Timer
T312
starte
d

Counter 1001C10 is
L1 Synchronization established BTS
incremented
sends NBAP: SYNCHRONIZATION
-Synchronisation indication
IDICATION -message
N_INSYNC_IND indicators on L1
Timer
received within supervision
T_RLFAILU
time
RE
started
-no rrc connection setup
RRC Connection Setup
complete reeived within
Complete not received
Timer
N_OUTSYNC_IN
supervisiontimer
time
RRCconnRepTimer1 expires
T_RLFAILU
D indicators on
T_RRC_Resp_
-> RRC Connection Setup is
RE expires
L1
CCH expired
resent to the UE
RRC Connection Setup
Complete not received
RRCconnRepTimer2 expires
-> RRC Connection Setup is
resent to the UE and
RRCconnRepTimer2 reset
RRCconnRepTimer1 stopped

Timer T312
expired => L1
Synchronization
not established

UE initiates physical
dedicated channel
establishment before
sending e.g. RRC
Connection Setup Complete
message on DPDCH
Physical channel
establishment
Less thanN312 L1 in sync
Other cases HHO,
54
NOKIA
Call Performance.PPT / indicators
16-09-2003 /Reunanen Jussi Company Confidential failed => UE stops
FACH=>DCH-state

RNC sends
NBAP:RADIO
LINK
DELETION
message
RNC receives
NBAP:RADIO LINK
FAILURE message form
BTS with a cause value

UE

Node B

RNC

[RACH] RRC:RRC Connection Request


NBAP: RL Setup Request

Start
Start TX/RX
TX/RX
NBAP: RL Setup Response
ALCAP:ERQ
ALCAP:ECF
[FACH] RRC: RRC Connection Setup

Start
Start TX/RX
TX/RX
L1 Synchronisation
NBAP: Synchronisation Indication

Call Setup Performance


After 1.5.2 CD17
-RRC Connection Establishment-

In case the BTS does not send NBAP: RL Failure


message to RNC but the RNC does not
receive RRC Connection Setup Complete
message from UE within timer
T_RRC_Resp_CCH, the RNC initiates RRC
Connection
Release
procedure
Counter
1001C10 is
L1 Synchronization
incremented
-Synchronisation
indication received
within supervision
time
RNC sends
-no rrc connection
NBAP:RADIO
setup complete
LINK DELETION
reeived within
message
supervision time
RRC entity of the UE is
timer
deleted at the same time
T_RRC_Resp_CCH
with the radio link deletion
expired
procedure

established BTS sends


NBAP:
SYNCHRONIZATION
IDICATION -message
N_INSYNC_IND indicators on L1
Timer
T312
starte
d

RRC Connection Setup


Complete not received
RRCconnRepTimer1 expires
-> RRC Connection Setup is
resent to the UE
RRC Connection Setup
Complete not received
RRCconnRepTimer2 expires
-> RRC Connection Setup is
resent to the UE and
RRCconnRepTimer2 reset
RRCconnRepTimer1 stopped

UE initiates physical
dedicated channel
establishment before
sending e.g. RRC
Connection Setup Complete
message on DPDCH
Less thanN312 L1 in sync
Other cases HHO,
FACH=>DCH-state
55
NOKIA
Call Performance.PPT / indicators
16-09-2003 /Reunanen Jussi Company Confidential

RNC sends RRC:RRC CONNECTION RELEASE


message on DCH to UE, timer
T_RRC_Resp_CCH restarted
No RRC CONNECTION
RELEASE COMPLETE
message from UE within
within the message
repetition timer (100ms)
the RNC retransmits the
message
L3Rep_RRC_Conn_Rel times

Call Setup Performance


Longer time (T312 high and high N312) the UE has to establish
the L1 synchronization the higher probability successful physical
channel establishment is and better call set up success rate
However longer the time for L1 synch the longer is call setup
time
During the test it was
AMR performance
noted that setting N312 to
2 or 4 (between 1 and 2
there is notable difference)
does not have any
significant effect on the call
set up success rate
100%

90%
80%
70%
60%
50%
40%
30%
20%
10%

Percentile MOC call setup delay @5 sec


Percentile MTC call setup delay @5 sec

N31
2=1
56

NOKIA

N31
2=4

17/03/03

07/03/03

06/03/03

01/03/03

15/02/03

12/02/03

10/02/03

05/02/03

01/02/03

28/01/03

25/01/03

24/01/03

22/01/03

20/01/03

18/01/03

16/01/03

0%
14/01/03

MOC / MTC call setup delay 5 sec Percentile

-RRC Connection Establishment-

But the effect on the call


set up time is significant
and therefore N312 value
of 2 was selected to be
used

N31
2=2

Call Performance.PPT / 16-09-2003 /Reunanen Jussi

Company Confidential

Call Setup Performance

-RRC Connection Establishment: Power Offset for SCCPCH Pilot &


TFCI bits -

Secondary CCPCH support multiple transport format combinations using TFCI.


For slot formats using TFCI, the TFCI value in each radio frame corresponds to a
certain transport format combination of the FACHs and/or PCHs currently in use.
This correspondence is (re-)negotiated at each FACH/PCH addition/removal.
For example (currently used, should be noted that pilot bits are not used
according to 3GPP R99 release -> thereofre 0 bits for pilot):
slot format 8 (SF64) has 0 pilot bits and 8 TFCI bits (for FACH or combined
FACH/PCH)
slot format 0 (SF256) has 0 pilot bits and 0 TFCI bits (for PCH)
Change of offset allowing better detection on SCCPCH which carries FACH/PCH
(MOC/MTC CSSR (call setup success rate) affected)
TFCI
NTFCI bits

Data
Ndata1 bits

Pilot
Npilot bits

Tslot = 2560 chips, 20*2k bits (k=0..6)

Slot #0

Slot #1

Slot #i

Slot #14

1 radio frame: Tf = 10 ms

57

NOKIA

Call Performance.PPT / 16-09-2003 /Reunanen Jussi

Company Confidential

Call Setup Performance

-RRC Connection Establishment: Power Offset for SCCPCH Pilot &


TFCI bits -

power offset TFCI


= 0 dB for 60
ksps SCCPCH

power
offset TFCI
= 4 dB for
60 ksps
SCCPCH

58

NOKIA

Call Performance.PPT / 16-09-2003 /Reunanen Jussi

Company Confidential

Failure due to
FACH/RACH is lower from
xxx test in Apr

Call Setup Performance


-RRC Connection Establishment-

RRC Connection establishment success rate and


delay
cannot be improved by changing timer T300 or counter N300
to more aggressive
Way to improve RRC Connection success rate

Improve the UE idle mode behavior by making the intrafrequency measurements earlier and cell reselection more
aggressive (see previous slide on idle mode performance)
Typically
RRC Connection Setup Success rate (=
(RRC_CONN_ACC_COMP/ RRC_CONN_STP_ATT) *100) is in the
region of 80% - 90%
Adjust the T312 and N312 values (current defaults N312=4 and
T312=10s should be ok but alternatives to try in case of call
setup delay is an issue N312=2 or even N312=1)
Use Nokia defaults for SCCPCH offsets

RRC Connection Success rate can be low due to:

59

NOKIA

UE changing cell during RRC Connection Setup phase (cell A ->


cell B) usually counter 1001C9 is updated in the cell A
RRC Connection Setup message is not heard by the UE (no
retransmission) -> usually counter 1001C9 is updated in the cell
Call Performance.PPT / 16-09-2003 /Reunanen Jussi

Company Confidential

Call Setup Performance


-RRC Connection Establishment-

It should be noted while making the idle mode performance more


aggressive that the when the UE is making measurements it is
also using more battery so therefore the UE should not make
measurement all the time (can be done by not setting any value
for Sintrasearch)
Also common channel power settings need to be optimised in
order to really optimise the idle mode and RRC Connection Setup
performance
Common channel power optimisation is to happen during
2H2003

60

NOKIA

Call Performance.PPT / 16-09-2003 /Reunanen Jussi

Company Confidential

Call Setup Performance


-Common Channel Power Tuning-

Call Setup Success Statistics


Call setup success rate has improved after parameter change
by 4~5%.
Channels
Default
case4
CPICH Ptx
PCCPCH Ptx
PSCH
SSCH
PICH
SCCPCH
AICH

27
24
25
25
20
29
21

27
25
27
27
24
27
27
Call Setup Success Rate

AMR MTC
Parameter set # of call attempts Setup success Setup success rate
Default
484
412
85.1%
Case4
625
556
89.0%
UDI MOC
Parameter set # of call attempts Setup success Setup success rate
Default
320
262
81.9%
Case4
527
451
85.6%
61

NOKIA

Call Performance.PPT / 16-09-2003 /Reunanen Jussi

100.0%
95.0%
90.0%
85.0%
80.0%
75.0%
70.0%
65.0%
60.0%
55.0%
50.0%
45.0%
40.0%
35.0%
30.0%
25.0%
20.0%
15.0%
10.0%
5.0%
0.0%

Success rate

AMR MOC
Parameter set # of call attempts Setup success Setup success rate
Default
572
473
82.7%
Case4
771
678
87.9%

Company Confidential

Default
Case4

Services

Call Setup Performance


-Common Channel Power Tuning-

Classified the call failures according to the phases where call


setup has failed
Step A, B are related to paging, AICH, and RACH/FACH phases.
affected by PICH, AICH, and SCCPCH (PCH and FACH) increase
Index of Call Results Breakdown
A
Call Setup Failure during RRC Connection Setup Phase (2.1)
B
Call Setup Failure during Access Phase due to RACH/FACH (2.2)
C
Call Setup Failure due to RRC Connection Reject (2.3)
D
Call Setup Failure due to NO RRC Connection Setup Complete (2.9)
E
Call SetUp Failure due to No CM service request (2.8)
F
Call Setup Failure during Security Procedure (2.5)
G
Call Setup Failure during RAB Setup procedure(Without Complete) (2.6)
H
Call Setup Failure during RAB Setup procedure(With Complete) (2.7)
I
Call Setup Success
J
Call Setup Failure due to Registration (2.4)

In Step A & B, the performance after parameter change has


AMR MOC
AMR MTC
UDI MOC
improved
Category Default Case4
Default (%) Case4 (%) Category Default Case4
Default (%) Case4 (%)
A
8
0
1.40%
0.00% A
3
0
0.62%
0.00%
B
19
26
3.32%
3.37% B
5
15
1.03%
2.40%
C
0
0
0.00%
0.00% C
0
1
0.00%
0.16%
D
36
32
6.29%
4.15% D
21
20
4.34%
3.20%
E
7
4
1.22%
0.52% E
2
7
0.41%
1.12%
F
11
4
1.92%
0.52% F
28
10
5.79%
1.60%
G
5
8
0.87%
1.04% G
6
4
1.24%
0.64%
H
13
17
2.27%
2.20% H
7
12
1.45%
1.92%
I
473
678
82.69%
87.94% I
412
556
85.12%
88.96%
J
0
2
0.00%
0.26%
J
0
0
0.00%
0.00%
62
NOKIA
Call Performance.PPT / 16-09-2003 /Reunanen Jussi Company Confidential

Category Default Case4


Default (%) Case4 (%)
A
0
0
0.00%
0.00%
B
18
26
5.63%
4.93%
C
0
0
0.00%
0.00%
D
18
25
5.63%
4.74%
E
7
3
2.19%
0.57%
F
6
5
1.88%
0.95%
G
3
6
0.94%
1.14%
H
6
11
1.88%
2.09%
I
262
451
81.88%
85.58%
J
0
0
0.00%
0.00%

Call Setup Performance

-RRC Connection Establishment: CountersUE

RNC

BTS

Failure Case:
RRC Connection Setup fa
Due to
Triggers:
1001C2 Handover control
1001C3 Admission control
1001C4 BTS
1001C5 Transmission netwo
1001C6 RNC internal reason
1001C7 Frozen BTS

CN

RRC: Connection Request


Correctly Received message
RRC Connection Setup Starts
Triggers:
1001C0 RRC Connection Attempt

RRC Connection Setup phase

Resource Reservation in RNC, BTS, Transmission


RRC: RRC Connection Request Setup
RRC Connection setup completed
Triggers:
1001C1 RRC Con Setup Compl

RRC Connection Access phase


RNC waits reply from UE

RRC: RRC Connection Completed

Failure Case:
RRC Connection Access fails
Due to
Triggers:
1001C9
L1 synchronization
1001C10 UU interface
1001C11 RNC internal reason

Failure Case:
RRC Connection Active f
Due to
Triggers:
1001C15 Iu interface
1001C16 Radio interface
RANAP: Initial UE Message 1001C17 BTS
1001C18 Iur interface
1001C19 Ciphering failur
1001C20 Integrity check
1001C21 RNC internal re

RRC Connection Access completed


Triggers:
1001C8 RRC Con Acc Comp

RRC: Initial Direct Transfer

RRC Connection Active phase


UE-CN Signaling

RANAP: Iu Release Command


RRC Connection Active completed
Triggers:
1001C12 RRC Con Active Comp

63

NOKIA

Call Performance.PPT / 16-09-2003 /Reunanen Jussi

Company Confidential

Special Reason:
RRC Connection Active
Release Due to
Triggers:
1001C13 SRNC Relocat
1001C14 Pre-emption

RRC Connection Setup due to


Cause
Triggers:
1001C22 MOC conv call att
1001C24 MOC stream call att
1001C26 MOC interac call att
1001C28 MOC backgr call att
1001C30 MOC subscribed call att
1001C32 MTC conv call att
CN 1001C34 MTC stream call att
RNC
UE
BTS
1001C36 MTC interac call att
1001C38 MTC backgr call att
RRC: Connection Request
1001C40 Emergency call att
1001C42 Inter RAT cell reselection
Correctly Received message
1001C44 Inter RAT cell chng ord
Triggers:
1001C46 Registration attempt
1001C0 RRC Connection Attempt
1001C48 Detach attempt
1001C50 MOC high prior signaling
RRC Connection Setup phase
1001C52 MTC high prior signaling
Resource Reservation in RNC, BTS, Transmission
1001C54 MOC low prior signaling
RRC: RRC Connection Request Setup
1001C56 MTC high prior signaling
RRC Connection setup completed
1001C58 Call re establishment att
Triggers:
1001C60 Terminating cause
1001C1 RRC Con Setup Compl
unknown
Failures by RRC Connection
RRC Connection Access phase
Setup Causes
RNC waits reply from UE
Triggers:
1001C23 MOC conv call failure
RRC: RRC Connection Completed
1001C25 MOC stream call failure
RRC Connection Access completed
1001C27 MOC interac call failure
Triggers:
1001C29 MOC backgr call failure
1001C8 RRC Con Acc Comp
1001C31 MOC subscribed call fail
RRC: Initial Direct Transfer
1001C33 MTC conv call failure
RANAP: Initial UE Message1001C35 MTC stream call failure
1001C37 MTC interac call failure
1001C39 MTC backgr call failure
RRC Connection Active phase
1001C41 Emergency call failure
UE-CN Signaling
1001C43 Inter RAT cell reselec fail
1001C45 Inter RAT cell chng ord fail
RANAP: Iu Release Command1001C47 Registration attempt fai
1001C49 Detach attempt failure
1001C51 MOC high prior sig failure
RRC Connection Active completed
1001C53 MTC high prior sig failure
Triggers:
1001C55 MOC low prior sig failure
1001C12 RRC Con Active Comp
1001C57 MTC high prior sig failure
1001C59 Call re establishment fail
1001C61 Terminating cause
unknown failure

Call Setup Performance

-RRC Connection Establishment: Counters-

64

NOKIA

Call Performance.PPT / 16-09-2003 /Reunanen Jussi

Company Confidential

Call Setup Performance

-RRC Connection Establishment: PIsUE

RNC

BTS

CN

RRC: Connection Request


Correctly Received message
RRC Connection Setup Starts
Triggers:
1001C0 RRC Connection AttemptRRC Setup Complete Ratio
100*[Sum(1001C1)/Sum(1001C0)]

RRC Connection Setup phase

Resource Reservation in RNC, BTS, Transmission


RRC: RRC Connection Request Setup
RRC Connection setup completed
Triggers:
1001C1 RRC Con Setup Compl

