Professional Documents
Culture Documents
W Access Problem Optimization Guide 20070825 A 3.2
W Access Problem Optimization Guide 20070825 A 3.2
Approved by Date
Contents
1 Introduction..................................................................................................................................14
2 Evaluating Access Performance................................................................................................15
2.1 Accessibility.....................................................................................................................................................15
2.2 System Availability..........................................................................................................................................16
2.3 Access Delay....................................................................................................................................................16
6 Summary.......................................................................................................................................91
7 Appendix 1: Paging Process......................................................................................................92
7.1 Paging Origination...........................................................................................................................................92
7.1.1 Paging by CN.........................................................................................................................................92
7.1.2 Paging by UTRAN.................................................................................................................................92
7.2 Paging Flow.....................................................................................................................................................92
7.2.1 Paging Type 1.........................................................................................................................................92
7.2.2 Paging Type 2.........................................................................................................................................94
7.3 Behaviors of UE after Receiving Paging.........................................................................................................94
Figures
Figure 3-1 Overall flow for analyzing call failure problems in DT/CQT.............................................................20
Figure 3-11 Functions and process of RRC direct retry and re-direction during setup of the RRC connection. .34
Figure 3-13 Signaling for direct retry of a R99 subscriber after admission rejection..........................................36
Figure 4-3 Position for counting point by counter for paging loss by idle UE.....................................................43
Figure 4-4 Position for counting point by counter for paging loss by UE in PCH state......................................44
Figure 4-5 Position for counting point by counter for RRC connection rejection................................................46
Figure 4-6 Position for counting point by counter for CS RAB assignment failure in RNC traffic statistics
starting counting....................................................................................................................................................50
Figure 4-7 Position for counting point by counter for PS RAB assignment failure in RNC traffic statistics......54
Figure 4-8 Position for counting point by counter for RB setup failure in traffic statistics.................................60
Figure 5-1 Originating signaling flow of paging failure due to UE location area update....................................62
Figure 5-2 Content of the Disconnect message in paging failure due to UE location area update......................63
Figure 5-3 Terminating signaling flow of paging failure due to UE location area update...................................63
Figure 5-5 Signal quality when the UE sends the RRC connection request message..........................................66
Figure 5-6 Signal quality when the UE repeats to send the RRC connection request..........................................67
Figure 5-13 Signal strength upon the first sending of RRC connection request..................................................72
Figure 5-15 Signaling and signal strength upon the second sending of RRC connection request.......................73
Figure 5-29 ciphering mode information configured in previous security mode command................................85
Figure 5-32 Signaling of UE upon failure in receiving RRC Connection Setup message...................................87
Figure 8-3 Timing relation between PRACH and AICH as seen at the UE........................................................112
Figure 8-4 Definition of access timeslot set (taking the uplink and downlink access timeslot fixed difference p-a
7680 chips as example)........................................................................................................................................115
Tables
Table 2-1 Indexes and recommended values for accessibility related to DT........................................................15
Table 2-2 Indexes and reference values for accessibility related to traffic statistics............................................15
Table 2-4 Indexes and reference for access delay related to DT...........................................................................16
Table 4-5 Counters related to RRC connection request rejection due to lub interface failure..............................46
Table 4-6 Traffic statistics counters related to RRC connection request rejection due to network congestion....48
Table 4-8 Counters related to RRC connection setup rejection due to redirection...............................................49
Table 4-9 Traffic statistics counters related to CS RAB assignment setup failure due to radio network problems
...............................................................................................................................................................................51
Table 4-10 Traffic statistics counters related to CS RAB assignment setup failure due to insufficient capability
...............................................................................................................................................................................52
Table 4-11 Counter related to CS RAB assignment setup failure due to transmission network problems...........53
Table 4-12 Traffic statistics counters related to PS RAB assignment setup failure due to radio network
problems................................................................................................................................................................55
Table 4-13 Traffic statistics counters related to PS RAB assignment setup failure due to insufficient capability
...............................................................................................................................................................................56
Table 4-14 Counter related to PS RAB assignment setup failure due to transmission network problems...........57
Table 4-15 Counter related to PS RAB setup failure due to no resource available..............................................58
Table 8-3 Broadcast parameters and description of cell reselection in system information...............................108
Table 8-4 Relationship among the access subchannel, access timeslot, and SFN..............................................115
Abstract
The document describes how to locate and solve access problems in WCDMA network
optimization, the definition of access problems, test methods, analysis flows, and solutions.
Finally, the appendix provides the fundamental knowledge necessary for analyzing access
problems by RNO engineers.
DT Drive Test
1 Introduction
This document aims to meet the requirements on solving access problems by on-site engineers
during RNO. It details the methods for evaluating network access performance, test methods,
data analysis methods, FAQs, and solutions. The appendix provides the fundamental
knowledge about access problems, description of principles, related parameters, and data
processing tools. It guides engineers to locate and solve access problems during optimizing
network KPI indexes and network O&M. The RRM algorithms and product implementation
in this document are for RNC V16, unless specified particularly. 3.3.7, 3.3.8, and 3.4.5 are
updated.
This document excludes the usage of tools.
It contains 10 chapters, with the structure as below:
Chapter 1: Introduction
Chapter 2: Evaluating Access Performance
Chapter 3: Analyzing DT/CQT Data
Chapter 4: Analyzing Traffic Statistics Data
Chapter 5: Solving Access Problems
Chapter 6: Summary
Chapter 7: Appendix 1: Paging Process
Chapter 8: Appendix 2: Access Process Analysis
Chapter 9: Appendix 3: Authentication Flow
Chapter 10:Appendix 4: Description of Access-related Parameters
On-site engineers must adjust network optimization parameters according to the importance levels of
parameters, the impact of adjustment on network services and network equipment in a proper time. The
whole adjustment must follow Radio Network Planning Online Data Modification Regulations and data
backup and emergency solutions are necessary. Immediate verification must be performed after
adjustment so that the adjustment is correct.
The access process of HSPA service is similar to that of original R99 service. There are only some minor
differences.
The access performance includes three aspects: accessibility, system availability, and access
delay. The specified indexes for measuring access performance are obtainable by DT and
traffic statistics. For the definition of indexes, see the UMTS Radio Network KPI Baseline
V3.3.
