1. wcdma rno paging problem analysis guidance 20041101-a-1.0

Product name Confidentiality level
WCDMA RNP For internal use only
Product version
Total 29 pages
V100R001
WCDMA RNO Paging Problem
Analysis Guidance
(For internal use only)
Prepared by: URNP -SANA Date: 2003-12-27
Reviewed by: Date:
Reviewed by: Date:
Approved by: Date:
Huawei Technologies Co., Ltd.
All rights reserved

WCDMA RNO Paging Problem Analysis Guidance For Internal Use Only
2004-07-09 Confidential Page 2 of 29
Revision Record
Date Revision
version
Description Author
2003-12-27 1.00 Initial transmittal Jiao Anqiang

Table of Contents
1Overview.................................................................................................................................................... 7
2Paging Problem Analysis Process.............................................................................................................. 9
2.1Problem Analysis Flow......................................................................................................................... 9
2.2Network Information Collection ......................................................................................................... 10
2.2.1Traffic Statistics............................................................................................................................ 10
2.2.2Alarm............................................................................................................................................ 13
2.2.3Subscriber Complaints.................................................................................................................. 13
2.2.4History Record of Network Planning Optimization ..................................................................... 14
2.2.5Radio Parameter Configuration .................................................................................................... 14
2.3Determining Optimization Goal.......................................................................................................... 15
2.4Locating Paging Problems .................................................................................................................. 15
2.4.1Determining Basic Location Direction......................................................................................... 15
2.4.2Direct Causes of Paging Loss ....................................................................................................... 16
2.4.3Further Analysis on Paging Loss Cause ....................................................................................... 17
2.4.4Other Cause Analysis.................................................................................................................... 17
2.5Paging Problems Optimization............................................................................................................ 17
2.6Optimization Verification ................................................................................................................... 17
3Typical Paging Problem Analysis............................................................................................................ 18
3.1Too Large Paging Area Planned ......................................................................................................... 18
3.1.1Problem Analysis.......................................................................................................................... 18
3.1.2Optimization Measures................................................................................................................. 19
3.2Improper Setting of CN Paging retransmission Times and Interval ................................................... 19
3.3Improper Setting of UTRAN Paging Retransmission Times and Interval.......................................... 21
3.4CN Adopts Whole-Network Paging.................................................................................................... 21
3.5Improper Setting of DRX Paging Cycle Coefficient........................................................................... 22
3.6Improper setting of Np value ............................................................................................................... 23
3.7UE Identifier Utilized by CN Paging .................................................................................................. 24
3.8UTRAN Should Activate IMSI ATTACH and DETACH Functions ................................................. 25
3.9Too Low Power allocation of Paging Channels.................................................................................. 26
3.10 Existence of Coverage Dead Zones.............................................................................................. 27
3.10.1 Problem Analysis ................................................................................................................... 27
3.10.2 Optimization Measures........................................................................................................... 27
3.11 UE Performance Problems............................................................................................................ 27
3.11.1 Problem Analysis ................................................................................................................... 27
3.11.2 Optimization Measures........................................................................................................... 27
4Outstanding Problems.............................................................................................................................. 27

List of Figures
Figure 1 Typical called flow of UE ...................................................................................................... 8
Figure 2 Paging Problem Analysis Flow.......................................................................................... 10
Figure 3 System Message 1 Resolution.......................................................................................... 26

List of Tables
Table 1 UMSC Paging Traffic Statistics Indexes................................................................................ 12
Table 2 SGSN Paging Traffic Statistics .............................................................................................. 12
Table 3 Subscriber complaint information.......................................................................................... 14
Table 4 Calculaiton Result of Paging Area When Cn Id Adopting IMSI ........................................... 19
Table 5 IMSI ATTACH and DETACH Identifier............................................................................... 25

WCDMA RNO Paging Problem Analysis Guidance
Key words: Paging, paging area, paging retransmission
Abstract: This document describes the general flow of solving paging problems, analyzes typical
problems during paging in details and provides optimization measures.
List of abbreviations:
Abbreviations Full spelling
DRX Discontinuous Reception
LA Location Area
PCH Paging Channel
PI Paging Indication
PICH Page Indication Channel
RA Route Area
RAN Radio Access Network
RNC Radio Network Controller
RNO Radio Network Optimization
WCDMA Wideband Code Division Multiple
Access