RRC Setup and


Access Complete Ratio [%

100*[Sum(1001C8)/Sum(1001C0

RRC Connection Access phase

Special Reason:
RRC Connection Active
Release Due to
Triggers:
1001C13 SRNC Relocatio
1001C14 Pre-emption

RNC waits reply from UE

RRC: RRC Connection Completed


RRC Connection Access completed
Triggers:
1001C8 RRC Con Acc Comp

RRC: Initial Direct Transfer

RANAP: Initial UE Message

RRC Drop Ratio [%]

RRC Connection Active phase

100-100*[Sum(1001C12+100
+1001C14)/Sum(1001C8)]

UE-CN Signaling

RANAP: Iu Release Command


RRC Connection Active completed
Triggers:
1001C12 RRC Con Active Comp

65

NOKIA

Call Performance.PPT / 16-09-2003 /Reunanen Jussi

Company Confidential

Call Setup Performance

-RRC Connection Establishment: PIs RRC Setup and Access Complete Ratio [%]
It should be noted that this PI is not call setup success rate (counted
from the user perspective) as:
the RAB setup and access phases not included
the users UE can send several RRC Connection Request
messages until the connection is successfully established
Also it should be noted that different UEs are using the RRC
Connection establishment cause values differently (counters 1001C22
1001C61)
E.g. Sanyo UE classifies UDI video call as MOC/MTC Conversational
call where as NEC VTX classifies the same UDI video call as
MOC/MTC High Priority Signaling
RRC Access Failure counters
The difference between counters 1001C9 and 1001C10 is:
1001C9 is incremented in case Synchronization Indication IS
received from BTS but no RRC Connection Setup Complete
message is received from the UE
1001C10 is incremented in case Synchronization Indication is NOT
received from BTS and no RRC Connection Setup Complete
message is received from the UE
66

NOKIA

Call Performance.PPT / 16-09-2003 /Reunanen Jussi

Company Confidential

Call Setup Performance

-RRC Connection Establishment: PIs-

The RRC Connection Setup and


Access Complete Ratio is showing
very poor results
This however does not give correct
picture from end user point of view as
the UE can send several RRC
Connection Requests and currently we
do not know how many RRC
Connection Requests the UE sends
before success

Therefore maybe better would be to calculate the RRC Connection Setup


and Access Complete Success Rate by:
In case the ratio 1001C0/1001C8 < (N300+1) then we could claim that
on average there are no call setup failures due to RRC Connection
In case the ratio 1001C0/1001C8 > (N300+1) then the RRC Connection
Setup and Access Complete success rate could be calculated as
(N300+1)/(1001C0/1001C8)
67

NOKIA

Call Performance.PPT / 16-09-2003 /Reunanen Jussi

Company Confidential

Call Setup Performance

-RRC Connection Establishment: Questions Why there is a problem with RRC Connection establishment?
What is the difference between counters 1001C9 and 1001C10 and which
one most probably is having higher value at the moment?
What are the actions with which the RRC Connection Setup and Access
Complete ration can be improved?

83

NOKIA

Call Performance.PPT / 16-09-2003 /Reunanen Jussi

Company Confidential

Contents
Call Setup Performance

Idle Mode Performance

Paging

RRC Connection Establishment Performance

RAB Establishment Performance

Call Setup Failure & Time

UL Interference

Dropped Call Performance

SHO Performance (round-the-corner)

DL Power Allocation Parameters

UL Power Allocation Parameters

PS Call Performance
84

NOKIA

Call Performance.PPT / 16-09-2003 /Reunanen Jussi

Company Confidential

Call Setup Performance


-RAB Establishment Performance-

The RAB establishment procedure starts right after the RRC


Connection establishment, signalling connection establishment,
authentication & security procedures and call setup signaling
UE
UE

RNC
RNC

Node
Node B
B

CN
CN

UE
UE

RNC
RNC

Node
Node B
B

RRC:Initial Direct Transfer(MM:CM Service Request)

CN
CN

RANAP:Direct Transfer(CC:Call Proceeding)

RANAP:Initial UE Message(MM:CM Service


Request)
SCCP:CC

RRC:DL Direct Transfer(CC:Call Proceeding)


RANAP:RAB Assignment Request
NBAP:Radio Link Reconfiguration Prepare

RANAP:Direct Transfer(MM:Authentication Request)

NBAP:Radio Link Reconfiguration ready

RRC:DL Direct Transfer(MM:Authentication Request)

ALCAP:ERQ
ALCAP:ECF

RRC:UL Direct Transfer(MM:Authentication


Response)
RANAP:Direct Transfer(MM:Authentication
Response)
RANAP:Common ID
RANAP:Security Mode Command

ALCAP:ERQ
ALCAP:ECF
NBAP:Radio Link Reconfiguration Commit
RRC:Radio Bearer Setup
RRC:Radio Bearer Setup Complete

RRC:Security Mode Command

RANAP:RAB Assignment Response

RRC:Security Mode Complete


RANAP:Security Mode Complete

RRC:DL Direct Transfer(CC:Alerting)

RRC:UL Direct Transfer(CC:Setup)


RANAP:Direct Transfer(CC:Setup)

RRC:DL Direct Transfer(CC:Connect)

RANAP:Direct
Transfer(CC:Alerting)
RANAP:Direct
Transfer(CC:Connect)

RRC:UL Direct Transfer(CC:Connect Acknowledge)


RANAP:Direct Transfer(CC:Connect Acknowledge)
85

NOKIA

Call Performance.PPT / 16-09-2003 /Reunanen Jussi

Company Confidential

Call Setup Performance

-RAB Establishment Performance: Counters; RAB Setup &


AccessRAB Setup starts
BTS

UE

RRC Connection Active


UE-CN Signaling

Depending on RAB Type


Triggers:
RNC
CN
1001C66 RAB setup att for CS Voice
1001C67 RAB setup att for CS Data conv
1001C68 RAB setup att for CS Data strea
phase
1001C69 RAB setup att for PS Data conv
1001C70 RAB setup att for PS Data strea
RANAP: RAB Assignment Request 1001C71 RAB setup att for PS Data intera
1001C72 RAB setup att for PS Data backg

RAB Setup phase

Resource Reservation in RNC, BTS, Transmission


RRC: Radio Bearer Setup

RAB Connection Access phase


RNC waits reply from UE

RRC: Radio Bearer Setup Completed


RANAP: RAB Assignment Response

RAB Active phase

User Plane Data Transfer

86

NOKIA

RAB Access Completed


Triggers:
1001C115 RAB setup acc comp
1001C116 RAB setup acc comp
conv class
1001C117 RAB setup acc comp
stream class
1001C118 RAB setup acc comp
conv class
1001C119 RAB setup acc comp
stream class
1001C120 RAB setup acc comp
Call Performance.PPT / 16-09-2003 /Reunanen Jussi Company Confidential
intera class

for CS Voice
for CS Data
for CS Data
for PS Data
for PS Data
for PS Data

Failure Case:
CS RAB Setup for Voice fails D
Triggers:
1001C80 Admission control
1001C81 BTS
1001C82 Transmission network
1001C83 RNC internal reason
1001C84 Frozen BTS
Note: Each RAB type
has identical failure counters

RAB Setup Completed


Depending on RAB Type
Triggers:
1001C73 RAB setup comp for
1001C74 RAB setup comp for
1001C75 RAB setup comp for
1001C76 RAB setup comp for
1001C77 RAB setup comp for
1001C78 RAB setup comp for
1001C79 RAB setup comp for

CS Voice
CS Data conv
CS Data strea
PS Data conv
PS Data strea
PS Data inter
PS Data back

Failure Case:
CS RAB Access for Voice fails
Due to
Triggers:
1001C122 UE
1001C123 RNC internal
Note: Each RAB type
has identical failure counters

Call Setup Performance

-RAB Establishment Performance: Counters RAB Active PhaseUE

RNC

BTS

CN

RAB Active phase

Special Reason:
RAB Active Release Due to
Triggers:
RANAP: RAB Assignment Request, with IE: RAB reconfiguration
1001C143 SRNC Relocation for CS Vo
1001C144 Pre-emption for CS Voice
For all types of RABs

User Plane Data Transfer

RAB Reconfiguration

Resource Reconfigured in RNC, BTS, Transmission


RRC: Radio Bearer Reconfiguration
RRC: Radio Bearer Reconfiguration Complete

RAB Reconfiguration Attemp


Triggers:
1001C197 RAB reconf Att
For all types of RABs

RANAP: RAB Assignment Response

RAB Reconfiguration Failure


Triggers:
1001C198 RAB reconf Fail
For all types of RABs
RANAP: RAB Assignment Request, with IE: RAB Release
RRC: Radio Bearer Release
RAB Active Completed
Depending on RAB Type
RANAP: RAB Assignment Response Triggers:
1001C136 RAB Act Comp for
1001C137 RAB Act Comp for
RAB Release
1001C138 RAB Act Comp for
Resource Released in RNC, BTS, Transmission
1001C139 RAB Act Comp for
1001C140 RAB Act Comp for
1001C141 RAB Act Comp for
1001C142 RAB Act Comp for

RRC: Radio Bearer Reconfiguration Complete

87

NOKIA

Call Performance.PPT / 16-09-2003 /Reunanen Jussi

Company Confidential

CS Voice
CS Data conv
CS Data strea
PS Data conv
PS Data strea
PS Data inter
PS Data back

Call Setup Performance

-RAB Establishment Performance: PIsBTS

UE

RNC

CN

RANAP: RAB Assignment Request

RAB Setup phase

Resource Reservation in RNC, BTS, Transmission

RAB Setup Complete Ratio [%

100*[Sum(1001C7379)/Sum(1001C66

RRC: Radio Bearer Setup

RAB Setup and


Access Complete Ratio [%]

RAB Connection Access phase

100*[Sum(1001C115..121)/Sum(1001C66

RNC waits reply from UE

RRC: Radio Bearer Setup Completed


RANAP: RAB Assignment Response

RANAP: RAB Assignment Request, with IE: RAB Release


RRC: Radio Bearer Release
RRC: Radio Bearer Reconfiguration Complete
RANAP: RAB Assignment Response

88

NOKIA

Call Performance.PPT / 16-09-2003 /Reunanen Jussi

Company Confidential

RAB Drop Ratio [%]

100-100*[Sum(1001C136142+1001
1001C144+1001C151154+1001C16
/Sum(1001C115121)]

Call Setup Performance

-RAB Establishment Performance: PIs RAB Setup and Access Complete Ratio [%]
This PI gives success rate for the RAB establishment this
however is not Call Setup Success Rate as it does not include
RRC phase
It is calculated as the ratio between received RANAP:
Assignment Request messages and sent RANAP: Assignment
Response (showing successful case) -messages
RAB Drop Ratio [%]
This PI can be used as dropped call rate
And it is calculated as sent RANAP: RAB Assignment Response
(showing successfully released RAB) messages summed with
RAB releases due to SRNC relocation (separate counters for
each traffic class) and RAB releases due to pre-emption
(separate counters for each traffic class) divided by amount of
RANAP: RAB Assignment Response (separate counters for each
traffic class)

89

NOKIA

Call Performance.PPT / 16-09-2003 /Reunanen Jussi

Company Confidential

Call Setup Performance


-RAB Establishment Performance-

UE
UE

RNC
RNC

Node B

CN
CN

RANAP:Direct Transfer(CC:Call Proceeding)


RRC:DL Direct Transfer(CC:Call Proceeding)
RANAP:RAB Assignment Request
NBAP:Radio Link Reconfiguration Prepare
NBAP:Radio Link Reconfiguration ready
ALCAP:ERQ
ALCAP:ECF
ALCAP:ERQ
ALCAP:ECF
NBAP:Radio Link Reconfiguration Commit
RRC:Radio Bearer Setup
RRC:Radio Bearer Setup Complete
RANAP:RAB Assignment Response

RRC:DL Direct Transfer(CC:Alerting)

RANAP:Direct
Transfer(CC:Alerting)
RANAP:Direct
Transfer(CC:Connect)

RRC:DL Direct Transfer(CC:Connect)


RRC:UL Direct Transfer(CC:Connect Acknowledge)

It should be noted that during


RAB Assignment procedure (i.e.
between RAB Assignment
Request and RAB Assignment
Response) the SHO activation is
not possible -> measurement
reports are rejected (e1a and
e1c) or buffered (e1b)
From UE logs point of view
between CALL PROCEEDING
and ALERTING
This phenomena (parallel
procedure not allowed) is not
taken into account by the
counters i.e. all the failure
counters for SHO are updated if
measurement reports are
received during this period

RANAP:Direct Transfer(CC:Connect Acknowledge)


90

NOKIA

Call Performance.PPT / 16-09-2003 /Reunanen Jussi

Company Confidential

Call Setup Performance


-RAB Establishment Performance-

As it can be seen the


RAB Setup Complete
ratio and RAB Setup and
Access Complete ratio
are extremely good
showing basically 100%
success rate for RAB
establishment

Therefore in case the whole call setup success rate is evaluated as RRC
Setup and Access Success rate * RAB Setup and Access Success rate the
results are not so good due to RRC Connection Setup and Access complete

91

NOKIA

Call Performance.PPT / 16-09-2003 /Reunanen Jussi

Company Confidential

Call Setup Performance

-RAB Establishment Performance : Questions What needs to be remembered concerning SHO during call setup?
What is RAB?
What is given in RAB attributes?

102

NOKIA

Call Performance.PPT / 16-09-2003 /Reunanen JussiCompany

Confidential

Contents
Call Setup Performance

Idle Mode Performance

Paging

RRC Connection Establishment Performance

RAB Establishment Performance

Call Setup Failure & Time

UL Interference

Dropped Call Performance

SHO Performance (round-the-corner)

DL Power Allocation Parameters

UL Power Allocation Parameters

PS Call Performance
103

NOKIA

Call Performance.PPT / 16-09-2003 /Reunanen JussiCompany

Confidential

Failure Case:
CS RAB Setup for Voice fails Due
to Triggers:
1001C80 Admission control
1001C81 BTS
1001C82 Transmission network
1001C83 RNC internal reason
1001C84 Frozen BTS
Note: Each RAB type has identical
failure counters

Call Setup Performance


-Failure due to AC-

Failure Case:
CS RAB Access for Voice fails
Due to
Triggers:
1001C122 UE
1001C123 RNC internal
Note: Each RAB type
has identical failure counters

If the reason for failure is Admission Control, it can be due to


Uplink: PrxTarget is reached for PrxTotal or PrxTarget+PrxOffset is
reached for average PrxTotal
Traffic,
External Interference,
See next slides for
Cabling,
details
MHA settings,
UE Power Control.
Downlink: PtxTarget is reached for PtxTotal or
PtxTarget+PtxOffset is reached for average PtxTotal
Traffic
Cabling
Size of cell and UE locations (all at cell edge ?)
DL Power allocation
See next slides for

The reason can be identified in the traffic table


details
counters
Signalling link: DCH_REQ_LINK_REJ_UL_SRNC & DCH_REQ_LINK_REJ_DL_SRNC
Conversational: REQ_CS_CONV_REJ_UL_SRNC & REQ_CS_CONV_REJ_DL_SRNC
104

NOKIA

Call Performance.PPT / 16-09-2003 /Reunanen JussiCompany

Confidential

Call Setup Performance


-Investigating Call Setup FailureMT
C

MTC/MOC ?

Cause:
Trans, BTS

Investigate
paging : See
chapter Paging

Cause: AC

ServLev:
RRC Setup failures ?

Investigate
transmission
or BTS
Cause:
Trans, BTS

Failure Case:
CS RAB Access for Voice fails
Due to
Triggers:
1001C122 UE
1001C123 RNC internal
Note: Each RAB type
has identical failure counters

105

NOKIA

ServLev:
RAB Setup failures ?

Cause: AC

ServLev:
RRC/RAB Access
failures ?

to MS
InvestigateDue
radio
conditions,
synchronisation : See
chapter RRC Connection
Establishment
Call Performance.PPT / 16-09-2003 /Reunanen Jussi
Company Confidential

UL

Investigate
interference :
See chapter
UL
Interferenc
e

Traffic:
UL/DL ?