2.1 Accessibility
Table 1.1 lists the indexes and recommended values for accessibility in DT.
Table 1.2 shows the indexes and reference values for accessibility related to traffic statistics.
Table 1.2 Indexes and reference values for accessibility related to traffic statistics
Index Service Statistics method Reference
The values previously mentions are just for reference. Determine the actual values according
to the detailed requirements of projects or requirements of commercial network contracts.
The values previously mentions are just for reference. Determine the actual values according
to the detailed requirements of projects or requirements of commercial network contracts.
The values previously mentions are just for reference. Determine the actual values according
to the detailed requirements of projects or requirements of commercial network contracts.
Step 4 The failure after RAB assignment: after the UE sends the RB SETUP COMPLETE message:
The originated UE receives the DISCONNECT/RELEASE message from CN
The originated UE waits for the CONNECT or ALERTING message until expiration, so
the call clearing process is originated. According to the protocols, after the UE sends the
CM SERVICE REQUEST message, the timer T303 starts. If the UE fails to receive the
CALLPROCEEDING, ALERTING, CONNECT, OR RELEASE COMPLETE message
before expiration of T303, the clearing process starts.
Before receiving alerting message, the UE enters the idle state and starts to receive
system information.
As strictly defined, after the MS enters the CELL_DCH state and before it receives the alerting
message, it must send the cell update message with the cause RLC unrecoverable error/ Radio link
failure.
Take the greater value of the maximum waiting time configured at RLC layer as default and the
synchronization time as the judgment time. It is unclear that the UE can report the RLC layer message,
so the maximum waiting time is neglected.
----End
As strictly defined, after the UE enters the CELL_DCH state, it sends the cell update message with the
cause RLC unrecoverable error/ Radio link failure before receiving the alerting message.
The judgment time should be the greater of maximum waiting time and asynchronization time configure
by default at the RLC layer. It is unknown that the test UE can report RLC layer messages, so neglect
the strict definition.
----End
Figure 1.1 Overall flow for analyzing call failure problems in DT/CQT
By DT data analyzing tool, such as Actix Analyzer and GENEX Assistant, determine the time
for Call Fail and obtain the following information:
Pilot information collected by scanner before and after Call Fail
Information about active set, monitor set, and signaling flow collected by UE
Match the signaling collected by UE and the time of single subscriber tracing by messages.
Meanwhile locate the points when problems occur in single subscriber tracing on RNC.
Based on signaling of single subscriber tracing on RNC and UE's signaling flow, determine
the point where call fails according to Figure 1.1. Analyze and solve problems according to
following sub-flows. The problems include:
Paging problems
RRC setup problems
RAB and RB setup problems
Authentication and encryption problems
1
The CC CALL PROCEDING message should be after CC SETUP.
Cell Reselection by UE
If the signals of the cell where the UE camps are weak while the signals of monitored cell are
strong, the problem might be due to cell reselection. When the UE has its location area (LA)
or route area (RA) updated upon paging, the paging message is sent to the original LA or RA,
so the UE fails to receive paging message.
Figure 1.1 shows the flow for analyzing RRC connection setup problem.
After UE Send RRC Connection Request Message, the RNC Fails to Receive It
If the Ec/Io of downlink CPICH is over low, the problem is about coverage.
If the Ec/Io of downlink CPICH is not over low (such as higher than –14 dB), the problem is
about RACH, with the following causes:
The power step of preamble is small
The output power of UE is lower than required.
NodeB is problematic with standing wave.
The parameter of cell radius is improperly configured.
If the power ramp of preamble is small, you can add the preamble ramp times. For example,
increase it from 8 to 20.
If the output power of UE is lower than required, there are no specific methods to solve it due
to the limitation of UE performance.
For NodeB problems, check whether there is standing wave alarm on NodeB.
If the parameter of cell radius is set over small, the NodeB cannot synchronize with the UEs
beyond cell radius. This causes access failure. This usually occurs in wide coverage scenarios
like rural and suburban areas.
After the RNC Receives the RRC Setup Request Message from UE, It Sends the
RRC Connection Setup Message Which Is Not Received by UE
The causes of this problem include:
Weak coverage
Improper parameters of cell selection and reselection
Check the CPICH Ec/Io. If it is lower than –12 dB (the default value is configured based on
Ec/Io as –12 dB) and
If there is no more qualified cell listed in the monitor set, the problem is about coverage.
If there is more qualified cell listed in the monitor set, the problem is about cell
reselection.
Solutions are as below:
When the coverage is weak:
If conditions permit, solve coverage problems by enhancing coverage, such as adding
sites to cover blind areas and adjusting engineering parameters. If you cannot enhance
the coverage, you can increase FACH power. Adjust FACH power according to the
coverage conditions of PCPICH Ec/Io. For example, the pilot Ec/Io in all coverage areas
after network optimization is larger than –12 dB, the success rate for UE to access from
3G idle mode can be guaranteed if the common channel power is allocated on the
condition of Ec/Io equal to –12 dB. When Ec/Io is smaller than –14 dB, the UE reselects
a GSM cell. The success rate of RRC setup in weak coverage areas after inter-RAT
reselection by UE can be guaranteed if the common power is allocated on the condition
of Ec/Io larger than –14 dB.
Cell selection and reselection:
Adjusting cell selection and reselection parameters accelerates cell selection and
reselection and helps solve RRC connection setup failure problems caused by improper
parameters of cell selection and reselection.
The RRC CONNECTION SETUP message is carried by FACH. After the UTRAN side receives the
PRACH preamble, the UE sends the RRC CONNECTION REQUEST message on RACH based on
current preamble power. The preamble transmit power keeps increasing until response is received
(restricted by maximum retransmission times of preamble). Therefore, in poor coverage areas,
unbalanced coverage by RACH and FACH is probable. Consequently the UTRAN side can receive the
RRC CONNECTION REQUEST message while the UE fails to receive the RRC CONNECTION
SETUP message.
After the RNC Receives the RRC Setup Request Message from UE, It sends the
RRC Connection Reject Message
When the RRC Connection Reject message is present, check the cause values, which include
congestion and unspecified.
If the cause value is congestion, the network is congested. Check the network load conditions,
including utilization of power, code, and CE resources. Determine the type of resource that
causes congestion and provide ways of network expansion. For details, see W-Network
Expansion Guide.