1 Overview
If it is necessary for network side to contact with UE which is in idle mode or in
CELL_PCH or URA_PCH state, a paging flow will be initiated. Paging is an
important way for network to contact with UE. Compared with other flows, the paging
flow has the characteristics of high frequency, heavy traffic and strong outburst in
radio network. Paging performance is closely related to the performances of the
whole radio network. Therefore, the study of paging problems is of practical
significance for radio network performances.
From the aspect that UE receives paging messages, paging messages consist
of PAGING TYPE1 and PAGING TYPE2. Which type of paging messages will be
sent to UE is decided by UTRAN. PAGING TYPE1 is used to page UEs in IDLE,
CELL_PCH and URA_PCH states through PCCH logical channel. PAGING TYPE2
is used to page UEs in CELL_FACH or CELL_DCH states through DCCH. This
document mainly discusses PAGING TYPE1, while PAGING TYPE2 can be
regarded as the common RRC signaling to process and will not be discussed in this
document.
Paging will be initiated on network side in the following cases:
 When UE is called: In order to set up a call, the core network (CN) sends
paging message to UTRAN through Iu interface, and UTRAN sends the
paging message from CN to UE through the paging process at Uu interface,
which will have UE initiate a signaling connection setup process with CN.
 Cell system message update: When system messages change, UTRAN will
trigger paging process in order to inform UE in idle, CELL_PCH and
URA_PCH state to carry out the system message update. Therefore, UE
can read the updated system message. For UE in CELL_FACH state, in
order to inform it to carry out corresponding system update, it is necessary
to send SYSTEM INFORMATION CHANGE INDICATION message to it
through BCCH (V1.2 version does not support the message, so UE in
CELL_FACH state will not be processed).
 UE state migration: In order to trigger UE in CELL_PCH and URA_PCH
state to carry out state migration (for example, migration to CELL_FACH
state), UTRAN will perform a paging process. Meanwhile, UE will initiate a
cell update or URA update, as a reply to the paging.
A typical called flow caused by paging is shown in Figure 1 : CN sends a paging
message to UTRAN, UTRAN calculates the paging time and obtains the target cell

after receiving the paging message, and delivers the paging message at the air
interface upon the paging time. If CN receives the paging response message from
UE, the UE paging is successful. The whole process includes paging delivery and
UE access. For UE access, refer to Access Process Guidance. There are probably
various problems in the paging process, which will cause that the target UE cannot
receive the paging message properly. For instance, during short messages
multi-casting and full-text paging in network, unreasonable paging strategies will
result in paging channel congestion, so large amount of messages will be lost, in
more serious case, the system will be caused over-loaded for long time and the
paging channel power allocation is too low, which will lead to low paging success
rate. This document will make a further discussion on these problems leading to
paging abnormality and present corresponding solutions.
UE NAS UE AS NSS MSC
paging
AUTHENTICATION REQUEST
AUTHENTICATION RESPONSE
RR_SECURITY_CONTROL_REQ
(IK CK)
Security mode control
SETUP
CALL CONFIRM
ALERT
CONNECT
CONNECT ACKNOWLEDGE
RAB setup process
paging
RR_EST_REQ (PAGING RESPONSE)
RR_PAING_IND
INITIAL_DIRECT_TRANSFER
(PAGING RESPONSE)
RANAPRANAP
RRC setup process
Figure 1 Typical called flow of UE
The general structure of this document is as follows:
1. Overview: It briefly introduces the paging time, paging flow and general
paging problems;
2. Paging problem analysis process: It describes the problem analysis flow of
paging, how to analyze paging problems step by step and what to do in each step.

3. Typical problem analysis of paging: It analyzes typical problems of paging,
including the phenomenon, principles and optimization methods of the problems.
4. Outstanding problems: They refer to problems which cannot be solved in this
document.
2 Paging Problem Analysis Process
2.1Problem Analysis Flow
The paging problem analysis flow is as shown in Figure 2 . Similar to the
general analysis methods of network problems, there are four steps of paging
problem analysis, which are detailed as follows:
Network information collection: Collect traffic statistics information related to
network and paging, alarm information, complaints of subscribers, network planning
information, optimization records and configuration of network parameters;
Determination of optimization goal: Determine KPI of paging problem
optimization;
Paging problems location: Locate causes of paging problems
Paging problems optimization: Take corresponding optimizing and adjusting
measures according to location results;
Optimization verification: Verify whether the KPI meets the requirements and
whether other information about paging is normal after optimization.

Collect network infomration
Identify optimization goal
Locate paging problem
Optimize paging problem
Optimization verification
Optimization goal achieved?
END
YES
NO
Figure 2 Paging Problem Analysis Flow
2.2Network Information Collection
Network information collection is the first step of paging problem analysis. The
optimization personnel should obtain the paging-related information to-be-optimized,
such as traffic statistics, subscriber complaints, alarms, optimization history record of
network planning and radio parameters configuration, so as to prepare for the
subsequent analysis.
2.2.1 Traffic Statistics
The paging-related traffic statistics can be observed respectively on the traffic
statistics console of RNC, UMSC and SGSN according to different paging areas.
RNC traffic statistics corresponds to a RNC area, UMSC traffic statistics
corresponds to a location area and SGSN traffic statistics corresponds to a RA. In
the actual traffic statistics analysis process, the traffic statistics of CN is mainly
analyzed, with combination of the traffic statistics data of RNC and CN. If there is no
paging-related traffic statistics on the traffic statistics console, it is necessary for
equipment users to establish the paging traffic statistics tasks. The paging-related
traffic statistics of RNC, UMSC and SGSN are discussed below based on the paging
traffic statistics of our company.