DL

Investigate
Tx Power :
See chapter
Tx Power
PIs

Call Setup Performance


-Call Setup Success Rate : CSSR-

CSSR Formula is built from


two part, RRC setup and
RAB Setup phases.
RRC separation is done by
selecting only RRC setups
with reason MOC or MTC
call and corresponding
failures.
RRC setup reason is set by
mobile and there has been
noticed some
inconsistencies between
mobile vendors => this
formula might ignore or
take into notice also other
RRC activities than which
are aimed

106

NOKIA

Call Performance.PPT / 16-09-2003 /Reunanen JussiCompany

Confidential

Call Setup Performance


-Call Setup Success Rate : CSSR-

As it can be seen the


poor CSSR is almost all
due to poor RRC
Connection Setup and
Access complete
success rate

107

NOKIA

Call Performance.PPT / 16-09-2003 /Reunanen JussiCompany

Confidential

Call Setup Performance


-Call Setup Time-

Call Setup Time for AMR mobile 3G UE to PSTN


Time
58:30.7
58:31.2
58:31.4
58:31.7
58:32.0
58:32.4
58:32.6
58:32.6
58:33.0
58:33.2
58:33.9
58:34.8
58:35.1
58:35.3
58:35.3

Direction
UL CCCH
DL CCCH
UL DCCH
UL DCCH
DL DCCH
UL DCCH
DL DCCH
UL DCCH
UL DCCH
DL DCCH
DL DCCH
UL DCCH
UL DCCH
UL DCCH
DL DCCH

Message
rrcConnectionRequest
rrcConnectionSetup
rrcConnectionSetupComplete
initialDirectTransfer
downlinkDirectTransfer
uplinkDirectTransfer
securityModeCommand
securityModeComplete
uplinkDirectTransfer
downlinkDirectTransfer
radioBearerSetup
measurementReport
radioBearerSetupComplete
measurementReport
downlinkDirectTransfer (CC: Alerting)

Call Setup Delay (PDF CDF)


100

PDF
CDF

90
80
70
60
50
40
30
20
10
0
0

Average call
setup time
~4.6seconds
108

NOKIA

Call Performance.PPT / 16-09-2003 /Reunanen JussiCompany

Confidential

0 to 3000

3000 to 5000

5000 to 8000 8000 to 10000

Setup Time [ms]

> 10000

UE

RNC

Node B

CN

RRC:RRC Connection Request


NBAP:RL Setup Request
NBAP:RL Setup Response

Call Setup
Performance
-Call Setup Time-

ALCAP:ERQ
ALCAP:ECF

Call Setup Time for UDI

RRC:RRC Connection Setup


NBAP:RL Restore Indication
RRC:RRC Connection Setup Complete
RRC:Initial Direct Transfer(MM:CM Service Request)
RANAP:Initial UE Message(MM:CM Service Request)
SCCP:CC
RANAP:Direct Transfer(MM:Authentication Request)
RRC:DL Direct Transfer(MM:Authentication Request)

UDI calls are performed between 2


3G UE
The test results below show that
M-M case time is longer
especially in CN due to B party
waiting time.
Application

Time
in UE
(sec)

Time in
CN (sec)

Total
time
(sec)

MOC UDI
server

2.9

0.2

5.21

MOC TV call 2.9

5.17

10.0

MTC TV call

0.1

5.0

RRC:UL Direct Transfer(MM:Authentication Response)


RANAP:Direct Transfer(MM:Authentication Response)
RANAP:Common ID
RANAP:Security Mode Command
RRC:Security Mode Command
RRC:Security Mode Complete
RANAP:Security Mode Complete

2.9

RRC:UL Direct Transfer(CC:Setup)


RANAP:Direct Transfer(CC:Setup)
RANAP:Direct Transfer(CC:Call Proceeding)

CN delay

RRC:DL Direct Transfer(CC:Call Proceeding)


109

NOKIA

Call Performance.PPT / 16-09-2003 /Reunanen JussiCompany

Confidential

UE delay

UE

RNC

Node B

CN

RANAP:RAB Assignment Request


NBAP:Radio Link Reconfiguration Prepare
NBAP:Radio Link Reconfiguration ready
ALCAP:ERQ
ALCAP:ECF
ALCAP:ERQ
ALCAP:ECF
NBAP:Radio Link Reconfiguration Commit
RRC:Radio Bearer Setup

Call Setup
Performance
-Call Setup Time-

Call Setup Time for UDI


UDI calls are performed between 2
3G UE
The test results below show that
M-M case time is longer
especially in CN due to B party
waiting time.
Application

Time
in UE
(sec)

Time in
CN (sec)

Total
time
(sec)

MOC UDI
server

2.9

0.2

5.21

MOC TV call 2.9

5.17

10.0

MTC TV call

0.1

5.0

RRC:Radio Bearer Setup Complete


RANAP:RAB Assignment Response
RANAP:Direct Transfer(CC:Alerting)
RRC:DL Direct Transfer(CC:Alerting)
RANAP:Direct Transfer(CC:Connect)
RRC:DL Direct Transfer(CC:Connect)
RRC:UL Direct Transfer(CC:Connect Acknowledge)

2.9

RANAP:Direct Transfer(CC:Connect Acknowledge)

CN delay
110

NOKIA

Call Performance.PPT / 16-09-2003 /Reunanen JussiCompany

Confidential

UE delay

Call Setup Performance


-Call Setup Time-

Call Setup Time for UDI mobile 3G UE to stationary 3G


UE
16:29.0
16:29.0
16:30.0
16:30.0
16:31.0
16:31.0
16:31.0
16:31.0
16:31.0
16:31.0
16:32.0
16:33.0
16:34.0
16:39.0

111

NOKIA

rrcConnectionRequest
rrcConnectionSetup
rrcConnectionSetupComplete
initialDirectTransfer
downlinkDirectTransfer
uplinkDirectTransfer
securityModeCommand
securityModeComplete
uplinkDirectTransfer
downlinkDirectTransfer
radioBearerSetup
radioBearerSetupComplete
measurementControl
downlinkDirectTransfer

Call Performance.PPT / 16-09-2003 /Reunanen JussiCompany

Call Setup Delay (PDF CDF)


100
90
80
70
60
50
40
30
20
10
0

PDF
CDF

0 to 3000

3000 to
5000

5000 to
6000

6000 to
7000

Setup Time [ms]

Confidential

7000 to
8000

8000 to
10000

> 10000

Call Setup Performance


-Call Setup Time-

Call Setup Time can be decreased by adjusting the MSC parameters for
the Authentication i.e. each call is not authenticated

The calls that are not authenticated there is ~600ms reduction in call setup
time
For example following MSC/VLR parameter set for authentication means that
only every 10th Mobile Originated call is authenticated

AUTHENTICATION LOC UP NEW VIS: 1

when a new visitor makes LU authentication


is always used

LOC UP: 1

when LU is made authentication is always


used
For periodical location updates
authentication is not used
when IMSI Attach procedure is made
authentication is always used

MO CALL: 10

112

only every 10th MO call is authenticated


NOKIA

only every 10th MO SMS procedure is


authenticated

Call Performance.PPT / 16-09-2003 /Reunanen JussiCompany

For MT calls authentication is not used at all

MT SMS: 0

IMSI ATTACH: 1

MT CALL: 0

PER UP: 0

MO SMS: 10

For MT SMS procedures authentication is not


used

MT LOC REQ: 0

Authentication not used for Mobile Terminated


location request procedures

MT USSD: 0

Authentication not used for Mobile Terminated


USSD

Confidential
SS OPER: 0

Call Setup Performance


-Call Setup Time-

Call Setup Delay can be improved currently by


Using 13.6kbps SRB bitrate but then Iub capacity is
decreased as the bitrate for DCCH (SRB) is fixed to be
13.6kbps irrespective of the state of the call (call setup phase
or active phase)
The call setup time improvement is in the region of 0.75
seconds compared to 3.4 kbps SRB bit rate allocation

113

NOKIA

Call Performance.PPT / 16-09-2003 /Reunanen JussiCompany

Confidential

Call Setup Performance


-Call Setup Time-

Improvements in 1.5.2 ED + CDs (available April September)


SRB BLER optimisation
The BLER target of SRB change from 1% to 0.1% decreases
the RLC PDU retransmissions
Each RLC PDU retransmission increases set up delay by
100-200ms
Decrease the number of long set up delays caused by PDU
retransmissions (minimum delay not decreased)
Call Set up Time Optimisation
Dynamic setting of ActivationTimeOffset enables 200 to
500ms reduction for set up delay
Improvements in RAN04 available 2004 (1Q)
Dynamic 13.6 kbps SRB at RRC connection establishment
SRB 13.6 kbps usage possible without loosing the Iub
capacity

114

NOKIA

Call Performance.PPT / 16-09-2003 /Reunanen JussiCompany

Confidential

Call Setup Performance


-Call Setup Time-

With SRB 13.6kbps it can be seen that


average call set up time of <4s (3G
PSTN) can be achieved
Results are from laboratory testing
showing 83 percentile share of better
than 4seconds call set up time

In case higher percentile for call set-up


delay for e.g. 5 seconds is needed then
it is utmost important that there is only
one RRC Connection Request made as
shown in the picture on the left
Call Setup Delay vs. #RRC
Connection Request

115

NOKIA

Call Performance.PPT / 16-09-2003 /Reunanen JussiCompany

Confidential

Call Setup Performance


-Call Setup Time : PIs-

RRC Setup time: time between RRC: RRC CONNECTION REQUEST


and RRC: RRC CONNECTION SETUP COMPLETE
AVE SETUP TIME FOR RRC (M1001C221)
Average setup time for RRC
This counter, divided by the denominator, gives the
Average Setup Time for RRCs on a measurement
period
UPDATED when RRC connection is successfully set up
UNIT: 10 ms
DENOMINATOR FOR AVE SETUP TIME FOR RRC (M1001C222)
Denominator for Average setup time for RRC
UPDATED: When RRC connection is successfully set up
this counter is incremented by 1
UNIT: 10 ms

116

NOKIA

Call Performance.PPT / 16-09-2003 /Reunanen JussiCompany

Confidential

Call Setup Performance


-Call Setup Time : PIs-

RAB setup time: the time between RANAP: RAB ASSIGNMENT REQUEST and RANAP: RAB
ASSIGNMENT RESPONSE
AVE SETUP TIME FOR CS VOICE RAB (M1001C223)
Average setup time for a CS Voice RAB
This counter, divided by the denominator, gives the Average Setup Time for
CS VOICE RABs on a measurement period
UPDATED: When a RAB is successfully set up.
DENOMINATOR FOR AVE SETUP TIME FOR CS VOICE RAB (M1001C224)
Denominator for average setup time for a CS Voice RAB
UPDATED when a RAB is successfully set up, this counter is incremented by 1
AVE SETUP TIME FOR CS DATA CONV RAB (M1001C225)
Average setup time for CS DATA Conversational RAB
This counter, divided by the denominator, gives the Average Setup Time for
CS DATA CONV RABs on a measurement period
UNIT: 10 ms
DENOMINATOR FOR AVE SETUP TIME FOR CS DATA CONV RAB (M1001C226)
Average setup time for CS DATA Conversational RAB
UPDATED when a RAB is successfully set up, this counter is incremented by 1
UNIT: 10 ms
AVE SETUP TIME FOR CS DATA STREAM RAB (M1001C227)
Average setup time for CS DATA Streaming RAB
This counter, divided by the denominator, gives the Average Setup Time for
CS DATA STREAM RABs on a measurement period
UNIT: 10 ms

117

NOKIA

Call Performance.PPT / 16-09-2003 /Reunanen JussiCompany

Confidential

Call Setup Performance


-Call Setup Time : PIs-

RAB setup time: the time between RANAP: RAB ASSIGNMENT REQUEST and
RANAP: RAB ASSIGNMENT RESPONSE
DENOMINATOR FOR AVE SETUP TIME FOR CS DATA STREAM RAB
(M1001C228)
Average setup time for CS DATA Streaming RAB
UPDATED: When a RAB is successfully set up, this counter is
incremented by 1.
UNIT: 10 ms
AVE SETUP TIME FOR PS DATA CONV RAB (M1001C229)
Average setup time for PS DATA Conversational RAB
This counter, divided by the denominator, gives the Average
Setup Time for PS DATA CONV RABs on a measurement period
UPDATED: When a RAB is successfully set up
DENOMINATOR FOR AVE SETUP TIME FOR PS DATA CONV RAB
(M1001C230)
Average setup time for PS DATA Conversational RAB
UPDATED:When a RAB is successfully set up, this counter is
incremented by 1.
AVE SETUP TIME FOR PS DATA STREAM RAB (M1001C231)
Average setup time for PS DATA Streaming RAB
This counter, divided by the denominator, gives Average Setup
Time for PS DATA STREAM RABs on a measurement period
UPDATED: When a RAB is successfully set up
118

NOKIA

Call Performance.PPT / 16-09-2003 /Reunanen JussiCompany

Confidential

Call Setup Performance


-Call Setup Time : PIs-

RAB setup time: the time between RANAP: RAB ASSIGNMENT


REQUEST and RANAP: RAB ASSIGNMENT RESPONSE
DENOMINATOR FOR AVE SETUP TIME FOR PS DATA STREAM RAB
(M1001C232)
Average setup time for PS DATA Streaming RAB
UPDATED: When a RAB is successfully set up, this counter is
incremented by 1
AVE SETUP TIME FOR PS DATA INTERACTIVE RAB (M1001C233)
Average setup time for PS DATA Interactive RAB
This counter, divided by the denominator, gives the
Average Setup Time for PS DATA INTERA RABs on a
measurement period
UPDATED: When a RAB is successfully set up
UNIT: 10 ms
DENOMINATOR FOR AVE SETUP TIME FOR PS DATA INTERACTIVE
RAB (M1001C234)
Average setup time for PS DATA Interactive RAB
UPDATED: When a RAB is successfully set up, this counter is
incremented by 1
UNIT: 10 ms
119

NOKIA

Call Performance.PPT / 16-09-2003 /Reunanen JussiCompany

Confidential

Call Setup Performance


-Call Setup Time : PIs-

RAB setup time: the time between RANAP: RAB ASSIGNMENT


REQUEST and RANAP: RAB ASSIGNMENT RESPONSE
AVE SETUP TIME FOR PS DATA BACKGROUND RAB (M1001C235)
Average setup time for PS DATA Background RAB
This counter, divided by the denominator, gives the
Average Setup Time for PS DATA BACKGROUND RAB
on a measurement period
UPDATED: When a RAB is successfully set up
UNIT: 10 ms
DENOMINATOR FOR AVE SETUP TIME FOR PS DATA
BACKGROUND RAB (M1001C236)
Average setup time for PS DATA Background RAB
UPDATED: When a RAB is successfully set up, this counter
is incremented by 1
UNIT: 10 ms

120

NOKIA

Call Performance.PPT / 16-09-2003 /Reunanen JussiCompany

Confidential

Call Setup Performance


-Call Setup Time : PIs-

Call Setup Time:


For Voice = Average RRC Connection Setup time + Average RAB
Setup time for Voice Setup time
AVE SETUP TIME FOR RRC (M1001C221)
AVE SETUP TIME FOR CS VOICE RAB (M1001C223)

DENOMINATO R FOR AVE SETUP TIME FOR RRC (M1001C222) DENOMINATO R FOR AVE SETUP TIME FOR CS VOICE RAB (M1001C224)

For CS Data Conversational Class = Average RRC Connection


Setup time + Average RAB Setup time for CS data
AVE
SETUP TIME FOR RRC (M1001C221)
AVE SETUP TIME FOR CS DATA CONV RAB (M1001C225)
Conversational

DENOMINATO R FOR AVE SETUP TIME FOR RRC (M1001C222)

DENOMINATO R FOR AVE SETUP TIME FOR CS DATA CONV RAB (M1001C226)

For CS Data Streaming Class = Average RRC Connection Setup


time + Average RAB Setup time for CS data Streaming

AVE SETUP TIME FOR RRC (M1001C221)


AVE SETUP TIME FOR CS DATA STREAM RAB (M1001C227)

DENOMINATO R FOR AVE SETUP TIME FOR RRC (M1001C222) DENOMINATO R FOR AVE SETUP TIME FOR CS DATA STREAM RAB (M1001C228)

For PS Data Conversational Class = Average RRC Connection


Setup time + Average RAB Setup time for PS data
AVE
SETUP TIME FOR RRC (M1001C221)
AVE SETUP TIME FOR PS DATA CONV RAB (M1001C229)
Conversational

DENOMINATO R FOR AVE SETUP TIME FOR RRC (M1001C222)

121

NOKIA

DENOMINATO R FOR AVE SETUP TIME FOR PS DATA CONV RAB (M1001C230)

Call Performance.PPT / 16-09-2003 /Reunanen JussiCompany

Confidential

Call Setup Performance


-Call Setup Time : PIs-

Call Setup Time:


For PS Data Streaming Class = Average RRC Connection Setup
time + Average RAB Setup time for PS data Streaming
AVE SETUP TIME FOR RRC (M1001C221)
AVE SETUP TIME FOR PS DATA STREAM RAB (M1001C231)

DENOMINATO R FOR AVE SETUP TIME FOR RRC (M1001C222) DENOMINATO R FOR AVE SETUP TIME FOR PS DATA STREAM RAB (M1001C232)

For PS Data Interactive Class = Average RRC Connection Setup


time + Average RAB Setup time for PS data Interactive

AVE SETUP TIME FOR RRC (M1001C221)


AVE SETUP TIME FOR PS DATA INTERACTIVE RAB (M1001C233)

DENOMINATO R FOR AVE SETUP TIME FOR RRC (M1001C222) DENOMINATO R FOR AVE SETUP TIME FOR PS DATA INTERACTIVE RAB (M1001C234)

For PS Data Background Class = Average RRC Connection


Setup time + Average RAB Setup time for PS data Background

AVE SETUP TIME FOR RRC (M1001C221)


AVE SETUP TIME FOR PS DATA BACKGROUND RAB (M1001C235)

DENOMINATO R FOR AVE SETUP TIME FOR RRC (M1001C222) DENOMINATO R FOR AVE SETUP TIME FOR PS DATA BACKGROUND RAB (M1001C236)

122

NOKIA

Call Performance.PPT / 16-09-2003 /Reunanen JussiCompany

Confidential

Call Setup Performance


-Call Setup Time : Questions-

How to improve the Call Setup Time?