The admission of RRC connection for HSDPA subscribers is consistent with that for T99
subscribers, including power, code, and CE resources. Pay special attention to code
admission. If the code word of HSDPA subscribers is statically assigned, and excessive codes
are assigned to HSDPA subscribers, the RRC connection of HSDAP or R99 subscribers fails
probably. This is due to that the codes of downlink signaling channel for HSDPA or R99
subscribers are inadequate.
If the cause value is unspecified, check the logs to determine causes of failure.
After Receiving RRC Connection Setup Message, the UE Does Not Send Setup
Complete Message
If the downlink signals are normal, the UE might be abnormal. Otherwise initial power of
downlink DCH is over low so the downlink cannot synchronize. You can solve the problem
by adjusting uplink Eb/No of the service.
After the UE Sends the RRC Setup Complete Message, the RNC Fails to Receive
It
It seldom occurs that uplink initial power control leads to increment of UE transmit power.
Upon presence of the problems, you can properly raise the Constant Value of DCH so that
the initial transmit power of uplink DPCCH of UE increases.
This problem is related to whether the initial target value of uplink SIR is rational and has
great impact on uplink initial synchronization at the beginning of link setup.
If it is set over large, the uplink interference from initial link setup of subscriber becomes
over large.
If it is set over small, the uplink synchronization time increases, and consequently the
initial synchronization fails.
This parameter is an RNC-level parameter. It has great impact on network performance, so
engineers must be cautious upon adjustment.
The RRC Connection Setup Complete message is sent on uplink DPCH. The UE calculates the initial
power of DPCCH according to received IE DPCCH_Power_offest and measured CPICH_RSCP.
DPCCH_Initial_power = DPCCH_Power_offset - CPICH_RSCP
Wherein, DPCCH_Initial_power = Primary CPICH DL TX Power + UL Interference + Constant Value
Constant Value can be configured at OMC. If it is set over small, the UE has lower power to send the
RRC CONNECTION SETUP COMPLETE message than required. Current default configuration of
Constant Value (the default value of it in version V13C03B151 is –20) usually prevents this problem
from happening.
MAC Failure
Check the AUTN parameter in the authentication request message send by network side upon
the authentication of network by UE. If the MAC information is incorrect, the UE send the
authentication failure message with the cause value MAC failure, shown as in Figure 1.1.
Sync Failure
When the UE detects that the SQN of AUTN message is incorrect, so the authentication fails.
The cause value is Synch failure (synchronization failure), as shown in Figure 1.1.
UE Capability Problems
To check the UE capability, refer to the RRC Connect Setup CMP message.
Currently the following UEs fail to support encryption algorithm:
NEC single-mode UEs
NEC C606
NEC C616
The following UEs support encryption algorithm:
Nokia 7600
Nokia 6650
Moto A835
Qualcomm 6200
Qualcomm 6250
Siemens U15
To solve the security mode reject problems due to UE capability, change the UE.
Admission Rejection
For non-HSDPA subscribers, when the system resource (power, code, channel code, Iub
transmission resource, and Credit) is inadequate, the admission is rejected and consequently
call setup fails. Now you must check the uplink and downlink load, code resource, Iub
transmission resource, and CE resource, determine the type of resource that causes
congestion, and provide corresponding expansion methods.
When excessive codes are statically assigned to HSDPA subscribers, the admission fails
due to inadequate downlink channel code resource for non-HSDPA subscribers. When
the system resource is inadequate and admission fails, the V1.5 or higher RNC conducts
different operations according to RAB Downsizing. For details, see the description of
solving inadequate lub bandwidth.
If the cell does not support the HSDPA service, the admission of R99 subscribers
depends on the set R99 admission threshold. If the cell supports the HSDPA service and
the downlink power of HSDPA and R99 subscribers is statically assigned, the power
admission of non-HSDPA subscribers is judged by (total power of cell - the power
statically assigned t HSDPA) * admission threshold. When the power of HSDPA and
R99 subscribers is dynamically assigned, the power admission of non-HSDPA
subscribers is consistent with that of original R99 subscribers. The uplink admission
decision is based on the RTWP or equivalent number of subscribers. If the uplink load is
too high, the admission of non-HSDPA subscribers is rejected.
If the bandwidth of the lub interface is inadequate, activation of R99 high-speed data
services fails due to the limited bandwidth. For example, the AAL2 bandwidth for
service on lub interfaces in many cells can support only a 384 Kbps service. If a 12.2
Kbps voice service already exists, the lub interface fails to provide enough bandwidth for
a 384 Kbps PS service. In the case of RNC V1.3, the RNC returns an SGSN RAB
assignment failure because the requested rate is unavailable. The SGSN then originates
RAB assignment through re-negotiation. In the case of RNC V1.5 or later versions, the
RNC lowers the rate first if the RAB Downsizing switch is on. If the lub resource is
available after the rate is down, the RNC sends a RAB assignment success message to
the SGSN. If the lub resource is not available even though the rate is down to 8 Kbps,
the RNC returns an SGSN RAB assignment failure. The SGSN then decides whether or
not to originate renegotiation based on its internal parameters. If the RAB Downsizing
Switch is off, the processing is the same as that in the case of RNC1.3.
The admission control of NodeB Credit resources is similar to the power admission
control. Whether the available Credit can support the currently requested service depends
on the spreading factor of the new subscriber. If the current Credit is not adequate, the
RNC performs different processing depending on state of the RAB Downsizing switch in
the case of RNC V1.5 or later versions. For details, see the handling in the case of
inadequate bandwidth of the lub interface as described earlier in this document.
For the admission rejection of HSDPA subscribers, consider the following aspects:
In the method for statically assigning power of HSDPA and R99 subscribers, consider:
− HSDPA subscribers supported by NodeB
− HSDPA subscribers supported by cell
− Total bit rate of cell
− Total guaranteed bit rate
− Whether the cell transmit power guaranteed bit rate exceeds the prescribed threshold
In the method for dynamically assigning power of HSDPA and R99 subscribers,
consider:
− HSDPA subscribers supported by NodeB
− HSDPA subscribers supported by cell
− Whether the guaranteed bit rate exceeds the prescribed threshold
For HSDPA subscribers, when the configured bandwidth at lub interface is inadequate,
admission rejection will not occur, but the rate become lower. In addition, the AAL2PATH of
HSDPA and R99 is respectively configured, and HSDPA AAL2PATH must be configured to
HSDPA RT or HSDPA NRT type. If the HSDPA AAL2PATH is configured to R99
AAL2PATH RT or NRT type, RAB assignment will not fail, but the RNC will directly set up
HSDPA service to R99 384kbps.