The traffic statistics of RNC is shown in Table 1 . Attention should be paid to
CN_PAGE_IDLE_UE_SUCC_RATE (paging success rate of UE in idle state initiated
by CN) and UTRAN_PAGE1_SUCC_RATE (The success rate of Paging Type 1
initiated by UTRAN). These two indexes indicate the page success rate of the
corresponding paging area of RNC. CN_PAGE_IDLE_UE_SUCC_RATE indicates
the page success rate from the aspect of CN, while UTRAN_PAGE1_SUCC_RATE
includes not only CN paging, but also UTRAN system information update and UE
state migration (the paging information of UE is cell updating). The two indexes can
be used to analyze the paging performance of an RNC area which includes one or
more location areas.
Table 1 RNC Paging Traffic Statistics Index
Name Description Standard measuring point
CN_PAGE_REQ
Count paging times at
IU interface
Receive the PAGING messages
initiated by CN
CN_PAGE_IDLE_UE_REQ
Count paging times at
IU interface of idle
subscribers
Receive the PAGING messages
initiated by CN when UE is in idle
status
CN_PAGE_IDLE_UE_SUCC
Count paging
success times of idle
subscribers
Receive RRC connection request
message of UE and the reason
belongs to the called type, such as
“Terminating Conversational Call”
UTRAN_PAGE1_REQ
Count times of initiating
PAGING TYPE 1
initiated from UTRAN
side
PAGING TYPE 1 message initiated
from UTRAN side
UTRAN_PAGE1_SUCC
Count PAGING
TYPE 1 message
initiated from UTRAN
side and the times of
UE success receive
response received
UE paging response message
received at UTRAN side .
CN_PAGE_IDLE_UE_SUCC_
RATE
Count paging success
rate of idle UE initiated
by CN
The calculation index, obtained by
the formula of
[CN_PAGE_IDLE_UE_SUCC]/
[CN_PAGE_IDLE_UE_REQ]
UTRAN_PAGE1_SUCC_RAT
E
Count the paging
success rate of Page
Type 1 initiated by
UTRAN
The calculation index, obtained by
the formula of
[UTRAN_PAGE1_SUCC]/[UTRAN_P
AGE1_REQ]
The relevant traffic statistics of UMSC paging are all based on a location area,
as shown in Table 1. Generally, the location area will not be configured cross RNC
and BSC, so we can carry out statistics on the paging success rate, the first paging
success rate and non-first paging success rate of a location area. The location area
paging success rate focuses on the paging status of a location area, not on paging

repeat times, while first paging success rate and non-first paging success rate focus
on the influence of paging repeat times on paging success rate.
ion timestransmisspagingFirst
messagesresponse-nodeliveredofnumberion timestransmisspagingFirst
ratesuccesspagingLA
－

ion timestransmisspagingFirst
timesresponsepagingFirst
ratesuccesspagingFirst 
timesresponserepitetionpagingInterface
timesrepitetionpagingInterface
ratesuccesspagingfirst-Non 
Table 1 UMSC Paging Traffic Statistics Indexes
Name Description Standard measure point
The first paging times
The first sending times of
Paging Req by MSC
Carry out statistics when MSC sends
PAGING message to RNC/BSC
The first paging
response times
Successfully receiving
response times after the first
sending of Paging message
by MSC
Carry out statistics when MSC
receives PAGING RESPONSE after
the first sending of Paging message
Interface paging repeat
times
The non-first paging times of
MSC
PAGING message to RNC/BSC
Interface paging repeat
response times
Successfully receiving
response times after the
non-first sending of Paging
message by MSC
Carry out statistics when MSC
receives PAGING RESPONSE after
the non-first sending Paging
message
The first paging times of
Iu interface
The first sending Paging
times of Iu interface
PAGING message to RNC
Lu interface paging
repeat times
The non-first sending Paging
times of Iu interface
PAGING message to RNC not at the
first time
Delivery paging
no-response times
Times of having no paging
response messages received
Paging timer stay-over statistics
The relevant traffic statistics indexes of SGSN paging are all based on a RA, as
shown in Table 2, the paging success rate of a certain RA can be obtained.
questTimesingRAGroupPag
TimesingFailureRAGroupPagquesTimesingRAGroupPag
ccessRateRAPagingSu
Re
Re －
＝
Table 2 SGSN Paging Traffic Statistics
Name Description Standard measure point
Group paging request
times of each RA
It provides delivery group
paging request times in the
SGSN sends Iu interface paging
request message (PAGING), and its

specific RA, not including
retransmission messages
CN Domain is PS.
Group paging failure
times of each RA
It provides group paging failure
times in the specific RA
Paging retransmission message times
reach maximum.
During traffic statistics analysis, attention should be paid to “location area
paging success rate” and “ RA paging success rate”, which are KPI measuring
paging performance. The RNC paging traffic statistics
CN_PAGE_IDLE_UE_SUCC_RATE and UTRAN_PAGE1_SUCC_RATE can be
taken as reference analysis.
2.2.2 Alarm
Based on the current realization in our company, CN is only “RNC overload”
alarm related to paging, alarm occurs when CN receives the overload message of
RNC. In order to guarantee the system operation stability and avoid the strike on
system by outburst messages, RNC flow-controls the paging-included message with
high processing frequency. When RNC receives paging messages from CN, the
judging system will discard the paging message and record the quantity of paging
message loss if it is in paging flow control state. If the loss proportion has reached a
certain threshold, RNC will send Overload message to CN, then CN will reduce the
message flow by certain step. If CN has not received Overload message within a
certain period, the message flow of IU will increase gradually till it get normal.
The paging-related alarm of RNC is “the flow control alarm”. When RNC is in
paging flow control state, the paging message will be lost unconditionally. When the
CPU occupation or the message packet occupation of a board sub-system exceeds
the threshold, the system will be switched into the flow control state, and the flow
control alarm will occur. When the system returns to normal operation state from the
flow control state, the flow control recovery alarm will occur. However, it is not
absolute that paging will be lost when “flow control alarm” occurs, because RNC
owns different flow control thresholds for different flow control objects (paging, serial
port printing and message tracing), as long as there is flow control, alarm will occur;
but paging loss will occur only when paging loss occurs.
2.2.3 Subscriber Complaints
If paging messages are lost, UE will not be called. The calling party will hear the
system prompt tone “the subscriber you dialed is out of service”. We can learn about
the subscriber complaints from 1860 Customer Service Center of pilot office or learn
about the cases of “out of service” by contacting with the complainer directly.