What needs to be noted from the Call Setup Time PI?
What are the current 3G PSTN, PSTN 3G and 3G 3G call setup
times?

123

NOKIA

Call Performance.PPT / 16-09-2003 /Reunanen JussiCompany

Confidential

Contents
Call Setup Performance

Idle Mode Performance

Paging

RRC Connection Establishment Performance

RAB Establishment Performance

Call Setup Failure & Time

UL Interference

Dropped Call Performance

SHO Performance (round-the-corner)

DL Power Allocation Parameters

UL Power Allocation Parameters

PS Call Performance
124

NOKIA

Call Performance.PPT / 16-09-2003 /Reunanen JussiCompany

Confidential

Call Setup Performance


-UL Interference-

By looking the service level measurements counters from the


system UL interference problems can be identified (or candidates
sites for UL interference problems)
From following statistics example problem sites can be identified

TIME

RNC_NAME WBTS_NAME WCEL_ID AVE_PRXTOT_CLASS_0 PRXTOT_DENOM_0 AVE_PRX_NOISE PRX_NOISE_DENOM_1 MAX_PRX_NOISE_VALUE MIN_PRX_NOISE_VALUE

xx/yy/2003

RNC0A

S1

11

-101.5303476

417890

-102.2032753

417890

-96.20651315

-104.0488615

xx/yy/2003

RNC0A

S2

21

-100.6359575

120433

-8.450980398

256097

-8.450980395

-8.450980399

xx/yy/2003

RNC0A

S3

31

-100.1797349

418038

-101.5176085

418038

-98.38141601

-102.7430289

xx/yy/2003

RNC0A

S4

41

-99.88490547

417646

-100.1795506

418082

-97.88041662

-102.6695255

xx/yy/2003

RNC0A

S5

51

-99.13808872

418011

-103.0792801

418011

-96.89290798

-106.3280456

xx/yy/2003

RNC0A

S6

62

-96.42019831

236918

-7.993405493

238251

-7.993405485

-7.993405494

xx/yy/2003

RNC0A

S7

72

-95.55816508

417988

-100.6589406

417988

-94.13858107

-105.7031607

xx/yy/2003

RNC0A

S8

81

-92.74359104

417909

-99.20143665

417996

-90.97058471

-105.9885019

xx/yy/2003

RNC0A

S9

82

-92.12983064

418028

-99.37448848

418028

-90.98810486

-106.1318252

xx/yy/2003

RNC0A

S10

101

-88.67772468

417997

-96.24871571

417997

-87.77917773

-104.9551382

xx/yy/2003

RNC0A

S11

111

-73.0103741

417889

-88.20890215

417912

-82.83331727

-101.3714613

125

NOKIA

Call Performance.PPT / 16-09-2003 /Reunanen JussiCompany

Confidential

Call Setup Performance


-UL Interference-

WBTS6 S6 WCEL62 UPLINK 10:10:10.0

Network statistics was found to show increased


daily Prx_Noise levels for sector S62
Call Setup problems and dropped calls under
this cell
In on-line monitoring sector 2 shows abnormal
behavior where Prx_tot is jumping up
periodically
Slight trace of the same WBTS6
behavior
also
seen
S6 WCEL61 UPLINK
10:10:10.0
sector 1 even though the level is much lower

Admission control will block all the call setup


attempts and even RRC Connection Request
during the time the UL noise rise is above the
set threshold value i.e. PrxTotal > PrxTarget +
PrxOffset and even RRC Connection Requests
are blocked when the UL noise rise is above
PrxTarget + PrxOffset
the
By setting:
the PrxTarget to maximum = 30dB
PrxOffset to maximum = 6dB
PrxNoise to autotuning with 20dB tuning window
The AC operation can be disabled however the calls cannot be setup because of the
coverage
126

NOKIA

Call Performance.PPT / 16-09-2003 /Reunanen JussiCompany

Confidential

Call Setup Performance


-UL Interference-

WBTS6 S6 WCEL62 UPLINK 10:10:10.0

The site was visited and investigated


and the following finding was made:
When the antenna line is
disconnected the phenomena
disappears.
When the antenna is reconnected
the same same behavior reappears.
This would suggest that the cause for
this behavior is rather external
interference than in the BTS.

WBTS6 S6 WCEL62 UPLINK 10:10:10.0

127

NOKIA

Call Performance.PPT / 16-09-2003 /Reunanen JussiCompany

However, the antenna installation of


this site more than suggests that this
site cannot perform normally.

Confidential

Call Setup Performance


-UL Interference-

Normal cell = UL & DL


coverage are balanced

Cell A

Cell suffering from UL


interference = DL
(CPICH) coverage
much bigger than UL
coverage

Cell B

Under the cell suffering from


excessive >>5dB noise rise due
to interference, it is extremely
difficult to initiate a call
RACH procedure might fail
due to not enough margin for
power ramp up (especially in
case the UE is having
problems with Open Loop PC)
Initial TX power for first PRACH
preamble

ptx = CPICHtransmissionPower-RSCP(CPICH)+RSSI(BS)
+"constant value
In case the cell having PrxNoise very high (and interference
conditions cannot be solved external source) and at the same
time no RRC Connection Requests received by the RNC (Counter
1001C0 is showing very low or 0 value) ->
PRACH_preamble_retrans and RACH_tx_Max could be tuned
128

NOKIA

Call Performance.PPT / 16-09-2003 /Reunanen JussiCompany

Confidential

Call Setup Performance


-UL Interference-

In Nokia RAN there are two sets of parameters related to antenna


line components to be set in order to guarantee proper operation.
Both sets of settings have to reflect the actual used BTS
configuration.
1. Feeder loss and MHA gain parameters (set in BTS
during commissioning)

129

NOKIA

BTS reports received wideband power back to the RNC. Feeder


loss and MHA gain parameters are used in calculating the value
of the received wideband power
MHA parameter is set according to the actual used MHA gain
(MHA gain is fixed)
MHA gain value can be checked from the MHA manufacturer's
data sheet
Feeder loss parameter includes feeder loss, diplexer loss and
other connector losses
If a value of 3dB for feeder loss for a cell is measured, then this
value must be entered in the BTS commissioning file
These parameter do not exist in the parameter dictionary
because they are not directly visible to RNC

Call Performance.PPT / 16-09-2003 /Reunanen JussiCompany

Confidential

Call Setup Performance


-UL Interference-

In Nokia RAN there are two sets of parameters related to antenna


line components to be set in order to guarantee proper operation.
Both sets of settings have to reflect the actual used BTS
configuration. N
2. MHA parameter and Cable loss parameter (WBTS
parameters in RNC)

These parameters are directly related


If MHA parameter set to OFF, Cable loss parameter is not used
If MHA parameter set to ON, Cable loss parameter is used
Cable loss parameter defines the difference in DL link loss in
relation to the UL link loss
Factors that contribute to UL/DL unbalance: MHA
Cable loss = Real MHA gain = Feeder loss parameter

The above mentioned parameters have to be set correctly in the


BTS and RNC. Otherwise there is a possibility that the cell will end
up with high noise

130

NOKIA

Call Performance.PPT / 16-09-2003 /Reunanen JussiCompany

Confidential

Call Setup Performance


-UL Interference-

RNC uses MHA parameters and cable loss parameter to adjust the
open loop PC for PRACH and initial UL DPCCH power
If parameter MHA used is set to yes, adjustment is done based
on the Cable Loss parameter value
If cable loss parameter is not correctly set, the initial PRACH
and UL DPCCH powers are not properly set:
If cable loss is set to less than the actual difference
between DL and UL link losses, the initial powers in PRACH
and UL DPCCH are too high
If cable loss is set to higher than the actual difference
between DL and UL link losses, the initial powers in PRACH
and UL DPCCH are too low
When considering the impact of this, one should bear in mind:
the relative inaccuracy of the open loop PC, and the fact
that the PRACH power is gradually ramped up, if no
acquisition indicator is received in AICH
UL DPCCH power (in case of first RL) is gradually
commanded (by TPCs) to be increased if no UL
syncronisation is obtained
131

NOKIA

Call Performance.PPT / 16-09-2003 /Reunanen JussiCompany

Confidential

Call Setup Performance


-UL Interference-

The impact of setting cable loss wrongly in RNC is probably not


fatal, definitely non optimum performance could be caused
In case we MHA is set yes in RNC then:
and for PRACH the cable loss is subtracted from
PtxPrimaryCPICH and result is transmitted in SIB 5/6 to the UE
for the open loop PC purposes
cable loss is used to adjust the TxCPICH in the following
formula for initial UL DPCCH:
DPCCH_Power_offset TxCPICH RxRSSI SIRDPCCH 10 log( SFDPCCH )

132

NOKIA

where TxCPICH is then PtxPrimaryCPICH - Cable Loss

Call Performance.PPT / 16-09-2003 /Reunanen JussiCompany

Confidential

Call Setup Performance


-UL Interference-

When MHA and Cable data is wrong in commissioning file,


RRC connection reject due to congestion can be observed.
Re-commissioning with correct data is needed
After original commissioning -> no calls
through

After re-commissioning => call can be


made

PrxTotal ~ -73dBm
PrxTotal ~ -105dBm
133

NOKIA

Call Performance.PPT / 16-09-2003 /Reunanen JussiCompany

Confidential

Call Setup Performance

-UL Interference PIs : PrxTotal measurements PrxTotal values are presented


according to:
BTS reports total UL interference
(RSSI) with a resolution of 0.1 dB
with the range [-112, ...,-50] dBm

RSSI_LEV _000: RSSI < -112.0 dBm

RSSI_LEV _001: -112.0 dBm <= RSSI <


-111.9 dBm

RSSI_LEV _002: -111.9 dBm <= RSSI <


-111.8 dBm

Sample 1 : Unloaded

RSSI_LEV _619: -50.2 dBm <= RSSI < -50.1


dBm

Sample 2 : Feasible load area 1

RSSI_LEV _620: -50.1 dBm <= RSSI < -50.0


dBm

RSSI_LEV _621: -50.0 dBm <= RSSI

X = collected sample of PrxTotal measurement

Sample 3 : Feasible load area 2


Sample 4 : Marginal load area
Sample 5 : Overload area
134

NOKIA

Call Performance.PPT / 16-09-2003 /Reunanen JussiCompany

The correct value in dBm can be


obtained by formula:
-112.0dBm + (RSSI_LEV)/10
Confidential

Call Setup Performance

-UL Interference PIs : PrxTotal measurements The PrxTotal measurements are divided into classes according to
following criteria:
CLASS

AREA

INCREMENTED IF

CLASS
0

Unloaded

(Lrt=<UnloadedRT) AND (Lnrt=<UnloadedNRT)

CLASS
1

Feasible_Load_Area
_1

(PrxTarget -PrxOffset >= PrxTotal ) AND


((Lrt>UnloadedRT) OR (Lnrt>UnloadedNRT))

CLASS
2

Feasible_Load_Area
_2

(PrxTarget > PrxTotal > PrxTarget -PrxOffset) AND


((Lrt>=UnloadedRT) OR (Lnrt>= UnloadedNRT))

CLASS
3

Marginal_Load_Are
a

(PrxTarget + PrxOffset > PrxTotal >=PrxTarget) AND


((Lrt>UnloadedRT) OR(Lnrt> UnloadedNRT))

CLASS
4

Overload_Area

(PrxTotal >= PrxTarget + PrxOffset) AND


((Lrt>UnloadedRT) OR (Lnrt>UnloadedNRT))

Based on the INCREMENT IF criteria the CLASS is selected and the


measurement result is assigned to the corresponding class and
the amount of samples in the class is put into the
PRXTOT_DENOM_0
135

NOKIA

Call Performance.PPT / 16-09-2003 /Reunanen JussiCompany

Confidential

Call Setup Performance

-UL Interference PIs : PrxTotal measurements As en example for CLASS 0:


th at 11pm
Measurement start time is: 2003 March 11th
and as next was same date but 10pm the
measurement/reporting period is 60min

WCELL Name

Start Time

AVE_PRXTOT
CLASS 0

PRXTOT_DENO
M_0

PLMN-PLMN/RNC2/WBTS-1627/WCEL16271

2003031123

84

17972

PLMN-PLMN/RNC2/WBTS-1627/WCEL16271

2003031122

86

17984

PLMN-PLMN/RNC2/WBTS-1627/WCEL16271

2003031121

88

WCELL Name

Start Time

AVE_PRXTOT
CLASS 0

17977

This is already average


value
PRXTOT_DENO
M_0

[dBm]
PLMN-PLMN/RNC2/WBTS-1627/WCEL16271

2003031123

-112.0 + 84/10 =
-103.6

PLMN-PLMN/RNC2003031122
-112.0 + 86/10 =
2/WBTS-1627/WCEL-103.4
16271Call Performance.PPT / 16-09-2003 /Reunanen Jussi
136
NOKIA
Company Confidential

17972

17984

This is number of
samples during
measurement period
i.e. as the period is
one hour and RRI
period is 200ms ->
3600s/0.2s = 18 000,
some samples are not
there but in general
~18 000 per one hour

Call Setup Performance

-UL Interference PIs : PrxNoise measurements MAX_PRX_NOISE_VALUE


(M1000C12)
This counter is updated if the
value of the counter PrxTotal
is smaller than the current
value of PrxNoise threshold
but Lrt<UnloadedRT and
Lnrt<UnloadedNRT
The dBm value is obtained
by dividing with -100
X = collected sample of PrxTotal measurement

MIN_PRX_NOISE_VALUE (M1000C13)
This counter is updated if the value of the counter PrxTotal is
bigger than the current value of PrxNoise threshold but
Lrt<UnloadedRT and Lnrt<UnloadedNRT
The dBm value is obtained by dividing with -100

137

NOKIA

Call Performance.PPT / 16-09-2003 /Reunanen JussiCompany

Confidential

Call Setup Performance

-UL Interference PIs : PrxNoise measurements AVE_PRX_NOISE (M1000C10)


In case the PrxTotal result is
in the unloaded area i.e.
Lrt<UnloadedRT and
Lnrt<UnloadedNRT, the
received PrxTotal value is
used to update the PrxNoise
counter
X = collected sample of PrxTotal measurement