For V17, strategies of the RRM admission algorithm change as follows:
Downlink power admission control for HSDPA cells is supported. Only dynamic power
assignment is available. For the DCH service, consider whether load of the non-HSDPA
service (the R99 service) exceeds the admission threshold of the non-HSDPA service
(that is, the admission threshold of the original R99 service). In addition, consider
whether the non-HSDPA power and the HSDPA GBP (Power to meet GBR) exceed the
threshold of total power of the cell. For the HSDPA service, check whether the HSDPA
throughput provided by the cell exceeds the threshold of sum of Guaranteed Bit Rates
(GBR) of all subscribers, or whether the GBP of stream services and background
services exceeds the HSDPA power of the cell. In addition, consider whether the non-
HSDPA power and the HSDPA GBP exceed the threshold of total power of the cell.
lub interface admission: For the DCH service, the admission depends on the peak bit rate
multiplied by the activation factor of the service. For the HSDPA service, the admission
depends on the GBR. If the lub interface reaches the congestion threshold, DCCC
downsizing occurs. If the RLC_AM re-transmission ratio exceeds the specified
threshold, run the SET CORRMALGOSWITCH command to enable lub Overbooking.
In this case, TF of the R99 occurs or rate of the HSDPA service decreases based on the
related factor. Run the ADD AAL2ADJNODE command to set the service activation
factor and the lub congestion threshold. Run the ADD TYPRABRLC command to set
trigger and release thresholds of RLC_AM re-transmission.
low, the UE originates to access the network in the best server, set the threshold for starting
selection of intra-frequency to –4 dB and set Treselection to 1. For cells at edge of different
LACs, set the threshold lower to decrease signaling traffic of location area update.
The RB setup failure due to weak coverage includes unqualified uplink and downlink
coverage.
The RB setup failure due to downlink weak coverage is as below:
The UE fails to receive the RB setup command. Unqualified downlink coverage is partially
due to poor demodulation performance of UE. It must be solved by RF optimization.
The RB setup failure due to downlink weak coverage is as below:
The UE receives the RB setup command, but the RAN fails to receive the ACK message or
RB Setup Complete message for RB setup. This is probably due to uplink interference. Check
RTWP for this.
This section describes access process in the case of dualband networking only.
The access in the case of dualband networking involves direct retry and re-direction in the
RRC connection stage and RAB direct retry. RAB direct retry includes service-based direct
retry and that after admission failure. The direct retry and re-direction algorithms are used to
increase first put-through ratio of the UE.
The following figure shows functions and process of RRC direct retry and re-direction during
setup of the RRC connection.
Figure 1.3 Functions and process of RRC direct retry and re-direction during setup of the RRC
connection
If the RNC receives a RRC connection request, the admission algorithm decides whether a
RRC connection is allowed between the UE and the current cell based on the load over the
current cell.
If the RRC connection is allowed, the RNC sends a RRC CONNECTION SETUP
message to the UE, and then the UE sets up a RRC connection.
If the RRC connection is not allowed, the RNC direct retry algorithm module searches
for a cell that complies with the direct retry algorithm in the UE candidate list.
− If a suitable target cell exists, the RNC sends the target cell data to the UE through a
RRC CONNECTION SETUP message.
− If no suitable cell exists, the RNC re-direction algorithm selects another suitable
frequency or radio access system (such as GSM), and then notifies the UE of the
REDIRECTION cell through a RRC CONNECTION REJECT message. The UE
originates an access request in the specified frequency or system.
The RAB direct retry includes service-based direct retry and RAB direct retry after admission
failure.
Service-based direct retry
If a R99 cell and a HSPA cell are at different frequencies but with the same coverage,
subscribers who are requesting for the R99 service are assigned to the R99 cell and those
who are requesting for the HSPA service are assigned to the HSPA cell if possible. Thus,
the R99 service is separated from the HSPA service.
Direct retry after admission failure
After admission failure, the subscriber can be connected to a cell at another frequency
with the same coverage, a HCS cell at another frequency, or a cell in another system (for
the AMR service only).
If admission of the HSPA service fails and the direct retry fails in all cells, the service
returns to the DCH of the local cell and a RAB connection is set up on the DCH again.
In dualband scenario 1, the services are carried in R99 cells as the strategy. In this case, the
real-time services do not need inter-frequency direct retry during call setup. Thus, the impact
on real-time services decreases and the HSPA subscribers can access the cells that support
HSPA through service-based direct retry. To make the UE reside in F1, modify the cell
selection and reselection parameter Qoffset2,n of F1 to 50 dB and that of F2 to -50 dB, or bar
F2.
In dualband scenario 2, the strategy of random cell residence is used. The UE originates
service access in the serving cell. All cells that use two carriers adopt the default value of the
Qoffset2,n parameter.
In both scenarios, RRC or RAB direct retry is available for the R99 service and RAB direct
retry after admission rejection is available for the HSPA service. If the direct retry of the
HSDPA service fails, the service returns to the DCH of the local cell and an RAB connection
is set up on the DCH again.
Example 1: In scenario 1, the HSDPA data card resides in the R99 cell. If the PDP is
activated, the subscriber accesses the HSDPA cell through direct retry. Figure 1.4 shows the
signaling for a service-based direct retry.
Example 2: In scenario 1, the admission threshold of the R99 service is exceeded if several
R99 subscribers access the cell. If one more R99 subscriber tries to access the cell, the
admission is rejected in the R99 cells at F1. In this case, the subscriber accesses a
R99+HSDPA cell through direct retry. Figure 1.5 shows the signaling for direct retry after
admission rejection.