Attention should be paid to the cases of UE called failure, and information to be
collected is shown in Table 3, but not limited to it 【5】:
Table 3 Subscriber complaint information
Time
Name
(complainer)
UE No.
(complainer)
UE No.
(the
called)
UE
type
(the
called)
Place
Whether
it is
often
Description
Sort out complaint information, seek for rules and observe whether the
information belongs to the following cases:
 Whether complaints occurs when paging is busy in the daytime and idle at
night; if paging failure occurs at the peak traffic, paging congestion analysis
should be emphasized on; if not at the peak traffic, other factors should be
analyzed.
 Whether types of the called UE are the same. Problems probably exist in
UE itself.
 Whether the complaint place is centralized; it is possibly because of signal
coverage;
Find the rules and solve problems more quickly.
2.2.4 History Record of Network Planning Optimization
Obtain the network planning report, and pay attention to the division of the
paging area (location area and RA). Obtain network planning report: If it is the
optimization before network is put into service, the planning report can be obtained
from the planning manager responsible for the project; if it is the optimization after
network is put into service, the planning report can be obtained from the client. The
planning report mentioned above should include planning reports after each network
capacity expansion.
For the network that has been put into service, it may have experienced several
optimization processes before this optimization. Then we should obtain the history
optimization records and learn about the network adjusting process and outstanding
problems. Emphasis should be put on optimization records about coverage absence,
system overload, paging loss and low paging channel powerallocation.
2.2.5 Radio Parameter Configuration
Parameters related to paging are as follows, which should be collected before
optimization:

 CN paging retransmission times and paging interval;
 UTRAN paging retransmission times and paging interval;
 DRX paging cycle coefficient k (DRX paging cycle = 2k
);

Paging indication quantity NP included in a PICH frame;
 Channel power allocation of PICH and PCH;
 Whether CN adopts global paging;
 UE ID used by CN paging (IMSI,TMSI or PTMSI)
2.3Determining Optimization Goal
Two KPI of “location area paging success rate” and “RA paging success rate”
should meet the optimization requirements. The paging success rate is
recommended to be over 86% 【3】.
2.4Locating Paging Problems
2.4.1 Determining Basic Location Direction
Paging problem optimization is intended to guarantee the paging KPI. Whether
UE can respond successfully is directly related to the paging KPI. From this aspect,
paging problems can be generally divided into three parts:
 Paging message is not delivered at the air interface at all. If paging
message is not delivered at the air interface at all, it is most likely that
paging is lost. Paging loss is a frequently-seen problem during paging
process and it is the emphasis of analysis in this document. For detailed
paging loss analysis, please refer to Section 2.4.2 and Section 2.4.3.
However, the reason why paging has not been sent out is probably the
transmission faults of IUB interface or faults of other equipments, which can
be seen through viewing the alarm information.
 Paging message has been delivered, but UE does not receive the message
or receives the wrong message. According to the specific subscriber
complaints, if problems exist on called UE only (prompt tone “the subscriber
you dialed is out of service”), because the reason is probably that the power
allocation of PICH and PCH is too low or there are problems with UE
performance; if problems exist in calling and called UE, the reason is
probably that signal coverage dead zones exist in the area.

 UE fails in responding after receiving paging messages. As shown in Figure
1,The problem is related to access failure, for solutions, please refer to
WCDMA RNP Access Process Problem Analysis Guidance.
During paging problem location, problems can be located by analyzing paging
traffic statistics, alarms and subscriber complaints. Paging loss generally occurs in
the period when traffic is heavy. Check “location area paging success rate” and “RA
paging success rate” of CN traffic statistics, if these two indexes are low (lower than
86%) when traffic is heavy, and subscriber complaints are centralized in this period,
paging loss is serious and attention should be paid to analyzing it. Meanwhile, check
whether RNC overload alarm or RNC flow control alarm occurs in CN. If these
alarms exist, the probability of paging loss is big.
If events that two indexes are low are evenly distributed on the time, and
subscriber complaints are regional, check other reasons except “paging loss”:
paging channel power allocation, signal coverage and UE performance.
It is necessary to make dialing test analysis on the site for further analysis on
paging problems. Select the period when traffic is heavy and the area where
subscriber complaints are centralized, trace paging message delivery and UE paging
response process on the CN and RNC maintenance consoles during the dialing test
process.
2.4.2 Direct Causes of Paging Loss
Paging will be lost in the following cases:
1. RNC system is in paging flow control state. Because paging message is
sent frequently and RNC flow-controls paging. If RNC detects that CPU
occupation or message queuing occupation reaches the preset
threshold, paging flow control will be triggered and paging message will
be lost unconditionally.
2. PCH capacity restriction; Based on the current PCH coding mode, one
TTI can only transmit 240bits. If IMSI paging is utilized, only three UEs
can be paged at the same paging period; and if TMSI and PTMSI
paging are utilized, only five UEs can be paged at the same paging
period【2】. If the quantity of UEs paged at the same paging period is
beyond the system processing capability, paging loss will occur.