This counter is a sum over a measurement period divided by a


denominator (averaged value)
UPDATED: In every radio resource indication period
dBm value value can be derived by multiplying with 100

138

NOKIA

Call Performance.PPT / 16-09-2003 /Reunanen JussiCompany

Confidential

Call Setup Performance

-UL Interference PIs : PtxTotal measurements PtxTotal means the total


transmitted power in a cell
The counter value is already
average dBm value and it is
multiplied by 100
PtxTotal values are presented as
follows:
X = collected sample of PtxTotal measurement

CLASS

AREA

INCREMENTED IF

CLASS
0

Unloaded

PtxTotal = PtxCPICH + PtxPCCPCH

CLASS
1

Feasible_Load_Area
_1

PtxTarget -PtxOffset > PtxTotal > (PtxCPICH +


PtxPCCPCH)

CLASS
2

Feasible_Load_Area
_2

PtxTarget > PtxTotal >= PtxTarget - PtxOffset

CLASS
3 NOKIA

Marginal_Load_Are PtxTarget + PtxOffset > PtxTotal >= PtxTarget


Call Performance.PPT / 16-09-2003 /Reunanen JussiCompany Confidential
a

139

Call Setup Performance

-UL Interference PIs : Lrt & Lnrt measurements There are specific counters for Lrt and Lnrt for each CLASS (0 4)
defined same way as for PrxTotal measurements.
Counter values are already
averaged and real % value is
obtained when the counter value
is divided by 100

When the PrxTotal value is inside


the specific CLASS range,this
counter is updated with the
estimated LRT value
At the same time as AVE
PRXTOT CLASS X counter is
X = collected sample of PrxTotal measurement
updated i.e. when Radio
Sample 1 : Unloaded
Resource Indication message
Sample 2 : Feasible load area 1
arrives

Sample 3 : Feasible load area 2


Sample 4 : Marginal load area
Sample 5 : Overload area
140

NOKIA

Call Performance.PPT / 16-09-2003 /Reunanen JussiCompany

Confidential

Call Setup Performance

-UL Interference PIs : Lrt & Lnrt measurements Lrt means RT traffic load in a cell. The value is between [0...1[
Lnrt means NRT traffic load in a cell. The value is between [0...1[
(1 i) Lown
By load we mean calculated fractional load
in a cell
(own and other cell user interference)
the uplink interference a user will cause in the cell can be
solved
P
R
R
Prx

R W

Prx _ total

rx

Prx _ total

R W

the load factor of all the users in the L


cell
own

Lown

Own _ cell _ users

141

NOKIA

Call Performance.PPT / 16-09-2003 /Reunanen JussiCompany

Prx _ own
Prx _ total

Lown

Confidential

L load factor

Call Setup Performance

-UL Interference PIs : Lrt & Lnrt measurements From obtained fractional load :
Lnrt

Lnrt

Lnrt

Lrt

N _ nrt _ active _ UL

i 1

Prx _ nrt
Prx _ total

Prx_nrt

1
W
1
i Ri

N _ nrt _ active _ UL

i 1

Lrt

1
Prx_total
W
1
i Ri

Lrt + Lnrt = Lown Lown

Lrt

N _ rt _ active _ UL

Prx _ rt
Prx _ total

i 1

Prx_rt

1
W
1
i Ri
N _ rt _ active _ UL

i 1

1
Prx_total
W
1
i Ri

Own _ cell _ users

Own cell real time user/ non-real time user load can be obtained by
multiplying Lrt/Lnrt by PrxTotal

142

NOKIA

Call Performance.PPT / 16-09-2003 /Reunanen JussiCompany

Confidential

Call Setup Performance


-Tx Power : PtxTotal PIs-

PtxTotal means the total


transmitted power in a cell
If raw counter value is
investigated (not through
EOSflex), the counter value is
already average dBm value
and it is multiplied by 100
PtxTotal values are presented as
follows:
X = collected sample of PtxTotal measurement

CLASS

AREA

INCREMENTED IF

CLASS
0

Unloaded

PtxTotal = PtxCPICH + PtxPCCPCH

CLASS
1

Feasible_Load_Area
_1

PtxTarget -PtxOffset > PtxTotal > (PtxCPICH +


PtxPCCPCH)

CLASS
2

Feasible_Load_Area
_2

PtxTarget > PtxTotal >= PtxTarget - PtxOffset

NOKIA
CLASS

Call Performance.PPT / 16-09-2003 /Reunanen


JussiCompany
Marginal_Load_Are
PtxTarget
+Confidential
PtxOffset > PtxTotal >= PtxTarget

143

Call Setup Performance


-Tx Power : Ptx_RT & Ptx_NRT PIs-

Ptx_RT means in downlink transmitted power for RT services in a cell


The value is achieved by calculating the RT traffic contribution
(traffic class conversational or streaming) to the average
transmitted code powers
The code powers are measured by the BTS and reported to the
RNC for each radio link of a cell
The value is produced in every NBAP Radio Resource Indication
period and expressed in dBm
Ptx_NRT means in downlink transmitted power for NRT services in a
cell
The value is achieved by calculating the NRT traffic contribution
(traffic class interactive or background) to the average transmitted
code powers
The code powers are measured by the BTS and reported to the
RNC for each radio link of a cell
The value is produced in every NBAP Radio Resource Indication
period and expressed in dBm

144

NOKIA

Call Performance.PPT / 16-09-2003 /Reunanen JussiCompany

Confidential

Call Setup Performance


-Tx Power : Ptx_RT & Ptx_NRT PIs-

Estimated average transmitted


power for downlink RT or NRT
users on the cell for CLASS X (0
4)
This counter is a sum over a
measurement period divided by
a denominator and it is average
value

X = collected sample of PtxTotal measurement

The counter for specific CLASS is updated at the same time as AVE
PTXTOT CLASS X counter i.e. when Radio Resource Indication
message arrives
For Example: When the PtxTotal value is inside CLASS 0 range,
the AVE PTX RT CLASS 0 counter is updated with the estimated
PTX RT value
Real dBm value is obtained when divided by 100!
CLASS 0 counters does not include situations when the BTS is not
Company Confidential
active

145

NOKIA

Call Performance.PPT / 16-09-2003 /Reunanen Jussi

Call Setup Performance


-Tx Power : Ptx_RT & Ptx_NRT PIs-

The Ptx_average values are obtained from the Radio Link


Measurements report sent by the BTS
RNC (AC/PS) is able to identify which of the Ptx_average values
belongs to which services (RT or NRT)
Based on the Radio Link ID information

Ptx, nrt

N _ nrt _ active _ UL

P
i 1

where

tx ,i

Ptx, i Ptx_average

Ptx, rt

N _ rt _ active _ UL

P
i 1

where

tx ,i

Ptx, i Ptx_average

From 25.133
Reported value
UTRAN_CODE_POWER _010
UTRAN_CODE_POWER _011
UTRAN_CODE_POWER _012

UTRAN_CODE_POWER _120
UTRAN_CODE_POWER _121
UTRAN_CODE_POWER _122
146

NOKIA

Measured quantity value


-10.0 Transmitted code power < -9.5
-9.5 Transmitted code power < -9.0
-9.0 Transmitted code power < -8.5

45.0 Transmitted code power < 45.5


45.5 Transmitted code power < 46.0
46.0 Transmitted code power < 46.5

Unit
dBm
dBm
dBm

dBm
dBm
dBm

Call Performance.PPT / 16-09-2003 /Reunanen JussiCompany

Confidential

reported-RL-Information-RL-Meas-Rep
- length (in bits): 00000
Reported-RL-InformationItem-RL-Meas-Rep
- extension flag: 0
- preamble: 00
- rL-ID: 1
- contents (in bits): 00001
measurement-1-Avail-Ind
- extension flag: 0
- choice index: 0
measurement-Available
- extension flag: 0
- average-DL-TransmittedPower: 80
- contents (in bits): 1010000
- trailing bits: 0

Contents
Call Setup Performance

Idle Mode Performance

Paging

RRC Connection Establishment Performance

RAB Establishment Performance

Call Setup Time

UL Interference

Dropped Call Performance

Loss of Synchronization DL/UL

SHO Performance

DL Power Allocation Parameters

UL Power Allocation Parameters

PS Call Performance
147

NOKIA

Call Performance.PPT / 16-09-2003 /Reunanen JussiCompany

Confidential

Dropped Call Performance


-Loss of Synchronization UL-

During the In-sync state, the BTS L1 monitors the synchronization status of
each radio frame
If N_OUTSYNC_IND (def. 20) consecutive Out-of-sync indications are received, L1
starts the timer T_RLFAILURE (def. 2seconds)
The timer is stopped and reset if N_INSYNC_IND (def. 4) consecutive In-sync
indications are received
If the timer T_RLFAILURE expires, the radio link is moved to an Out-of-sync state
When a radio link has been moved to the Out-of-sync state, L1 informs BTS
L3 about the synchronization failure
BTS L3 sends the NBAP:RADIO LINK FAILURE message to the RNC
BTS L3 repeats the NBAP:RADIO LINK FAILURE message periodically at
intervals defined by timer T_arlf (T_arlf = 5s) until L1 has re-established
the synchronization or the radio link is released by the RNC with the
NBAP:RADIO LINK DELETION message
If N_INSYNC_IND In-sync frames are received during the Out-of-sync state, L1
changes the radio link back to the In-sync state and inform BTS L3 about the reestablished synchronization
BTS L3 sends the NBAP:SYNCHRONIZATION INDICATION message to the RNC
During the Out-of-sync state, L1 keeps on searching the synchronization as long
as the synchronization has been re-established or the radio link is released by
the RNC with the NBAP:RADIO LINK DELETION message

148

NOKIA

Call Performance.PPT / 16-09-2003 /Reunanen JussiCompany

Confidential

Dropped Call Performance


-Loss of Synchronization UL-

The loss of UL synchronization can be noticed on L3Iub table


counters:
FAIL_SHO_SRNC_INI_SYN_FAIL for the initial synchronization
FAIL_SHO_SRNC_ACT_SYN_FAIL for active links
UL In-Sync/Out-of-Sync is based on the measured UL BER:
Out-of-Sync is entered if the BER goes below Q_OUT for 160ms
In-Sync is entered if the BER goes above Q_IN for 160ms
Hidden parameters:
Q_IN=20% BER (before R1.5.2 CD12 it was 12%)
Q_OUT=15%BER

149

NOKIA

Call Performance.PPT / 16-09-2003 /Reunanen JussiCompany

Confidential

Dropped Call Performance


-Loss of Synchronization DL/UL-

Radio link failure detection in DL is based on counter N313 (counting out


of sync indicator) and timer T313 in UE
In CELL_DCH State, after receiving N313 consecutive "out of sync"
indications from layer 1 for the established DPCCH physical channel in
FDD the UE
start timer T313
upon receiving N315 successive "in sync" indications from layer 1
and upon change of UE state:
stop and reset timer T313
Timer
T313
started

N_OUTSYNC_IND
out of sync
indicators

out of sync
indicators

Timer
T_RLFAILUR
E started

Timer T313
stopped and
reset

Timer
T_RLFAILUR
E stopped
and reset

N313 amount of out


of sync indicators

150

NOKIA

N_INSYNC_INDin
sync indicators

Call Performance.PPT / 16-09-2003 /Reunanen JussiCompany

in sync
indicators on
L1
N315 L1 in
sync indicators

Confidential

Dropped Call Performance


-Loss of Synchronization DL/UL-

if T313 expires:
consider it as a "Radio link failure"
N_OUTSYNC_IND
out of sync
indicators

N313 amount of
out of sync
indicators

151

NOKIA

Timer
T313
expires

Timer
T315
expires

RNC receives
RNC sends
NBAP:RADIO LINK
NBAP:RADIO
Radio Link Failure
FAILURE message
LINK
DELETION
form BTS with a cause
message
value
"Synchronization
Failure" Enter Idle mode in
case timer T302 or
Timer
Timer
T314 are not running.
T315
T314
T302 is the timer for
started
started
Cell/URA Update
When the criteria for
When the criteria for
retransmissions.
radio link failure are
radio link failure are
No refulfilled.
fulfilled.
establishment for
The timer is started
The timer is started if
RT services in
only if radio bearer(s)
radio bearer(s) that are
current release ->
that are associated
associated with T314
T314 is for Tr and
with T315 exist.
exist of if only RRC
UM bearers
Currently all the radio
connection exists.
links are associated
with T315
Not implemented in
Timer
T313
started

out of sync
indicators

Re-establishment
timer for NRT
services (AM
bearers)

Timer
T_RLFAILUR
E started

T_L1_Sync_ReEstab_W
ait timer started in
RNC
Timer
T_L1_Sync_ReEstab_
T_RLFAILURE
Wait timer expired
expires

NokiaConfidential
RAN currently

Call Performance.PPT / 16-09-2003 /Reunanen JussiCompany

Dropped Call Performance


-Loss of Synchronization DL/UL-

Periods in time where neither "in sync" nor "out of sync" is reported by
layer 1 do not affect the evaluation of the number of consecutive (resp.
successive) "in sync" or "out of sync" indications
Longer T313 timer values allow more time for the UE to recover from
possible L1 synchronisation problems
However the longer the timer value is the longer time the resources
are left hanging
Values 10 and 20 are tested and value 20 provides better dropped call
performance

152

NOKIA

Call Performance.PPT / 16-09-2003 /Reunanen JussiCompany

Confidential

Contents
Call Setup Performance

Idle Mode Performance

Paging

RRC Connection Establishment Performance

RAB Establishment Performance

Call Setup Time

UL Interference

Dropped Call Performance

Loss of Synchronization DL/UL

SHO Performance

DL Power Allocation Parameters

UL Power Allocation Parameters

PS Call Performance
155

NOKIA

Call Performance.PPT / 16-09-2003 /Reunanen JussiCompany

Confidential

Dropped Call Performance


-SHO Performance-

The SHO performance optimisation was done in


Nokia test network, Espoo, Finland in the area
where there were no coverage problems
Optimal values derived for the SHO parameters
in unloaded (no traffic) network
In the following slides the results are presented
leading to the optimal values in the table below

Parameters
Default value
Optimal Value
1.5
Addition Window (dB)
2
Addition Time (ms)
100
100
Drop Window (dB)
4
2.5
Drop Time (ms)
640
640
Replacement Window (dB)
2
2
Replace Time (ms)
100
100
MaximumActive Set Size
3
3
CPICH Ec/No filter Coefficient
3
3
Active Set Weighting Coefficient
0
0
156

NOKIA

Call Performance.PPT / 16-09-2003 /Reunanen JussiCompany

Confidential

CPICH Ec/No [dB]

Dropped Call Performance


-SHO Performance-

CPICH Ec/No Filter Coefficient:Parameters

Value

Addition Window (dB)


Addition Time (ms)
Drop Window (dB)
Drop Time (ms)
Replacement Window (dB)
Replace Time (ms)
Maximum Active Set Size
CPICH Ec/No filter Coefficient
Active Set Weighting Coefficient

3 (600ms), default
4 (800ms)
5 (1100ms)
6 (1600ms)

Number of dropped calls Average active set size


(RT)

Power (dBm)

157

NOKIA

By comparing
the results of
different
parameter sets,
the default value
3 should be the
optimal one--highest SHO
gain, good KPIs
and time
proper
Average
between SHO
UE Power (dBm)
ASoverhead.
updates (s)

Call Performance.PPT / 16-09-2003 /Reunanen JussiCompany

Confidential

0 (200ms)
2 (400ms)
3 (600ms), default

2.5
100
4
640
2
100
3
0, 1, 2, 3, 4, 5, 6
0

By comparing the
results of different
parameter sets, the
default value 3
should be the optimal
one---highest SHO
gain, good KPIs and
proper SHO
overhead.

Dropped Call Performance


-SHO Performance-

60ms
80ms
100ms, default
120ms
160ms

Parameters
Addition Window (dB)
Addition Time (ms)
Drop Window (dB)
Drop Time (ms)
Replacement Window (dB)
Replace Time (ms)
MaximumActive Set Size
CPICH Ec/No filter Coefficient
Active Set Weighting Coefficient

By looking at the No.


of dropped calls, only
80ms and 100ms
could be acceptable.
The SHO overhead
with AddTime= 80ms
is too high, so 100ms
was selected as the
best one.