Figure 1.5 Signaling for direct retry of a R99 subscriber after admission rejection
When DRX is 6, 7, and 8, the paging period is respectively 640ms, 1280ms, and 2560ms. In
terms of statistics probability, if enough UEs originate enough calls, the traffic is in Poisson
distribution and the average access delay keeps increasing. According to on-site test result,
when DRX is 8, most paging delays are between 1s and 1.5s. The longest paging delay is
even longer than 2.5s. When DRX is 6, the paging delay is evenly distributed between 0.35s
and 0.95s. Therefore, setting DRX to 6 can effectively lower proceeding delay.
Setting DRX to 6 leads to accelerated power consumption by UE, so you must consider this
with the actual network conditions. At the beginning of network operation, the key is to raise
the RAN performance. According to partners' signaling, such as Nokia, Ericsson, ZTE, and
Lucent, setting DRX to 6 is the majority.
Terminating call Starting assignment before the call is Starting assignment after
answered the call is answered
Originating call Starting assignment before the Starting assignment after
Alerting message the Alerting message
Early assignment increases call completion rate. Late assignment avoids occupation of TCH
resource during ringing, so it increases the utilization of TCH resource.
According to test result, the UE receives Alerting message 1.28s earlier in early assignment
than in late assignment. Using late assignment helps to receive response signals (ringing)
from network more quickly, so it is more rational. However, the UE might fail to put through,
so late assignment affects call completion rate. You must balance the advantages and
disadvantages before using it.
2
The data is from the research and test result of network optimization in a pilot office. The following data is
also from this source.
DCH3.4K and 0.491s shorter than that set up on DCH 13.6K. The signaling of RRC
connection set up on DCH 13.6K comparatively occupy more resource, it is recommendable
if the resource at early stage of network operation is adequate.
Note that after RRC connection setup, the signaling is reconfigured on DCH3.4K upon RB
setup.
3
Due to problems on traffic statistics, current cluster-level analysis is not concerned. It is the same as RNC-
level analysis.
Therefore, in RNC traffic statistics, the counters for failure cause type is usually cell-level. In
the following sections, priority is given to cell-level indexes.
If the new indexes meet requirements, the analysis ends. If there are still problems,
continue the analysis until the indexes meet requirements.
For the methods and flow for analyze problematic cells, see the flow for analyzing cell-level
traffic statistics data.
The flow for analyzing cell-level traffic statistics data proceeds as below:
Check whether there are cells with unsatisfied indexes
If no, the analysis ends.
If there are cells with unsatisfied indexes, analyze the detailed causes, find major causes
of indexes deterioration, and provide proper solutions.
After carrying out solutions, analyze the new data of traffic statistics until indexes meet
requirements.
Figure 1.2 Position for counting point by counter for paging loss by idle UE
When the RNC receives the paging message from CN, and the paged UE is in idle state, the
RNC takes statistics of VS.RANAP.Paging.Att.IdleUE at the point B upon sending PAGING
TYPE 1 message to the cells in paged area.
When the RNC receives the RRC CONNECTION REJECT message from UE, the RNC takes
statistics of VS.RANAP.Paging.Succ.IdleUE at the point C if the RRC connection setup
request is due to one of the following causes:
Terminating Conversational Call
Terminating Streaming Call
Terminating Interactive Call
Terminating Background Call
Terminating High Priority Signaling
Terminating Low Priority Signaling
Terminating cause unknown
The causes of paging loss by idle UE usually include:
Parameter configuration problem. For this problem, check the paging-related parameters
whether they are configured as the baseline parameters.
Weak coverage. For example, the RNC cannot page a UE in indoor UE without being
covered by signals, or a UE in blind elevator area.
Table 1.1 shows the counters related to paging loss for UE in PCH state.
VS.UTRAN.Paging1.Att It counts the times that the RNC sends the PAGING
TYPE 1 message.
VS.UTRAN.SuccPage1 It counts the times that the RNC succeed in sending the
PAGING TYPE 1 message.
VS.RRC.Paging1.Fail.PchUE It counts the times that the RNC fails to send the
PAGING TYPE 1 message.
Figure 1.2 Position for counting point by counter for paging loss by UE in PCH state
When the RNC sends the PAGING TYPE 1 message to the UE in CELL_PCH or URA_PCH
state, it takes statistics of VS.UTRAN.Paging1.Att at the point A.
When the UE in CELL_PCH or URA_PCH state receives the PAGING TYPE 1 message, it
sends RNC the CELL UPDATE message. The cause for cell update is paging response. When
the RNC receives the cell update message from UE, with the cause paging response, it takes
statistics of VS.UTRAN.SuccPage1 at the point B.
The RNC will not count the times of sending PAGING TYPE 1 message due to modification of system
information.
Flow Control
When the lu interface is in flow control state, it drops the paging messages from CN.
Table 1.1 shows the counter related to paging loss due to flow control.
When the RNC receives the paging message from CN, and the IU interface is in paging flow
control state, the IU interface drops the paging message and counts the time.
When paging messages are dropped due to flow control at IU interface, the traffic of network
must be heavy. Therefore, precaution to network expansion must be performed.
PCH Congestion
When the RNC receives paging message from CN, and the current paging flow exceeds the
maximum capacity of PCH, PCH is congested and paging messages are dropped.
Table 1.1 lists the counters related to PCH congestion.
If paging messages are dropped in the cell due to PCH congestion, the paging traffic of the
cell must have reached the maximum traffic. Check the parameters related to repeated paging
and PCH. If this problem is due to heavy traffic volume, split the location area.
Figure 1.2 Position for counting point by counter for RRC connection rejection
After the RNC sends the RRC CONNECTION SETUP message, it fails to receive the
RRC CONNECTION SETUP COMPLETE or RRC CONNECTION SETUP FAILED
message from UE. This corresponds to the third cause listed previously.
Table 1.1 Counters related to RRC connection request rejection due to lub interface failure
Counter name Counter description
VS.RRC.Rej.RL.Fail It counts the times that RRC connection setup is rejected due to
RL setup failure
VS.RRC.Rej.AAL2.Fail It counts the times that RRC connection setup is rejected due to
AAL2 synchronization failure
4
Others are calculated indexes, equal to total failure times (RRC.FailConnEstab.Cong) minus previous failure
times.
Table 1.1 lists the counters related to RRC connection request rejection due to network
congestion.