3. Other causes: Transmission faults of IUB interface and equipments
faults. Probability of those faults is small and can be viewed from the
alarm console.
2.4.3 Further Analysis on Paging Loss Cause
The direct causes of paging loss are system overload and paging channel
overload, after further analysis of its causes, it is found that CN and RNC have
adopted inappropriate paging strategies, for detailed analysis, please refer to
hyperlinks as follows:
 Too large paging area planned
 Improper setting of CN paging retransmission times and interval
 Improper setting of UTRAN paging retransmission times and interval
 CN adopts whole-network paging
 Improper setting of DRX paging cycle coefficient
 Improper setting of Np value
 CN adopts improper UE identifier
2.4.4 Other Cause Analysis
 Too low power allocation of paging channels
 Existence of coverage dead zones
 UE performance problems
2.5Paging Problems Optimization
The paging problem optimization methods are described in detail in Section 3.
2.6Optimization Verification
After network is optimized and adjusted, the optimization results should be
verified. The frequently-used verifying methods are available as follows: have it run
and check relevant paging traffic statistics, check whether there are paging alarms,
collect subscriber complaint information, and carry out UE called dialing test on site.
 Traffic statistics: Check mainly whether “location area paging success rate”
and “RA paging success rate” achieve the optimization goal 86% specified
in advance;

 Alarm: Check whether there is any “RNC overload” alarm in CN or whether
there is any flow control alarm in RNC;
 Subscriber complaint: Whether there is any subscriber complaint in a
certain period;
 Dialing test: Test UE called success rate when traffic is at the peak and in
the place where subscriber complaint occurs. During dialing test, it is
unnecessary to through-connect UE, just listen to ring back tone and
prompt tone.
3 Typical Paging Problem Analysis
3.1Too Large Paging Area Planned
3.1.1 Problem Analysis
In general, CN pages target UE in one paging area (location area or RA), which
is called intra-office paging. For CS domain service, CN identifies and pages UE with
location area. In protocols, location area is defined as the area where mobile
terminals can move freely without updating VLR. A location area may contain one or
more cells. For PS domain service, CN identifies and pages UE with routing area.
RA is defined as the area where mobile terminals can move freely without updating
SGSN in a specified operation mode. A RA may contain one or more cells. The
relationship between location area and RA is specified in GSM, namely, RA can be
equal to location area in size, or only the subset of a certain location area.
If paging area is planned too large, the same paging message of a network
paging mobile station may be sent in multi cells, which will overload the paging
channel as well as increasing signaling flow at the Iub interface. If the paging
channel of a cell is overloaded in a certain period, the paging message sent to UE in
the cell will be caused lost and the power-on subscriber within the service area can
not be paged (The subscriber is out of service).
On the contrary, if paging area is planned too small, subscribers will update
locations frequently while moving, which will increase signaling flow of the system.
When network is just put into service, PS paging traffic is not heavy, RA may be
planned larger, it can be planned as per n=1. With the development of network, the
demands for PS paging traffic are increasing, in this case RA can be planned
smaller consequently. However, too small RA may cause more paging area update
events of subscribers while moving, and that signaling overhead on network side
becomes larger. Moreover, frequent location update will influence the standby time
of UEs.

The paging area maximum value is determined by the paging channel capacity.
For estimation methods of paging area capacity, please refer to【2】, for paging
mixture capacity in typical environments in【2】, please refer to Table 4:
Table 4 Calculaiton Result of Paging Area When Cn Id Adopting IMSI
n
Mmix
RA LA
1 3 1.738624042 0.002592996 133 133
2 3 1.741614489 0.003616996 95 190
3 3 1.74665366 0.004640996 74 222
4 3 1.744911078 0.005664996 61 244
5 3 1.745096566 0.006688996 51 255
6 3 1.745372494 0.007712996 44 264
7 3 1.745680526 0.008736996 39 273
8 3 1.745680526 0.009760996 35 280
9 3 1.746014764 0.010784996 32 288
10 3 1.746024614 0.011808996 29 290
Meaning of each field in the table:
n is the proportion of location area size to RA size, LA=nRA;
Mmix is the UEs paged quantity by each TTI of PCH;
H  is traffic supported by a specified paging module in a DRX cycle length
period;
is the identification of the supported by a specified
paging module in an RA during a DRX cycle length period.
For instance, n=2, LA=190 means 190 cells in a location area. If the actual
location area of network is larger than 190, it is necessary to divide the location area.
It is recommended by GSM network optimization experiences that the location
area capacity of each local network is larger than 300 carrier frequencies and the
location area with paging traffic exceeding 50,000 in busy hour is divided.
3.1.2 Optimization Measures
Divide location areas where network capacity or paging traffic is bigger than a
specified threshold, which can reduce paging message flow efficiently.
3.2Improper Setting of CN Paging retransmission Times and Interval
In order to guarantee the paging success rate, CN will retransmission paging
message at the IU interface. CN paging retransmission times and interval can be
H  RAPerDRXCycleLengthInNErlang
RAPerDRXCycleLengthInNErlang