Value
2.5
60, 80, 100, 120, 160
4
640
2
100
3
3
0

ower (dBm)

Parameters
Value
Addition Window (dB)
2.5
Addition Time (ms)
60
Drop Window (dB)
4
Drop Time (ms)
120, 240, 320, 640, 1280
Replacement Window (dB)
2
Replace Time (ms)
100
Maxim
Active Set
Sizeset size Average time between
r of dropped
callsumAverage
active
UE3Power (dBm)
(RT)
AS updates (s)
CPICH Ec/No filter Coefficient
3
Active Set Weighting Coefficient
0
158

NOKIA

Call Performance.PPT / 16-09-2003 /Reunanen JussiCompany

240ms
320ms
640ms, default
1280ms

Confidential

Only the value


640ms can
provide an
acceptable KPIs
by looking at the
number of
dropped calls.

Dropped Call Performance


-SHO Performance-

Parameters
Addition Window (dB)
Addition Time (ms)
Drop Window (dB)
Drop Time (ms)
Replacement Window (dB)
Replace Time (ms)
Maximum Active Set Size
CPICH Ec/No filter Coefficient
Active Set Weighting Coefficient

umber of dropped
calls

Average active set size


1.5 / 2.5
(dB)
(RT)

1.5 / 3 (dB)
1.5 / 3.5 (dB)
1.5 / 4.5 (dB)
1.5 / 7.5 (dB)

159
NOKIA
Power
(dBm)

Value
0.5, 1.5, 2.5, 3.5
100
2, 3, 4, 5
640
2
100
3
3
0

By
comparing
Average
time between
the
SHO(s)
AS
updates
overhead,
number of
drop calls and
UE Tx power,
the parameter
set 1.5/ 2.5 is
the best one.

0.5 / 2 (dB)
1.5 / 3 (dB)
2.5 / 4 (dB), default
3.5 / 5 (dB)

By considering
the SHO
overhead,
Number of
dropped calls and
UE Tx power at
the same time,
the parameter set
1.5 /3 is seen as
the best one.

UE Power (dBm)

Parameters
Addition Window (dB)
Addition Time (ms)
Drop Window (dB)
Drop Time (ms)
Replacement Window (dB)
Replace Time (ms)
MaximumActive Set Size
CPICH Ec/No filter Coefficient
Active Set Weighting Coefficient
Call Performance.PPT / 16-09-2003 /Reunanen JussiCompany Confidential

Value
1.5
100
2.5, 3, 3.5, 4.5, 7.5
640
2
100
3
3
0

Dropped Call Performance


-SHO Performance-

Parameters
Addition Window (dB)
Addition Time (ms)
Drop Window (dB)
Drop Time (ms)
Replacement Window (dB)
Replace Time (ms)
MaximumActive Set Size
CPICH Ec/No filter Coefficient
Active Set Weighting Coefficient

ropped

m)

2.5
100
3.5, 4, 4.5, 5.5
640
2
100
3
3
0

Average active set size Average time between


(RT)
AS updates (s)

1dB
2dB

160

Value

NOKIA

2.5
2.5
2.5
2.5

/
/
/
/

3.5 (dB)
4 (dB), default
4.5 (dB)
5.5 (dB)

The tested
sets did not
show better
performanc
e as the set
1.5/ 2.5.

UE Power (dBm)

The value 1 can


gives more SHO
gain (lower UE Tx
power) and better
performance (less
number of dropped
calls), but the effect
is not so significant.

Parameters
Addition Window (dB)
Addition Time (ms)
Drop Window (dB)
Drop Time (ms)
Replacement Window (dB)
Replace Time (ms)
MaximumActive Set Size
CPICH Ec/No filter Coefficient
Active Set Weighting Coefficient
Call Performance.PPT / 16-09-2003 /Reunanen JussiCompany Confidential

Value
1.5
100
3
640
1, 2
100
2
3
0

Dropped Call Performance


-SHO Performance-

Parameters
Addition Window (dB)
Addition Time (ms)
Drop Window (dB)
Drop Time (ms)
Replacement Window (dB)
Replace Time (ms)
MaximumActive Set Size
CPICH Ec/No filter Coefficient
Active Set Weighting Coefficient

er of dropped
calls

1.5
100
3
640
2
100
2, 3
3
0

Max AS size= 2
Max AS size= 3

Max AS size= 2
gives lower SHO
overhead, but the
value 3 provides
more SHO gain. No
big difference for
the other KPIs. Max
AS size= 3 was still
the better one.

Average active set size Better


Average time
between
UE Power (dBm)
performance
(RT)
AS updates (s)

100ms
640ms

Bm)
161

Value

NOKIA

with 100ms by
looking at the
number of dropped
calls. But both the
parameter have no
big affect to the SHO
gain and SHO
overhead

Parameters
Addition Window (dB)
Addition Time (ms)
Drop Window (dB)
Drop Time (ms)
Replacement Window (dB)
Replace Time (ms)
MaximumActive Set Size
CPICH Ec/No filter Coefficient
Call Performance.PPT / 16-09-2003 /Reunanen JussiCompany ConfidentialActive Set Weighting Coefficient

Value
1.5
100
3
640
1
100, 640
2
3
0

Dropped Call Performance


-SHO Performance-

Parameters
Addition Window (dB)
Addition Time (ms)
Drop Window (dB)
Drop Time (ms)
Replacement Window (dB)
Replace Time (ms)
MaximumActive Set Size
CPICH Ec/No filter Coefficient
Active Set Weighting Coefficient

Value
1.5
100
3
640
2
100
3
3
0, 1
ASWeightingCoeff= 0
ASWeightingCoeff= 1

Active Set Weighting


Coefficient is used to weight
either the measurement result of
the best active set cell (M_best) or
the sum of measurement results of
all active set cells (M_sum) when
the UE calculates the reporting
range for the events 1A and 1B.

verage active set size Average time between


(RT)
AS updates (s)
162

NOKIA

UE Power (dBm)

Call Performance.PPT / 16-09-2003 /Reunanen JussiCompany

Confidential

The value 1 can


give us lower
SHO overhead
and more gain
(lower UE Tx
power) at the
same time. But
this parameter
still need fullly
test.

Dropped Call Performance


-SHO Performance-

20 - 25

15 - 20

11 - 15

10 - 11

9 - 10

8-9

7-8

6-7

5-6

4-5

3-4

2-3

1-2

0-1

# of events

Difference in ASU period (measured as time between ASU Complete


messages) from two different environments: Case A; HongKong and
Case B;Time
Osaka
between "ASU complete" messages
Case A:
200
Average time between
150
Measurement 3
100
Measurement 2
Active Set Update is 2Measurement 1
50
3 seconds which is
0
very small compared
to the Case B
seconds

Case B:
Average time between
active set updates is 510 seconds

45
40
35
30
25
20
15
10
5
0

163

NOKIA

seconds from call setup

115 - 120

105 - 110

95 - 100

85 - 90

75 - 80

65 - 70

55 - 60

45 - 50

35 - 40

25 - 30

15 - 20

5 - 10

2-3

ASU time

0-1

# of ASUs

ASU time

Call Performance.PPT / 16-09-2003 /Reunanen JussiCompany

Confidential

It can be seen that the


smaller cells (a lot of
micro cells in Case A and
only macros in Case B)
shorter the time between
ASUs

Dropped Call Performance


-SHO Performance-

ASU period is measured as the time between Active Set Update


commands (including both soft and softer HOs)
For Case A the SHO optimisation must take the ASU period into account
as well i.e. necessarily most optimised network is not the one with the
lowest SHO overhead but combination of ASU period and SHO overhead
For Case B the SHO optimisation can concentrate on SHO overhead
optimisation with the cost of decreased ASU period

When optimising SHO check also ASU period as the shorter the ASU
period is the higher is the signaling load in RNC
The limit for ASU period is under study!

164

NOKIA

Call Performance.PPT / 16-09-2003 /Reunanen JussiCompany

Confidential

Dropped Call Performance


-SHO Performance-

ASU period can be measured based on the counters as:


Total time of measurement - (One cell in the active set for RT (M1001C0)
One cell in the active set for NRT (M1007C19))
Avg Time between AS Update
sum(successful Active Set Updates OnSHO For RT Traffic (M1007C15)
Successful Active Set Updates On SHO For NRT Traffic (M1007C32))

It should be noted that the total time of measurement also


includes the times when there are no calls so this PI value should
only be calculated for busy hour when there are a lot of calls
AS update AS update

No calls

AS update

AS size=1

165

NOKIA

AS update

AS size=2 AS size=3 AS size=2

AS update

AS size=1
AS size=1

Call Performance.PPT / 16-09-2003 /Reunanen JussiCompany

No calls

AS update AS update

Confidential

AS update

AS size=2 AS size=3 AS size=2


Total Time of Measurement

AS size=1

Dropped Call Performance


-SHO Performance-

Some of the UEs are suffering from the delayed measurement


report sending, which appears to be such that the higher the bit
rate is the more late the measurement report is sent by the UE
Measurement report is not sent even the reporting threshold
for e1a or e1c is reached
Ec/No

measurement
report sent :
AMR

measurement
report sent : PS
384

Addition window /
Replacement window

time

166

NOKIA

Ideal timing for


measurement
measurement report
report sent : PS
sending after e.g. Addition
64 or UDI (CS
time
T64)
Call Performance.PPT / 16-09-2003 /Reunanen JussiCompany Confidential

This kind of UE behavior


cannot be completely
fixed but results can be
improved by adjusting:
Addition window
Addition time
Replacement window
Replacement time

Dropped Call Performance

UE seem to have parallel processes


-SHO Performanceongoing I.e. measurement control

Time

ongoing
I.e. measurement
control
Plotted Data [ Line(6501 to 7000); Time(11:34:34
to 11:35:38)
]
and therefore there is 2sec. delay in
sending the measurement report
Monitored
Active Set
Set SC /
SC / Ec/No
Ec/No [dB]
[dB]

Message

100

80

11:35:20

Measurement
Control

60

11:35:20
40

11:35:21
20

188 / -8.5
35 / -13.5

309 / -9.5
97 / -18.5

188 / -3.5
35 / -8.5

309 / -11.5
97 / -23.5

-40

11:35:22

188 / -5.5
35 / -11.5

309 / -19.5
97 / <-24

188 / -5.5
35 / -14.5

309 / -21.5
97 / <-24

-60

11:35:23
-80

Measurement
Report e1c: SC188,
SC35, SC309, SC97

-120
CpichRscp
167

NOKIA

Sir

TargetSir

UeTxPower

HOIndication

Call Performance.PPT / 16-09-2003 /Reunanen JussiCompany

DpchRscp

Confidential

DlTransportChannelBler_1

N/A

N/A

N/A

N/A

N/A

N/A

0[A]

N/A

N/A

N/A

N/A

21[A]

N/A

N/A

21[A]

N/A

N/A

21[A]

N/A

N/A

N/A

N/A

SC309 =
-12.5 SC97
= -23.5

N/A

N/A

N/A

N/A

N/A

N/A

N/A

SC188 = -5.5
SC35 = -10.5

309 / -12.5
97 / -23.5

N/A

N/A

N/A

N/A

N/A

N/A

N/A

N/A

N/A

21[A]

N/A

N/A

N/A

N/A

N/A

N/A

N/A

Measurement
Report e1c: SC188,
SC35, SC309, SC97

188 / -5.5
35 / -10.5

-10

EcNo/ASS

-20

N/A

N/A

N/A

N/A

-5

11:35:22

-100

10

There is no
-15
response to
measurement
report e1c
-20
messages by the
network as the first
measurement -25
report is sent too
late by the UE and
-30
the call drops
CpichEcNO

ActiveSetSize

Dropped Call Performance


-SHO Performance-

Some of the UEs are suffering from the low PS performance (in terms of throughput)
While the bearer bit rate increases the throughput performance during SHO
performance -> the only way to fix this is to reduce the max. active set size for
PS data to e.g. 2

Poor
performa
nce due
to
coverage

168

NOKIA

High BLER, high


variance with DL SIR
and a lot of SHO
activity

Call Performance.PPT / 16-09-2003 /Reunanen JussiCompany

Confidential

Low variance
of SIR,
ASS=2,
stable BLER
and
throughput

Dropped Call Performance


-SHO Performance: UL Interference-

WBTS6 S6 WCEL62 UPLINK 10:10:10.0

Normal cell = UL & DL


coverage are balanced

Cell A

Cell suffering from UL


interference = DL
(CPICH) coverage much
bigger than UL coverage

Cell B

Cell B is having higher PrxNoise


level than Cell A which is causing
the UL and DL coverage to be very
much different.
It should be noted that even tough the AC is disabled by high enough
PrxTarget and PrxOffset parameters or PrxNoise autotuning feature
and calls can be set up under this cell, successful SHOs become
practically impossible between this kind of cell and normally
operating cell
This is due to that the UE Tx power is controlled by the original
cell (Cell A) and therefore there is not enough power from UE to
overcome the UL Noise rise in Cell A
169

NOKIA

Call Performance.PPT / 16-09-2003 /Reunanen JussiCompany

Confidential

Dropped Call Performance


-SHO Performance: UL Interference-

BTS

UE

RNC

Normal cell = UL
& DL coverage
are balanced

RRC: Measurement Report (e1a / e1c)

Cell suffering from


UL interference = DL
(CPICH) coverage
much bigger than UL
coverage

Decision to set up new RL

NBAP: Radio Link Setup Request

NBAP: Synchronization
Indication

Start TX/RX
UL Synchronisation Procedure starts

Cell A

NBAP: Radio Link Setup Response


ALCAP:ERQ
ALCAP:ECF
RRC: Active Set Update
UE Synchronisation Procedure starts: According to
Synchronisation Procedure B

RRC: Active Set Update Complete

Active Set
Update: 1a,
1c to add cell
B

NOKIA

Call Performance.PPT / 16-09-2003 /Reunanen JussiCompany

DL

UL & DL
Ec/No

UL

Addition window /
Replacement window

As the UE Tx power is not enough for


Cell B synchronisation, the SHO can
fail
170

Cell B

Addition/
Replacement
time

Confidential

UE -> RNC :
Measurement report :
e1a / e1c

distance
Active Set
Update
complete

Dropped Call Performance


-SHO Performance: UL Interference-

Normal cell = UL &


DL coverage are
balanced

Cell suffering from UL


interference = DL
(CPICH) coverage much
bigger than UL coverage

Cell A

Cell B
Less than N_INSYNC_IND
indicators on L1

Measureme
nt report
e1A: CellB

RNC sends RRC: Active


Set Update - message
to UE -> Starts timer
T_RRC_Resp_DCH

T313 expires -> UE


considers Radio Link
Failure for Cell B ->
UE drops the Cell B
from AS
UE starts T313 for
Cell B after
receiving N313
out-of-sync for Cell
B

BTS stops UL
Synchronization
procedure and Tx

L1 Synchronization not
established BTS does NOT
send NBAP:
SYNCHRONIZATION
IDICATION -message

BTS UL Synchronization
SHO Decision
After receiving
RNC sends
procedure continues until L1
algorithm Cell
ASU UE Starts DL
NBAP:RADIO
LINK
B -> Setup RL
Synchronization is achieved
Synchronization
DELETION

-> NBAP: Radio


or NBAP: RL Deletion
procedure -> and
message
Link Setup
sends ASU
message is received
Request (Chip
Complete
Offset)
Successful resource allocation in Cell B BTS ->
RNC does not receive
BTS Starts Tx (DPCH) -> BTS Starts UL
synchronization indication
-> timer
Synchronization
procedure
sends
NBAP:
171
NOKIA
Call
Performance.PPTand
/ 16-09-2003
/Reunanen
JussiCompany Confidential

Counter: RL DEL
ON SRNC DUE TO
INI SYN FAIL
(M1005C67) updated

Dropped Call Performance


-SHO Performance: UL Interference-

The ways to fix this case are:

172

Use CPICH Tx power in the BTS to balance the


Normal cell = UL & DL
UL and DL coverage i.e. decrease the CPICH Tx
coverage are balanced
power by the UL noise rise
+ Can be done cell by cell basis
+ Relatively easy to implement algorithm for
Cell A
autotuning
- Coverage is reduced also in the cases
when there is no UL interference
Active Set
Update: 1a,
- Requires autotuning feature otherwise cell
1c to add
UL & DLcell B
coverage is reduced in DL all the time
Use cell individual offset to lower the CPICH
Ec/No
Ec/No by the noise rise value
Addition window /
+ Can be done cell by cell basis
Replacement window
+ Does not affect to the coverage
- Needs some algorithm to tune the cell
Addition/
individual offset according to existing noise Replacemen
t time
rise
UE -> RNC :
- The UEs might have problem with too big
Measurement
report : e1a / e1c
difference between actual CPICH powers
also DL power drifting might cause
problems

NOKIA

Call Performance.PPT / 16-09-2003 /Reunanen JussiCompany

Confidential

Cell suffering from UL


interference = DL
(CPICH) coverage much
bigger than UL
coverage

Cell B
L1 synchronization
not achieved
between new cell
and UE, monitored
by the BTS

DL
UL
DL

Drop
window

distance
Active Set
Update
complete

RNC
sends
Active Set
Update :
e1b to
remove
the cell A

Dropped Call Performance


-SHO Performance: UL Interference-

The ways to fix this cases are:

173

Increase the Drop window setting so that UL


Normal cell = UL & DL
synchronization is achieved to the target cell
coverage are balanced
- New drop window is used for each neighbor
cell definition
- Needs some algorithm to tune the drop
window according to existing noise rise and Cell A
cell by cell basis somehow
+ Easy to test to see if situation can be
improved
Active Set
Update: 1a,
- The UEs might have problem with too big
1c to add
difference between actual CPICH powers
UL & DLcell B
also DL power drifting might cause problems
Change the drop timer to be long enough so that
Ec/No
UL synchronization is achieved to the target cell
Addition window /
- New drop time is used for each neighbor cell
Replacement window
definition
- Needs some algorithm to tune the drop time
Addition/
according to existing noise rise and cell by
Replacemen
t time
cell basis somehow
- Easy to test to see if situation can be
UE -> RNC :
Measurement
improved
report : e1a / e1c
- The UEs might have problem with too big
difference between actual CPICH powers
also DL power drifting might cause problems

NOKIA

Call Performance.PPT / 16-09-2003 /Reunanen JussiCompany

Confidential

Cell suffering from UL


interference = DL
(CPICH) coverage much
bigger than UL
coverage

Cell B
L1 synchronization
not achieved
between new cell
and UE, monitored
by the BTS

DL
UL
DL

Drop
window

distance
Active Set
Update
complete

RNC
sends
Active Set
Update :
e1b to
remove
the cell A

Dropped Call Performance

-SHO Performance : Cell Individual Offsets In case there is so called round


the corner effect between two
cells it might be tempting to
use cell individual offsets to
make the SHO to happen
earlier and therefore have the
connection on without dropped
calls

Cell A

Cell B

Without cell individual offsets

UL & DL

UL & DL
Ec/No

Cell A

NOKIA

route

Addition window /
Replacement
window

Cell B

174

The call drops due to


too rapid CPICH
coverage
degradation for Cell
A, and therefore
there is not enough
time for SHO

UE -> RNC :
Measurement report :
Call Performance.PPT / 16-09-2003 /Reunanen JussiCompany Confidential
e1a / e1c

Addition/
Replacement
time

Dropped Call Performance

-SHO Performance : Cell Individual Offsets By using cell individual offset this problem might be
overcome
However now we have different problem i.e. better DL
coverage than UL coverage
Better is to try to tune time to trigger (i.e. Addition /
Replacement time need to be shortened)

Cell A

Cell B

With cell individual offsets

See next case

Active Set
Update: 1a,
1c to add cell
B

UL & DL
Ec/No

UL
Cell individual
offset

Cell A
Addition window /
Replacement
window

175

NOKIA

UE -> RNC :
Measurement report :
e1a / e1c
Call Performance.PPT / 16-09-2003 /Reunanen JussiCompany Confidential

Cell B

DL

route
Active Set
Update
complete

Addition/
Replacement
time

Dropped Call Performance


-DL Coverage < UL Coverage-

In case there is a cell with


lower CPICH power than the
surrounding cells are having
then this cell is having too
good UL performance as
this cells UL cannot be used
efficiently due to SHO is
decided upon DL (CPICH
Ec/No)
To improve this and include
the cell into AS earlier than
CPICH Eke/No would allow, it
is possible to use cell
individual offsets for this
purpose

176

NOKIA

Normal cell = UL & DL


coverage are balanced

Cell having lower DL


coverage = DL (CPICH)
power is lower than
adjacent cells

Cell A

UL & DL
Ec/No

Cell B
Active Set
Update: 1a,
1c to add cell
B
Cell B starts to
suffer from UL
interference due
to UE getting
closer to Cell B
wo SHO

Addition window /
Replacement
window

UE -> RNC :
Measurement report :
Call Performance.PPT / 16-09-2003 /Reunanen JussiCompany Confidential
e1a / e1c

UL
DL

distance
Active Set
Update
complete
Addition/
Replacement
time

Dropped Call Performance


-DL Coverage < UL Coverage-

If Cell B is having cell


individual offset set
according to the difference
of CPICH powers between
Cell A and Cell B, the SHO
can happen correct time
and Cell B can be utilized
fully
However currently the
added cells initial Tx power
is set according to the
i.e.
existing cell
MAX R
TFCTFCS
In case
TFTFC
ILcell
the
best
Ptx, init
p ( c, Ptx_total ),
W
Ec/No is known
ILP ( c, Ptx_total)

177

NOKIA

1
Ptx, CPICH Ptx_total.
c

Normal cell = UL & DL


coverage are balanced

Cell having lower DL


coverage = DL (CPICH)
power is lower than
adjacent cells

Cell A

Cell B
Active Set
Update: 1a,
1c to add cell
B

UL

UL & DL

DL

Ec/No

Addition window /
Replacement
window

UE -> RNC :
Measurement report :
Call Performance.PPT / 16-09-2003 /Reunanen JussiCompany Confidential
e1a / e1c

distance
Active Set
Update
complete
Addition/
Replacement
time

Dropped Call Performance


-DL Coverage < UL Coverage-

However currently the


added cells initial Tx power
is set according to the
existing cell i.e.
In case the best cell
Ec/No
not known
Ptx ,initial ,new
Ptx ,ave,is
highest Ptx ,CPICH , new Ptx ,CPICH , highest
It should be noted that
neither of the formulas
include information about
the cell individual offset and
due to that the SHO to cell,
which is having cell
individual offset (high
value), might have problems
Synchronization in DL
direction difficult
178

NOKIA

Normal cell = UL & DL


coverage are balanced

Cell having lower DL


coverage = DL (CPICH)
power is lower than
adjacent cells

Cell A

Cell B
Active Set
Update: 1a,
1c to add cell
B

UL & DL
Ec/No

Addition window /
Replacement
window

UE -> RNC :
Measurement report :
Call Performance.PPT / 16-09-2003 /Reunanen JussiCompany Confidential
e1a / e1c

UL
Cell individual
offset
Tx power for the
Cell B DPCH is not
enough -> DL
synch not
achieved -> SHO
failure

distance
Active Set
Update
complete
Addition/
Replacement
time

DL

Dropped Call Performance

Received message RRC: Measurement Report (e1A).


Triggers for RT:
M1007C10 CELL ADDITION REQUEST ON SHO FOR RT
TRAFFIC
Received message RRC: Measurement Report (e1B).
Triggers for RT:
BTS
RNC
UE
M1007C11 CELL DELETION REQUEST ON SHO FOR RT
TRAFFIC
RRC: Measurement Report (e1a / e1c)
Received message RRC: Measurement Report (e1C).
Triggers for RT:
Decision to
M1007C12 CELL REPLACEMENT REQUEST ON SHO
set up new RL
FOR RT TRAFFIC
Received message RRC: Measurement Report (e1A).
NBAP: Radio Link Setup Request for Soft Handover
Triggers for NRT:
M1007C27 CELL ADDITION REQUEST ON SHO FOR NRT
NBAP: Radio Link Addition Request for Softer Handover
TRAFFIC
Received message RRC: Measurement Report (e1B).
Start TX/RX
Triggers for NRT:
M1007C28 CELL DELETION REQUEST ON SHO FOR
NBAP: Radio Link Setup Response
NRT TRAFFIC
NBAP: Radio Link Addition Response
Received message RRC: Measurement Report (e1C).
Triggers for NRT:
ALCAP:ERQ
For each
cell already
M1007C29 CELL REPLACEMENT
REQUEST
ON SHO in AS
ALCAP:ECF
FOR NRT TRAFFIC
the above counters are

-SHO Performance : PIs-

RRC: Active Set Update


B

RRC: Active Set Update Complete


A

179

NOKIA

Call Performance.PPT / 16-09-2003 /Reunanen JussiCompany

Confidential

updated i.e. in this case 2


cells

At first (UE in point A) AS


includes 2 cells
Then UE moves to point B
and reports e1A to add
blue cell

Dropped Call Performance

Number of DCH requests for Diversity HO (SHO) for


voice calls in SRNC
M1002C16 RT DCH DHO REQ FOR CS VOICE CALL IN
SRNC
Number of DCH requests for a CS Voice Call rejected by
BTS
RNC
the SRNC for reasons caused by radio resources in the
UE
target cell of diversity handover
RRC: Measurement Report (e1a / e1c)
M1002C17 RT DCH DHO REQ FOR CS VOICE CALL
REJECT IN SRNC
Decision to set up new RL Number of DCH requests for a transparent CS Data Call
with conversational class due to diversity handover in
NBAP: Radio Link Setup Request
the SRNC
M1002C58 RT DCH DHO REQ FOR CS DATA CALL CONV
UL Synchronisation Procedure starts CLASS IN SRNC
Start TX/RX
Number of DCH requests for a transparent CS Data Call
(on SRNC side) rejected for reasons caused by radio
NBAP: Radio Link Setup Response
resources in the target cell of diversity handover
ALCAP:ERQ
M1002C59 RT DCH DHO REQ FOR CS DATA CALL CONV
CLASS REJECT IN SRNC
ALCAP:ECF
SRNC A total number of DCH requests for a
RRC: Active Set Update
nontransparent CS Data Call with streaming class due
to diversity handover in the SRNC
UE Synchronisation Procedure starts:
M1002C60 RT DCH DHO REQ FOR CS DATA CALL
ccording to Synchronization Procedure B
STREAM CLASS IN
Number of DCH requests for a nontransparent CS Data
RRC: Active Set Update Complete
Call with streaming class (SRNC side) rejected for
reasons caused by radio resources in the target cell of
diversity handover
In case BTS responses back to RNC by
M1002C61RT DCH DHO REQ FOR CS DATA CALL
NBAP: RL Setup Response/RL Addition
STREAM CLASS REJECT IN SRNC
Response
In case
RNC sends to BTS NBAP: RL Setup
SAME
FORof
DRNC!
M1005C6: RL SETUP SUCC FOR SHO ON
Request/RL Addition Request
SRNC
M1005C1: RL SETUP ATT FOR SHO ON SRNC
M1005C8:RL SETUP SUCC FOR SHO ON
M1005C3:RL SETUP ATT FOR SHO ON DRNC
DRNC
M1005C42: RL BRANCH ADD ATT FOR SHO ON SRNC
M1005C44: RL BRANCH ADD SUCC FOR SHO
M1005C43: RL BRANCH ADD ATT FOR SHO ON DRNC
ON SRNC

-SHO Performance: PIs-

NBAP: Synchronization Indication

M1005C45: RL BRANCH ADD SUCC FOR SHO


ON
180SRNC
NOKIA
Call Performance.PPT / 16-09-2003 /Reunanen JussiCompany Confidential

Dropped Call Performance


-SHO Performance : PIs-

BTS

UE

RNC

RRC: Measurement Report (e1a / e1c)


Decision to
set up new RL
NBAP: Radio Link Setup Request
Start TX/RX
NBAP: Radio Link Setup Response
ALCAP:ERQ
ALCAP:ECF
RRC: Active Set Update
RRC: Active Set Update Complete
RRC: Active Set Update Failure

181

NOKIA

Call Performance.PPT / 16-09-2003 /Reunanen JussiCompany

Received message RRC: Active Set Update Complete.


Triggers:
M1007C15 SUCCESSFUL ACTIVE SET UPDATES ON SHO
FOR RT TRAFFIC
Received message RRC: Active Set Update Complete.
Triggers:
M1007C32 SUCCESSFUL ACTIVE SET UPDATES ON SHO
FOR NRT TRAFFIC
Received message RRC: Active Set Update Failure.
Triggers:
M1007C16 UNSUCCESSFUL ACTIVE SET UPDATES ON
SHO FOR RT TRAFFIC
Received message RRC: Active Set Update Failure.
Triggers:
M1007C33 UNSUCCESSFUL ACTIVE SET UPDATES ON
SHO FOR NRT TRAFFIC

For each cell in the AS the above


counter(s) are updated (including the
cell to be deleted from AS in case of
e1B)

Confidential

Dropped Call Performance


-SHO Performance : PIs-

UE is in Spot A i.e.
not in soft handover
-> AS size = 1

2
1

UE in Spot B sends
Measurement
report indicating
e1A to add cell 2
-> M1007C10 CELL
ADDITION
REQUEST ON SHO
FOR RT TRAFFIC is
updated in cell 1

UE

BTS

RNC

RRC: Measurement Report (e1a / e1c)

Decision to set up new R


NBAP: Radio Link Setup Request

UL Synchronisation
Start TX/RX
Procedure starts
NBAP: Radio Link Setup Respons
ALCAP:ERQ
ALCAP:ECF

RNC Makes AC/RM for the SHO updates


M1002C16 RT DCH DHO REQ FOR CS VOICE
CALL IN SRNC counter in cell 2
RNC sends NBAP: RL Setup Request to BTS BTS
receives the NBAP; RL Setup
updates M1005C1: RL SETUP ATT FOR SHORequest, if RL Setup Is successful,
starts TX/RX and UL Synchronization
ON SRNC counter in cell 2
procedure and sends NBAP: RL Setup
Response to RNC
182

NOKIA

Call Performance.PPT / 16-09-2003 /Reunanen JussiCompany

Confidential

Dropped Call Performance


-SHO Performance : PIsUE

RNC Sends RRC: Active Set Update


message to the UE

BTS

RNC

RRC: Measurement Report (e1a / e1c)

Decision to set up new R


NBAP: Radio Link Setup Request

UL Synchronisation
Start TX/RX
Procedure starts
NBAP: Radio Link Setup Respons
ALCAP:ERQ
ALCAP:ECF
RRC: Active Set Update

UE Synchronisation Procedure starts:


After receiving RRC: Active Set Update from According
to Synchronization Procedure B
RNC the UE checks the Update message and
RRC: Active Set Update Complete
if ok UE starts Synchronization Procedure B
and and sends RRC: Active Set Update
Complete message without waiting for theWhen receiving RRC: Active Set Update
completion of the Physical Layer
Complete, the RNC updates M1007C15
synchronization procedure B (according toSUCCESSFUL ACTIVE SET UPDATES ON
25.214/25.331)
SHO FOR RT TRAFFIC or M1007C32

183

NOKIA

SUCCESSFUL ACTIVE SET UPDATES ON


SHO FOR NRT TRAFFIC counter. Counter
Call Performance.PPT / 16-09-2003 /Reunanen JussiCompany Confidential
updated for cell 1.