Table 1.1 Traffic statistics counters related to RRC connection request rejection due to network
congestion
Counter name Counter description
RRC.FailConnEstab.NoReply It counts the times that RRC connection setup fails due to
no response
Table 1.1 Counters related to RRC connection setup rejection due to redirection
Counter name Counter description
Other problems5
When CS RAB assignment fails, the counter starts counting at the point B shown in Figure
1.2.
Figure 1.2 Position for counting point by counter for CS RAB assignment failure in RNC traffic
statistics starting counting
At the point B in Figure 1.2, when the RNC sends CN the RAB ASSIGNMENT RESPONSE
message with the cause failure, the corresponding counter starts working according to
specific failure causes. The RB SETUP process is marked in broken line and is optional.
5
This index is a calculated index, equal to RAB request times minus the sum of RAB success times and other
failure times.
6
This index is a calculated index, equal to total failure times (VS.RAB.FailEstabCS.RNL) minus previous
failure times.
Table 1.1 Traffic statistics counters related to CS RAB assignment setup failure due to radio
network problems
Counter name Counter description
VS.RAB.FailEstabCS.RNL It counts the times that CS RAB assignment setup fails due
to radio network problems.
It counts the total failure times.
VS.RAB.FailEstCS.Relo It counts the times that CS RAB assignment setup fails due
to relocation.
VS.RAB.FailEstCS.RIPFail It counts the times that CS RAB assignment setup fails due
to air interface failure.
VS.RAB.FailEstCS.Unsp It counts the times that CS RAB assignment setup fails due
to insufficient capability.
The indexes listed in Table 1.1 are cell-level indexes. For detailed VS.RAB.FailEstCS.Unsp,
see Table 1.2.
CS RAB assignment setup fails due to relocation
When the RNC carries out relocation, it receives the RAB ASSISNMENT REQUEST
message, and then it will not respond to the message but respond the RAB
ASSISNMENT RESPONSE message directly to CN (due to Relocation Triggered). This
index is seldom present, so neglect it.
CS RAB assignment setup fails due to air interface failure
After the RNC receives the RB Setup Failure message from UE, it sends the RAB
Assignment Response message to CN due to Failure in the Radio Interface Procedure.
To analyze CS RAB assignment setup fails due to air interface failure, you must analyze
the causes of RB setup failure. For details, see RB setup failure in 4.3.5.
CS RAB assignment setup fails due to insufficient capability
The detailed causes of CS RAB assignment setup failure due to insufficient capability
include:
− Requested Traffic Class not Available (18)
− Requested Maximum Bit Rate not Available (20)
− Requested Maximum Bit Rate for DL not Available (33)
− Requested Maximum Bit Rate for UL not Available (34)
− Requested Guaranteed Bit Rate not Available (21)
− Requested Guaranteed Bit Rate for DL not Available (35)
− Requested Guaranteed Bit Rate for UL not Available (36)
− Requested Transfer Delay not Achievable (22)
CS RAB assignment setup fails due to insufficient capability when the cell is congested,
such as Requested Maximum Bit Rate not Available. Note that the causes of the indexes
include the following causes of failure due to radio resource congestion:
− CS RAB is rejected due to inadequate power
− CS RAB is rejected due to uplink CE resource
− CS RAB is rejected due to downlink CE resource
− CS RAB is rejected due to code resource
Table 1.2 Traffic statistics counters related to CS RAB assignment setup failure due to insufficient
capability
Counter name Counter description
7
This index is a calculated index, equal to total failure times (VS.RAB.FailEstCS.Unsp) minus times of other
four failures.
Table 1.1 Counter related to CS RAB assignment setup failure due to transmission network
problems
Counter name Counter description
Other Causes
The index seldom occurs, so this document neglects it.
Figure 1.1 Position for counting point by counter for PS RAB assignment failure in RNC traffic
statistics
8
This index is a calculated index, equal to RAB request times minus the sum of RAB success times and other
failure times
At the point B in Figure 1.1, when the RNC sends CN the RAB ASSIGNMENT RESPONSE
message with the cause failure, the corresponding counter starts working according to
specific failure causes. The RB SETUP process is marked in broken line and is optional.
Table 1.1 Traffic statistics counters related to PS RAB assignment setup failure due to radio
network problems.
Counter name Counter description
VS.RAB.FailEstPS.Par It counts the times that PS RAB assignment setup fails due
to parameter errors.
It counts the total failure times.
VS.RAB.FailEstPS.Relo It counts the times that PS RAB assignment setup fails due
to relocation.
VS.RAB.FailEstPS.RIPFail It counts the times that PS RAB assignment setup fails due
to air interface failure.
VS.RAB.FailEstPS.Unsp It counts the times that PS RAB assignment setup fails due
to insufficient capability.
The indexes listed in Table 1.1are cell-level indexes. For detailed VS.RAB.FailEstPS.Unsp,
see IStep 11Table 1.1.
PS RAB assignment setup fails due to parameter errors
The detailed causes include:
− Invalid RAB Parameters Value
− Invalid RAB Parameters Combination
− Condition Violation for SDU Parameters
− Condition Violation for Traffic Handling Priority
− Condition Violation for Guaranteed Bit Rate.
The index is seldom present, so this part neglects it.
PS RAB assignment setup fails due to relocation
9
This index is a calculated index, equal to total failure times minus other failure times.
When the RNC carries out relocation, it receives the RAB ASSISNMENT REQUEST
message, and then it will not respond to the message but respond the RAB
ASSISNMENT RESPONSE message directly to CN (due to Relocation Triggered).
This index is seldom present, so neglect it.
PS RAB assignment setup fails due to air interface failure
After the RNC receives the RB Setup Failure message from UE, it sends the RAB
Assignment Response message to CN due to Failure in the Radio Interface Procedure.
To analyze PS RAB assignment setup fails due to air interface failure, you must analyze
the causes of RB setup failure. For details, see RB setup failure in 4.3.5.