configured. Because paging is initiated specific to location area, paging
retransmission will increase paging traffic. Especially in case of public downlink
channel congestion at the air interface, paging traffic will double due to paging
repeat, which will waste downlink channel resources seriously. CN paging
retransmission configuration should be in accordance with UTRAN.
CN paging retransmission is the retransmission paging in case that UE has not
responded at the first time, which can increase paging success rate and connection
rate. But from data in GSM network, paging retransmission success rate is low,
especially at the second time or even more, which contributes a little to paging
success rate and connection rate. Experience data indicates that: Generally, the
success rate of the first paging is approaching to 86%, and the retransmission
paging rate is 1.8%. Herein, retransmission paging success is mainly the second
paging success, while success rate of the third paging is estimated less than 0.2%
calculated by the decreasing pattern, and the third paging (the second repeat) has
little influence on paging success rate【6】. The paging mode of WCDMA is similar
with that of GSM, from online statistics, frequent CN paging contributes little to the
paging success rate; contrarily, it increases the system load.
It is no good for CN paging interval to be too short. CN sends paging messages
to RNC through the IU interface, and RNC calculates the paging time according to
IMSI attached by the paging message and arranges a delivery at the corresponding
paging time of the nearest paging cycle. Currently, RNC of our company also adopts
the paging retransmission mode, the default is repeating once and the
retransmission interval is a paging cycle. From analysis above, the maximum interval
between the time that RNC receives CN paging and the time that message is
delivered to the air interface is a paging cycle. If RNC retransmits once, CN paging
interval should be longer than two paging cycles. The paging cycle realized by our
company is 2.56s. If CN paging interval is bigger than 2.56s and less than 5.12,
when CN retransmitted paging reaches, and RNC has not finished repeat paging,
RNC will arrange paging in the following paging cycle, Thus only three paging
messages are sent in fact, not four expected, which causes paging message loss
CN paging retransmission configuration should be in accordance with UTRAN.
When UTRAN retransmits paging once, it is suggested for CN configuration to
retransmit once (totally twice) at the interval longer than two paging cycles. If the
paging cycle is 2.56s, it is suggested that CN paging retransmission interval be 5s
and configuration be 5s with the consideration of paging loss during Iu/Iub flow
control【4】. Reduce paging repeat times and increase paging interval, keep the time

of no paging responses unchanged, interval of reporting prompt tones when there is
no paging response will not be influenced.
CN paging interval and repeat times can be configured through software
parameters.
3.3Improper Setting of UTRAN Paging Retransmission Times and Interval
In order to reduce paging message flow at the Iu interface and increase the
probability that UE receives paging message, UTRAN can retransmit paging
messages. UTRAN paging retransmission configuration should be in accordance
with CN paging retransmission.
Paging is delivered at the fixed time (a paging cycle), and UTRAN paging
interval is the integer multiple of a paging cycle, one paging cycle in general.
Therefore, we can adjust the UTRAN paging repeat interval by adjusting DRC
paging cycle coefficient k.
UTRAN paging repeat times should not be too many; otherwise, adding paging
repeats at Iu interface, the paging channel load at UU interface will increase rapidly.
In addition, UTRAN is realized on MACC layer, and MACC does not identify the
specific PRC message, MACC will go on repeating paging even if UE replies paging
response message. Too many UTRAN paging repeat times will cause the
unnecessary system overhead and message flow at UU interface.
It is reasonable for paging retransmission times to keep the current default
configuration. It is also reasonable to adjust paging retransmission times on RNC
maintenance console using MML command SET WFMRCFGDATA.
3.4CN Adopts Whole-Network Paging
Considering improving call connection rate, CN side can be configured with
whole-network paging, which bears the characteristics that paging has overridden
the concept of location area, and is initiated specific to all UTRANs suspended under
whole CN. In this case, paging traffic becomes larger, especially when multiple
location areas are suspended under CN, the location areas with smaller capacity will
be overloaded, and can not recover for a long time.

Global paging should be avoided in CN. From experiences on GSM network,
the whole-network paging of CN is the important cause of overload of location area
with smaller capacity; meanwhile, CN global paging contributes little to the
improvement of connection rate of long-distance incoming call.
CN global paging is useful only when UE location area is recorded as failure by
NLR, which hardly occurs, however, once it occurs, it means serious fault occurs;
even more CN global paging does not work【6】.
3.5Improper Setting of DRX Paging Cycle Coefficient
When UE is in IDLE and PCH state, UE will utilize the discontinuous reception
technology “Discontinuous Reception (DRX)” to reduce power overhead. According
to Protocol TS25.304, paging cycle length (DRX cycle length) = MAX (2K
，PBP).
where, K refers to DRX cycle length coefficient, PBP refers to Paging Block
Periodicity which is applied only in TDD mode, for FDD, PBP=1, so, DRX cycle
length = 2K
.
DRX paging cycle coefficient of UE has three sources: system message, CN
paging message and UU interface signaling of UTRAN, while the processing way of
CS and PS is different.
For PS domain, DRX paging cycle coefficient is obtained by the negotiation
between UE and SGSN through NAS layer message (attach process), and the
negotiation data is adopted as standard whether UE is in IDLE or connection state,
however, paging coefficient of CS domain is adopted if negotiation fails.
For CS domain, the minimum values of system message and CN paging
message are taken as DRX paging cycle coefficient if UE is in IDLE state. And the
minimum values of system message, CN paging message and signaling at UU
interface of UTRAN are taken as DRX paging cycle coefficient if UE is in connection
state.
Settings of DRX paging cycle coefficient K should be considered with the
following factors:
1. DRX paging cycle coefficient K determines the DRX cycle length, the larger
the value of K is, the longer the DRX cycle is. Meanwhile, UE power
consumption will be lower, however, the value of K makes UE paging cycle
longer. Namely, UE paging response time becomes longer. If K is set too
small, paging cycle will become smaller, and UE processing paging
overhead and power consumption will increase. In the protocol, the value is