Dropped Call Performance


-SHO Performance : PIs-

The Active Set Update success rate can be calculated as:


Active Set Update Success Ratio ( RT ) 100 *

Active Set Update Success Ratio( NRT ) 100 *

sum( Successful Active Set Updates On SHO For RT Traffic ( M 1007C15))


%
sum( Successful Active Set Updates On SHO For RT Traffic ( M 1007C15)
Unsuccessful Active Set Updates On SHO For RT Traffic ( M 1007C16))
sum( Successful Active Set Updates On SHO For NRT Traffic ( M 1007C 32))
%
sum( Successful Active Set Updates On SHO For NRT Traffic ( M 1007C 32)
Unsuccessful Active Set Updates On SHO For NRT Traffic ( M 1007C 33))
UE

Formula does not take into account


the cases where measurement
report is sent several times until
RNC responds with RRC: Active Set
Update -> Counters1007C16 and
1007C33 are updated only in case
UE responds to Active Set Update
message with Active Set Update
Failure message.
184

NOKIA

Call Performance.PPT / 16-09-2003 /Reunanen JussiCompany

BTS

RNC

RRC: Measurement Report (e1a / e1c)


Decision to set up new RL
NBAP: Radio Link Setup Request
UL Synchronisation
Start TX/RX
Procedure starts
NBAP: Radio Link Setup Response

ALCAP:ERQ
ALCAP:ECF
RRC: Active Set Update

UE Synchronisation Procedure starts:


According to Synchronization Procedure B

Confidential

RRC: Active Set Update Complete

Dropped Call Performance


-SHO Performance : PIs-

Therefore also the amount of


periodical measurement
reports must be checked i.e. :
CELL ADDITION FAILURE ON
SHO FOR RT TRAFFIC
M1007C13
CELL REPLACEMENT
FAILURE ON SHO FOR RT
TRAFFIC M1007C14
CELL ADDITION FAILURE ON
SHO FOR NRT TRAFFIC
M1007C30
CELL REPLACEMENT
FAILURE ON SHO FOR NRT
TRAFFIC M1007C31
It should be noted that the
counters M1007C13,
M1007C14, M1007C30 and
M1007C31 are all updated for

185

NOKIA

Call Performance.PPT / 16-09-2003 /Reunanen JussiCompany

Confidential

Dropped Call Performance


-SHO Performance : PIs-

In case also RL Setup Procedure and AC decision are needed to be included


then the SHO Success Rate could be counted as:

(M1007C15)/(M1007C10 + M1007C11 + M1007C12) for RT traffic


(M1007C32)/(M1007C27 + M1007C28 + M1007C29) for NRT traffic

SUCCESSFUL ACTIVE SET UPDATES ON SHO FOR RT TRAFFIC (M1007C15)


divided by ADDITION/DELETION/REPLACEMENT REQUEST ON SHO FOR RT
TRAFFIC (M1007C10 + M1007C11 + M1007C12)
Then in case the SHO Success Rate formula 1 and 2 are giving different results
the problem spot can be identified by analysing:

Iub Radio Link Set up & Addition success rate (either combined as below or
separately)

Radio Link Set Up / Addition Success ( SHO)


M1005C42 M1005C43 M1005C1 M1005C3
100 *
%
M1005C44 M1005C45 M1005C6 M1005C8

186

UE

Call Performance.PPT / 16-09-2003 /Reunanen JussiCompany

RNC

RRC: Measurement Report (e1a / e1c)


Decision to set up new RL
NBAP: Radio Link Setup Request

And separate failure reasons can be


identified based on M1005C16
M1005C21, M1005C28 M1005C33,
M1005C46 M1005C57 and M1005C58
M1005C65

NOKIA

BTS

UL Synchronisation
Start TX/RX
Procedure starts
NBAP: Radio Link Setup Response
ALCAP:ERQ
ALCAP:ECF
RRC: Active Set Update
UE Synchronisation Procedure starts:
According to Synchronization Procedure B

Confidential

RRC: Active Set Update Complete

Dropped Call Performance


-SHO Performance : PIs-

Then in case the SHO Success Rate formula 1 and 2 are giving
different results the problem spot can be identified by analysing:

DCH allocation successrate for SHO purposes

DCH Allocation Success ( SHO)

M 1002C17 M 1002C 59 M 1002C 61 M 1002C103 M 1002C105 M 1002C107 M 1002C109


100%
M1002C16 M1002C58 M1002C60 M1002C102 M1002C104 M1002C106 M1002C108

Where nominator is stating the


number of DCH allocation requests in
the target cell for Diversity
Handover (SHO) for RT traffic classes
(Speech+Conversational CS data +
Streaming CS Data) and NRT traffic
classes (Conversational PS data +
Streaming PS data + Interactive PS
data + Background PS data)
Where denominator is the allocation
rejections due to lack of radio
resources in target cell
187

NOKIA

Call Performance.PPT / 16-09-2003 /Reunanen JussiCompany

UE

BTS

RNC

RRC: Measurement Report (e1a / e1c)


Decision to set up new RL
NBAP: Radio Link Setup Request
UL Synchronisation
Start TX/RX
Procedure starts
NBAP: Radio Link Setup Response
ALCAP:ERQ
ALCAP:ECF
RRC: Active Set Update
UE Synchronisation Procedure starts:
According to Synchronization Procedure B

Confidential

RRC: Active Set Update Complete

Dropped Call Performance


-SHO Performance and AC-

When an event 1A or 1C is received, the radio load conditions in


the candidate cell are evaluated by Admission Control
Failure to add or replace a link can occur for the same reasons as
for an initial RRC Connection or RAB Setup (except that the UL
interference is not to checked as the UE transmission is already
included in UL noise rise)
DL Transmit Power
Transmission

214

NOKIA

Call Performance.PPT / 16-09-2003 /Reunanen JussiCompany

Confidential

Dropped Call Performance


-SHO Performance : Questions-

In case ASU period needs to be taken into account as well then


what would be the optimal set?
What are the difficulties in the SHO failure KPIs? And what is your
recommendation for SHO failure PI?
What would be your suggestion to handle the indoor <-> outdoor
SHO assuming that the indoor solution is:
AIR with 13dBm TX power
DAS with 43dBm TX power

215

NOKIA

Call Performance.PPT / 16-09-2003 /Reunanen JussiCompany

Confidential

Contents
Call Setup Performance

Idle Mode Performance

Paging

RRC Connection Establishment Performance

RAB Establishment Performance

Call Setup Time

UL Interference

Dropped Call Performance

Loss of Synchronization DL/UL

SHO Performance

DL Power Allocation Parameters

UL Power Allocation Parameters

PS Call Performance
216

NOKIA

Call Performance.PPT / 16-09-2003 /Reunanen JussiCompany

Confidential

Dropped Call Performance


-Power per Connection40

20

-40

Ec
217

NOKIA

EcIo

Sir

Call Performance.PPT / 16-09-2003 /Reunanen JussiCompany

90

1.

80

-60

15:45:16.00

15:45:16.00

15:45:15.00

15:45:15.00

15:45:14.00

15:45:13.00

15:45:12.00

15:45:11.00

15:45:10.00

15:45:10.00

15:45:09.00

15:45:08.00

e1b, 1 active set cell


dropped (become 1-way
active)

2.

70
60
50

DLTrchBLER
increase rapidly

3.

40
30

-80

This case can be improved by


allowing more power window in DL
for the PC to operate

Call drop

-20

As the SIR getting lower and lower


until the call drops this suggests
that the DL power has reached its
maximum

100

15:45:08.00

Observations from the case on the right:


1. SIR cannot be maintained at the
correct level
2. Ec/No is ok
3. BLER Increases to 100%
4. Target SIR increases at the same
time as BLER increases to 100%

4.

20

-100

10

-120

TargetSir
Confidential

ActiveSetSize

DlTrChBler

Dropped Call Performance


-Power per Connection-

DL power per connection can be increased by changing the parameter


CPICHtoRefRabOffset
the max DL TX power is determined by AC as:
Ptx ,max MIN ( RI max,eff Ptx ,ref , N dpch Ptx _ DPCH _ max )

DL tx pwr per
Connection [dBm]
/ BLER [%]

RI max,eff

Ptx ,CPICH
C pichTo Re fRabOffset

Max
power

Min
power

Min
power
time

time

Call drop
NOKIA

Ptx , ref

DL tx pwr per
Connection [dBm]
/ BLER [%]

Max
power

218

RI max
ref RI ref

Call Performance.PPT / 16-09-2003 /Reunanen JussiCompany

Confidential

Dropped Call Performance


-Power per Connection-

When increasing the max power per connection (putting


CPICHtoRefRabOffset parameter lower) it should be noted that the
AC limit might be reached quite fast:
CPICHtoRefRabOffset = 0dB
128kbps service with Eb/No = 4.5dB
Reference service AMR 12.2kbps with Eb/No = 8dB
PtxCPICH = 33dBm
Ptx_DPCH_max = -3dB (=40dBm = 10W)
=> Ptx,ref = 33dBm (0dB) = 33dBm
=> Rimax,eff = (4.5dB*128kbps)/(8dB*12.2kbps) = 4.6865
=> Ptx_Max = 9.35W
In case the PtxTarget is set to 40dBm then in case there is even one
more connection up and running the possibility that the PtxTarget
value is exceeded during the new 128kbps connection admission
control procedure and therefore it is denied

Therefore CPICHtoRefRABOffset should be tuned with PtxCPICH and


PtxTarget
219

NOKIA

Call Performance.PPT / 16-09-2003 /Reunanen JussiCompany

Confidential

Dropped Call Performance


-Power per Connection-

Another point to take into account is that the when the maximum
power per connection is increased also the minimum power limit
is increased (as the min power is max power PCrangeDL (15dB))
In case the max power increment is a lot (~3dB) then the
minimum power is increased by 3dB as well which can lead to the
minimum power problems (BTS sending too much power to the
UEs close to the BTS and therefore causing problems to the UE
and even dropped call)

Therefore the PCrangeDL parameter should be tuned according to


the CPICHtoRefRabOffset parameter tuning (from the default)

220

NOKIA

Call Performance.PPT / 16-09-2003 /Reunanen JussiCompany

Confidential

Contents
Call Setup Performance

Idle Mode Performance

Paging

RRC Connection Establishment Performance

RAB Establishment Performance

Call Setup Time

UL Interference

Dropped Call Performance

Loss of Synchronization DL/UL

SHO Performance

DL Power Allocation Parameters

UL Power Allocation Parameters

PS Call Performance
221

NOKIA

Call Performance.PPT / 16-09-2003 /Reunanen JussiCompany

Confidential

Contents
Call Setup Performance

Idle Mode Performance

Paging

RRC Connection Establishment Performance

RAB Establishment Performance

Call Setup Time

UL Interference

Dropped Call Performance

Loss of Synchronization DL/UL

SHO Performance

DL Power Allocation Parameters

UL Power Allocation Parameters

PS Call Performance
222

NOKIA

Call Performance.PPT / 16-09-2003 /Reunanen JussiCompany

Confidential

PS Call Performance
Test case was performed with RAN 1.5.2 software level
DL Power Control was enabled in the mobile terminal
Kenny SW version P 1.2-12.55
64 UL/64 DL kbps bearer services were used as per the following
profiles
QOS PROFILE INDEX & BLER target

123; 0.63%

123;1%

123; 5%

123; 20%

QOS PROFILE NAME

BLER 0.63

BLER 1

BLER 5

BLER 20

TRAFFIC CLASS

DELIVERY ORDER

DELIVERY OF ERRONEOUS SDU

ND

ND

ND

MAXIMUM SDU SIZE

1500

1500

1500

1500

MAXIMUM BIT RATE FOR DOWNLINK

64kbps

64

64

64

MAXIMUM BIT RATE FOR UPLINK

64kbps

64

64

64

RESIDUAL BER

9*

7*

7*

1*

SDU ERROR RATIO

7**

4**

4**

1**

TRANSFER DELAY
GUARANTEED BIT RATE FOR UPLINK
GUARANTEED BIT RATE FOR DOWNLINK
TRAFFIC HANDLING PRIORITY

223

NOKIA

Call Performance.PPT / 16-09-2003 /Reunanen JussiCompany

Confidential

PS Call Performance
BLER versus avg. RTT for different ping sizes using 64 kbps
bearer
900

RTT in milli seconds

800
700
600
32 bytes

500

512 bytes

400

1472 bytes

300
200
100
0
0

10

15
BLER %

224

NOKIA

Call Performance.PPT / 16-09-2003 /Reunanen JussiCompany

Confidential

20

25

PS Call Performance
Avg. Round Trip Times for different ping sizes at various BLER using
64kbps bearer
900
800

Delay in milli seconds

700
600
0.63% BLER

500

1% BLER
5% BLER

400

20% BLER
300
200
100
0
0

200

400

600

800

1000

1200

Ping Size in bytes


225

NOKIA

Call Performance.PPT / 16-09-2003 /Reunanen JussiCompany

Confidential

1400

1600

PS Call Performance
BLER versus Average FTP Upload Throughputs for various file sizes
using 64kbps bearer
60

Throughput in Kbit/s

50
40

16KB
64KB

30

500KB
1024KB

20
10
0
0

10

15
BLER %

226

NOKIA

Call Performance.PPT / 16-09-2003 /Reunanen JussiCompany

Confidential

20

25

PS Call Performance
FTP Upload Throughputs at different BLER & file sizes using
64kbps bearer
60

Throughput in kbit/s

50
40

0.63% BLER
1% BLER

30

5% BLER
20% BLER

20
10
0
0

200

400

600

800

File sizes in kilo bytes

227

NOKIA

Call Performance.PPT / 16-09-2003 /Reunanen JussiCompany

Confidential

1000

1200

PS Call Performance
Standard deviations in the Round Trip Times
(milliseconds) for each ping size at a certain BLER
levelPing Size
32 bytes
512 bytes
1472 bytes
BLER
%

228

NOKIA

0.63%

11ms

22ms

37ms

1%

25ms

36ms

42ms

5%

86ms

98ms

114ms

20%

172ms

214ms

144ms

Call Performance.PPT / 16-09-2003 /Reunanen JussiCompany

Confidential

PS Call Performance
0.63 % BLER
1MB
UL #
1
2
3

229

1 % BLER

0.63 % BLER

1MB

Total bytes Re-xmt bytes Re-xmt %


1048404
0
0,0 %
1048404
0
0,0 %
1048404
0
0,0 %

UL #

1 % BLER

1
2
3

Total bytes Re-xmt bytes Re-xmt %


1048404
0
0,0 %
1048404
0
0,0 %
1048404
0
0,0 %

DL5#% BLER
Total bytes Re-xmt bytes Re-xmt %
1
1048404
0
0,0 %
1MB
5
%
BLER
2
1048404
0
0,0 %
3
1048404
0
0,0 %
UL #
Total bytes Re-xmt bytes Re-xmt %
1
1048404
0
0,0 %
2
1048404
0
0,0 %
3
1048404
0
0,0 %

20# % BLER
DL
Total bytes Re-xmt bytes Re-xmt %
1
1048404
0
0,0 %
1MB
20
%
BLER
2
1048404
0
0,0 %
3
1048404
0
0,0 %
UL #
Total bytes Re-xmt bytes Re-xmt %
1
1048404
67160
6,4 %
2
1048404
67238
6,4 %
3
1048404
52560
5,0 %

DL #

DL #

NOKIA

Total bytes Re-xmt bytes Re-xmt %


1
1048404
0
0,0 %
2
1048404
0
0,0 %
3
1048404
0
0,0 %

Call Performance.PPT / 16-09-2003 /Reunanen JussiCompany

Confidential

Total bytes Re-xmt bytes Re-xmt %


1
1048404
0
0,0 %
2
1048404
0
0,0 %
3
1048404
0
0,0 %

PS Call Performance
-Conclusions-

A low BLER provides a much better PS performance


Lower RTT
Higher throughput
On the other hand, a low BLER can reduce the PS capacity
Higher transmit power
Higher interference
In the short term, low BLER is an acceptable solution
In the middle-long term, BLER could be tuned to balance QoS and
capacity

230

NOKIA

Call Performance.PPT / 16-09-2003 /Reunanen JussiCompany

Confidential

References
Jphone project experiences NWP Team in Japan
H3G project experiences NWP Team in HK & UK
Measurements in Nokia WCDMA Test Network, RAN1.5.2 SW,
February-March 2003: Soft Handover Optimised Performance Huibin Lin, Marja-Leena Lahti
3G Packet Data Performance Measurements Analysis: BLER test
cases - Anand Shah, Karl Tigerstedt, Octavio Garcia, Baris Sarer
SERVICE LEVEL MEASUREMENT in WCDMA RAN 1.5 and REPORTS
in OSS Reporter Juho Pirskanen & Tapio Toppari
RNC Performance Counters - DN99581291 Issue 2-4 en
All SFSs

231

NOKIA

Call Performance.PPT / 16-09-2003 /Reunanen JussiCompany

Confidential

Вам также может понравиться