PS RAB assignment setup fails due to insufficient capability
The detailed causes of PS RAB assignment setup failure due to insufficient capability
include:
Requested Traffic Class not Available (18)
Requested Maximum Bit Rate not Available (20)
Requested Maximum Bit Rate for DL not Available (33)
Requested Maximum Bit Rate for UL not Available (34)
Requested Guaranteed Bit Rate not Available (21)
Requested Guaranteed Bit Rate for DL not Available (35)
Requested Guaranteed Bit Rate for UL not Available (36)
Requested Transfer Delay not Achievable (22)
The index is present when the cell is congested. For detailed causes, such as Requested
Maximum Bit Rate not Available, see CDL. Note that the causes of the indexes include the
following causes of failure due to radio resource congestion:
PS RAB is rejected due to inadequate power
PS RAB is rejected due to uplink CE resource
PS RAB is rejected due to downlink CE resource
PS RAB is rejected due to code resource
Others10
Table 1.2 lists the traffic statistics counters related to PS RAB assignment setup failure due to
insufficient capability.
Table 1.2 Traffic statistics counters related to PS RAB assignment setup failure due to insufficient
capability
Counter name Counter description
10
This index is a calculated index, equal to total failure times (VS.RAB.FailEstCS.Unsp) minus times of
other four failures.
Table 1.1 Counter related to PS RAB assignment setup failure due to transmission network
problems
Counter name Counter description
No Resources Available
No resource available causes PS RAB setup failure
Table 1.1 shows the counter related to PS RAB setup failure due to no resource available.
Table 1.1 Counter related to PS RAB setup failure due to no resource available
Counter name Counter description
Other Causes
The index seldom occurs, so this document neglects it.
For HSDPA service, the cause of low success rate of RAB assignment is the same as that of
R99 PS RAB assignment. The traffic statistics indexes of PS RAB involve R99 PS service
and HSDPA service.
HSDPA RAB Setup Success Ratio
This KPI can be used to evaluate the RAB setup success ratio of the PS service carried by
HSDPA.
Measurement Cell
Scope
Notes The RNC level KPI is calculated by aggregating all the cell counters.
Measurement Cell
Scope
Notes The RNC level KPI is calculated by aggregating all the cell counters.
RB Setup failure
RB setup failure: after the RNC sends the RB Setup message, it receives the RB Setup Failure
message from UE.
The detailed causes of RB setup failure include:
Unsupported configuration
Physical channel failure
Cell update occurrence
Invalid configuration
In traffic statistics of RNC, the counter for RB setup failure starts counting at the point A
shown in Figure 1.1.
Figure 1.1 Position for counting point by counter for RB setup failure in traffic statistics
At the point A shown in Figure 1.1, when the RNC receives the RADIO BEARER SETUP
FAILURE message from UE, it takes statistics according to various causes of RB setup failure
in the cell where UE camps.
Table 1.1 lists the traffic statistics counters related to RB setup failure.
sends the Disconnect command with the cause Bearer capability not authorized after
call proceeding. Consequently, the UE receives the RB_SETUP message and has not
completed RB setup, so it responds RB setup failure upon receiving the Disconnect
message, and then the RNC responds RAB setup failure.
No response to RB Setup
Table 1.1 lists the traffic statistics counter related to no response to RB setup.
Figure 1.1 Originating signaling flow of paging failure due to UE location area update
According to Figure 1.1, the RNC receives the Disconnect message from CN.
Figure 1.2 shows the content of the Disconnect message.
Figure 1.2 Content of the Disconnect message in paging failure due to UE location area update
According to Figure 1.2, the cause value for the Disconnect message is no route to
destination. Therefore, the connection is released because the destination UE cannot be
paged.
Figure 1.3 shows the terminating signaling flow of paging failure due to UE location area
update.
Figure 1.3 Terminating signaling flow of paging failure due to UE location area update
According to Figure 1.3, the terminating UE has location area and route areas updated. During
the process, the UTRAN fails to page UE, so the call fails.
Solution
No general solution is for this problem. You can rationally configure the location area and
route area to avoid frequent location area update in hot spot areas.
The interval between two times of repeating to send request by UE is about 1.2s.
Figure 1.2 Signal quality when the UE sends the RRC connection request message.
Figure 1.2 Signal quality when the UE sends the RRC connection request message.
Figure 1.3 shows the Signal quality when the UE resends the RRC connection request.
Figure 1.3 Signal quality when the UE repeats to send the RRC connection request
Solution
To reduce the time for cell reselection as possible, modify Qhyst2 to 0, SintraSearch to 7.
During walking test, ping-pong cell reselection occurs without decrement of reselection time.
It is recommended that:
Qhyst2 remains 2 dB
After the table shown in Figure 1.3 is sorted by time order, you can see that the RNC responds
to the second RRC setup connection message from UE.
According to Figure 1.3, the downlink signals are strong, so the uplink signals should be
strong. Why does the first connection fail?
After static test, the problem reoccurs. The problem occurs every half hour. Sometimes, call
fails because repeating to send RRC connection request fails four times. According to analysis
of signaling for tracing TMSI by RNC, the RNC fails to receive the RRC Connection Request
message. The signal strength of the cell: RSCP ranges from –60 dBm to –70 dBm; Ec/No
ranges from –2 dB to –4 dB. According to previous interference analysis, regular interference
is present in the cell, as shown in Figure 1.4.
Interference lasts for 40s each time. The last peak shown in Figure 1.4 is not the periodic interference as
previous ones. It lasts for a short time (a sampling point). The interference is present from 8:00 to 21:00.
There is no interference in other time.
At first, engineers guess that the interference causes the problem. After sampling data, based
on RNC messages, UE messages, and recorded RTWP, there is no interference one minute
before and after call fails. Therefore the problem is irrelevant to interference.
The following tests are to locate the causes of problems:
Test with Qualcomm handset (6200). During one-hour call, no similar phenomenon is
present, but the interference is still present. This proves that Qualcomm test handset is
normal.
To check that the problem is not due to AICH, raise the AICH power to 0 dB. Test with
Moto handset and the problem is still present. Namely, the problem is irrelevant to AICH
power.
Restore the AICH power to –7 dB, restore the retransmission times of preamble from 8
to 20. During the test more than one hour, the problem is not present.
Signals are stable and strong during indoor static test: Ec/Io is about –3 dB and RSCP is
about –50 dBm. So engineers doubt that the power of Moto handset is problematic in
areas with strong signals. After engineers lower the Ec/Io to –7 dB by increasing
downlink load, the RRC setup problem remains.