within the range of 2 ~12, the current value of our company is 8, and the
paging cycle is 2.56 seconds.
2. According to the current coding mode of PCH, a TTI can transmit only 240
bits, if IMSI paging is utilized, only 3 UEs can be paged at the same paging
moment; if TMSI paging is utilized, only 5 can be paged at the same paging
moment. And if the number of UE paging at the same paging moment
exceeds the processing capacity of system, it will cause paging loss and
call loss. If K is set too small, the paging cycle will become shorter, the
probability of the same calculated UE paging moment will increase,
consequently, the probability of paging loss will increase.
3. Because UTRAN paging retransmission interval is a paging cycle, when
setting DRX paging cycle coefficient K, we should consider the
retransmission interval of UTRAN paging, and the K value should be set in
coordination with CN paging retransmission.
The lower limit of value K should be set with the consideration of saving UE
power consumption and reducing the probability of the same UE paging moment,
and the upper limit of value K should be set with the consideration of the repeat
interval of UTRAN paging. During the actual setting process, the value K should be a
value within 2 ~12, i.e., 5, 6, 7 and 8, and the corresponding paging cycles are
320ms, 640ms, 1280ms and 2560ms. Suppose that the CN paging repeat time is 1
(totally broadcasting 2 times), and the interval is 2s, so how should the UTRAN
repeat times and the intervals be set? Generally, the following rules should be
followed:
1. If the UTRAN retransmission time is 0 (no retransmission of UTRAN), the
DRX cycle of UTRAN shall be smaller than CN retransmission interval (2s) ,namely,
the value K at this time should be set to 7 or 8 (1.28s,2.56s), the setting to be 8 is
considered with the paging loss【4】during lu/luB paging flow control;
2. If the UTRAN retransmission time is 1, the DRX cycle of UTRAN shall be
smaller than half of the CN retransmission interval (1s) , namely, the value K at this
time should be set to 6 or 7 (0.64,1.28s), the setting to be 7 is considered with the
paging loss【4】during lu/luB paging flow control; (Configuration recommended)
3.6Improper setting of Np value

Np refers to PI paging indication number delivered in a certain frame by paging
indication channel PICH, the value of which is within the range of (18, 36, 72, 144),
which is indicated by Number of PI per frame in System Message 5. UE will receive
PICH frame at the specified paging time and find the corresponding PI indicator bit
(the qth
PI). Only when the corresponding PI indicator bit is effective, UE
demodulates the corresponding S-CCPCH frame.
         Np
Np
SFNSFNSFNSFNPIq mod
144
144mod512/64/8/18 










Where, PI = DRX index mod NP = (IMSI div 8192) mod NP, SFN refers to UE paging
moment.
Significance of Np in the actual network: IMSI is divided into Np groups by this
parameter, and all IMSIs in each group adopt the same PI. Influences of Np value on
network: If Np value is too small, the corresponding UE number in each group will be
larger. For each IMSI, if the probability that PI indication occurs increases, times that
IMSI is waken up will increase, which is no good in saving UE power. If Np value is
too large, the corresponding IMSI number in each group will be smaller. For each
IMSI, if the probability that PI indication occurs decreases, times that IMSI is waken
up will decrease. However, if Np is larger, the bit number of each PI will decrease
and PICH demodulation performance of UE will be required higher.
Np value is within the range of (18, 36, 72, 144), which should be determined
according to the current network subscriber quantity (subscriber traffic). Generally,
the value should be larger when subscribers are of large amount, and the value
should be smaller when there are fewer subscribers. In the actual network, the
intermediate 36 or 72 can be selected.
3.7UE Identifier Utilized by CN Paging
Network can page multi UEs at the air interface at the same paging time using
paging type 1 message. Because of PCH capacity restriction, UE quantity is closely
related to UE identifiers used in paging messages, namely, UE identifier influences
the paging channel capacity. Compared with CN paging, the probability that UTRAN
initiate paging is much smaller. Only the case that CN initiates paging is considered:
When UE is in IDLE state, CN can only page three UEs at the same paging time
using IMSI;

When UE is in IDLE state, CN can only page five UEs at the same paging time
using TMSI or PTMSI;
When UE is in CELL_PCH or URA_PCH state, no matter what paging identifier
CN uses, UTRAN transfers UE identifier to U-RNTI and pages, it can only page five
UEs at the same paging time【2】.
From analysis above, CN can increase PCH capacity using UE temporary
identifiers “TMSI and PTMSI”.
CN optimization pages by using UE temporary identifier TMSI and PTMSI, and it
can be adjusted with software parameters.
3.8UTRAN Should Activate IMSI ATTACH and DETACH Functions
After UE is powered on and registered successfully, MSC/VLR will set the
subscriber state as ATTACH state. IMSI DETACH means that after mobile
subscribers power off, MS initiates a DETACH flow and MSC/VLR sets the
subscriber state as IMSI detach. Generally, HLR will not be informed by this flow. In
DETACH state, if the MS is called, CN will directly inform the calling party that the
MS is powered off, which skips of the process of sending useless paging messages
to UE.
As shown in Figure 3 , UE determines whether IMSI ATTACH and DETACH
processes can be adopted through receiving the system message 1. Gsm MAP IE
consists of two octets, one is T3212, and bit1 of the other octet is ATT identifier. “0”
means that the network does not allow UE to adopt IMSI ATTACH and DETACH
processes, “1” refers to the admission of【1】. In Figure 3 , “0a 01” means that T3212
is 60 minutes, UE is admitted to adopt IMSI ATTACH and DETACH processes.
During actual network operation, UTRAN should activate IMSI ATTACH and
DETACH functions.
Table 5 IMSI ATTACH and DETACH Identifier
8 7 6 5 4 3 2 1
T3212 octet 1
Spare ATT octet 2

Figure 3 System Message 1 Resolution
Check and modify IMSI attachment and detachment admission indication by
using the Command “LST CNDOMAIN and MOD CNDOMAIN” on RNC
maintenance console.
3.9Too Low Power allocation of Paging Channels
Phenomenon: Subscriber complaints are centralized in some certain areas,
when UE is called, the prompt tone “the subscriber is out of service” will occur, while
UE can work normally while calling; the print message of paging decoding failure can
be seen at the UE daemon, which is the problem of the over-low paging channel
power allocation.
Paging channels include PICH and PCH. Too low PICH power will lead to UE
resolution PI indication error. Resolution of PI value from “0” to “1” will cause false
alarm and UE power waste; resolution of P1 value from “1” to “0” will cause paging
miss and loss. If PCH power is too low, UE will fail in decoding the paging
messages.
The current baseline power allocation: PICH is -3dB and PCH is -2dB, which are
both relative to pilot channels power.

Improve the power allocation between PICH and PCH properly. Check and
modify PICH power by using MML Command LST CHPWROFFSET and MOD
PICHPWROFFSET on RNC maintenance console, check and modify PCH power by
using Command LST SCCPCH and MOD SCCPCH.
3.10 Existence of Coverage Dead Zones
Phenomenon: Subscriber complaints are centralized in a certain area, UE fails
both in calling and being called, the signal strength shown in UE panel is low, and
RSCP and EC/IO of pilot signals tested with the drive test equipment are lower than
indexes required for UE normal access, by which coverage dead zone is identified in
this area.
Carry out coverage optimization, for specific optimization measures, please
refer to relevant Guidance.
3.11 UE Performance Problems
UEs of different subscribers may be made by different manufacturers, the
receiving performance and demodulation performance of different UEs are also
different. From GSM network experiences, some paging problems are located in
UEs themselves.
Sort out and analyze subscriber complaint materials. If the called UE is of the
same type, problem may exist in UEs.
Take a called verification test on UEs from different manufacturers, if conditions
permit, compare UE receiving performance parameters and demodulation
performance parameters, including testing RSCP and EC/IO of pilot signals received
at the same place and learn about the service demodulation thresholds of UEs.
4 Outstanding Problems
Currently, there are mainly the following outstanding problems existing:
1. Relationship between paging retransmission times and interval of CN and
UTRAN is complicated. In case of PCH capacity, paging success rate and paging

loss rate specified, it is necessary to make simulation analysis on how to work out
the best paging retransmission strategy.
2. The upper limit of location area division is closely related to the service model
and the traffic model of location area. Currently, the theoretical value can be worked
out by using paging area planning tool. Since 3G service model and traffic model are
very complicated, some problems still exist: how big is the difference between
preconditions of paging area planning tool and actual cases, how is the agreement
of the theoretical value and the actual value, which require further theoretical study
and experiences from the network optimization practices.

List of references:
[1] 3GPP R1999 25_series, 2002/09
[2] Miao Jiashy, WCDMA RNP Paging Area Planning Guidance, 2003/03
[3] Chen Qi, Dong Yan, Zhou Xinjie, Zhi Bo, WCDMA RNP Network Optimization Method Soft
Subject Research Report, 2003/08
[4] Zhang Jiayi, WCDMA Paging Strategy Study, 2003/12
[5] Li Peng, Guidance that Subscribers Are out of Service, 2001/11
[6] Huawei GSM Paging Tackling Data, 2003/06
[7] Jiao Anqiang, WCDMA RNP Special Technologies Key Signaling Process Analysis Paging
Process, 2003/11

1. wcdma rno paging problem analysis guidance 20041101-a-1.0

More Related Content

What's hot (20)

Viewers also liked (20)

Similar to 1. wcdma rno paging problem analysis guidance 20041101-a-1.0 (20)

Recently uploaded (20)

1. wcdma rno paging problem analysis guidance 20041101-a-1.0