To further confirm that the problem is not caused by interference, test after 22:00.
According to the RTWP, there is no interference, but repeating to send request message
occurs fours times in the test longer than one hour, with two times of Call Fail.
According to comparative analysis of interference record by NodeB and system
information by UE, the interference value is changed. During repeating to send request
message, the interference before and after system information remain the same (–105
dB), which proves that the repeating to send request problem of Moto handset is
irrelevant to external interference.
According to previous test, a conclusion is drawn that the problem is irrelevant to uplink
interference and power configuration of AICH. Since the Qualcomm handset (6200) is normal
in the test, the problem must be with uplink RACH of Moto handset.
Solution
After engineers change the retransmission times of preamble from 8 to 20, the problem never
occurs again.
Figure 1.2 shows the signal strength upon the first sending of RRC connection request.
Figure 1.2 Signal strength upon the first sending of RRC connection request
In Figure 1.2,
The second column is the signal strength of serving cell
The third column is the scramble of the serving cell
The fourth and fifth column is the signal strength and scramble of the best monitored
cell.
The signals of these two cells keep fluctuating.
Figure 1.3 shows the single subscriber tracing signaling by RNC
Because the downlink coverage is weak, so the UE originates the RRC connection request
message. Consequently, the RNC receives the RRC connection request message and sends the
RRC connection setup message. However, the downlink signals are weak, so the UE fails to
receive the RRC connection setup message.
Figure 1.4 shows the signaling and signal strength upon the second sending of RRC
connection request after 2s.
Figure 1.4 Signaling and signal strength upon the second sending of RRC connection request
When the UE sends the RRC connection request message the second time, the downlink
signal strength is about –13, so the connection succeeds. According to the Figure 1.4, when
the Ec/Io of downlink signals is lower than –12 dB, it is not guaranteed that the UE can
correctly demodulate data from downlink FACH.
The current FACH power is –1 dB, which is provided based on the relationship curve of
FACH Ec/No and power allocation rate tested on field on the assumption that Ec/Io at cell
edge is –12 dB. To raise the receiving success rate when Ec/Io power is –14 dB, raising the
FACH power by 2 dB is recommended out of the consideration for the threshold for starting
inter-RAT measurement.
Solution
After the FACH power is set to 1 dB, the problem no longer exists that RRC connection fails
because the UE cannot receive RRC Setup message in downlink.
Figure 1.2 shows the traced signaling at RNC side. According to UE side, after the UE sends
the initial direct transfer message, it does not receive any message. Therefore, it repeats to
resend the RRC connection request message after 5s.
According to UU interface of RNC, after the RNC receives the initial direct transfer message,
it sends the authentication request message (probably there is no authentication, so the RNC
directly sends the Security mode command). But there is no response, so the RNC release
RRC connection after expiration.
By comparison of messages at UE side and RNC side, the UE fails to receive the
measurement control and authentication request message (sometimes Security mode
command). According to the BLER statistics at UE side, after the UE sends the initial direct
transfer message, the BLER with 32 as the Trch ID is 100%. Therefore decoding signaling RB
on downlink transport channel is all wrong, so the UE fails to receive any message from
downlink DCH. The RRC setup message is sent on CCH, so the UE can receive it.
From previous analysis, the downlink DCH might be problematic.
According to Figure 1.4, when the cell of PSC 205 is listed in active set and multiple braches
of SHO are combined, the BLER increases but call drop does not occur. When signals of
other cell are weak, call drops easily, as shown in Figure 1.5.
The signals from the cell of SC 205 are strong, but the downlink BLER is 100%, so the call
drops. Therefore, the cell of SC 205 is problematic in downlink.
Meanwhile, another cell under the same NodeB is normal. Therefore engineers doubt that a
DSP of downlink NDLP board on NodeB is problematic. After connect a normal cell to the
DSP, the problem is still present. Therefore, the DSP must be problematic. After NDLP reset,
the DSP becomes normal. The DSP is fixed not after activation and deactivation, but after
reset.
After tracing performance of cells under RNC and tracing assignment of code tree, engineers
find:
CCH uses a code word (SF = 32) (PCPICH SF=256, PCCPCH SF=256, AICH SF=256,
PICH SF=256, and SCCPCH SF=64)
Four HS-SCCHs use a code word (SF = 32)
15 HS-PDSCHs use 30 codes (SF=32)
When an HSPDA subscriber accesses the network, a code word (SF = 128) is necessary for
13.6K signaling. However, no code word is available now, so the RRC connection is rejected.
Figure 1.1 shows the assignment of HSDPA code tree.
According to the assignment of HSDPA code tree in Figure 1.1, the number of codes assigned
for HSDPA subscribers is clear, as well as the rest codes and the occupation by R99
subscribers. Figure 1.1 shows a sample.
Besides tracing code tree, engineers can obtain the cause of admission failure from RNC logs
based on the time for admission rejection, and IMSI.
Solution
After engineers assign 14 codes for HS-PDSCH, the HSDPA subscriber succeeds in access the
network.
When codes are statically assigned to HSDPA subscribers, the admission is usually rejected
due to code word restriction. The PS 384Kbps R99 subscribers and other R99 subscribers in
the cell use most codes, so the admission of HSDPA subscribers in downlink is rejected
because the DCH obtains no code word when HSDPA subscribers connect to the network.
According to Figure 1.1, the UE receives the Disconnect message after completion of RB
setup. The CN releases the connection, so the connection fails. The cause value of Disconnect
is requested circuit channel not available.
The terminating UE is in the equipment room of RNC. There is indoor coverage system in the
equipment room, so the coverage is good. But excessive subscribers are using the network, so
the network is congested and the connection fails.
Solution
Solve access failure problems due to inadequate capacity by network expansion.
This failure occurs under a special background, because excessive subscribers use the network
in the equipment room. Therefore, the test becomes problematic. To guarantee normal test,
engineers must restrict the number of subscribers using the network in equipment room. This
problem must be noticed during optimization.
Figure 1.1 shows the signaling of UE upon a connection failure according to Analyzer
software.
According to RNC signaling, the signals from the cell of SC 121 attenuate sharply. The cell of
SC 56 needs adding to the active set, the UE cannot receive the ActiveSet Update message
from RNC.
Solution
To solve the problem, adjust the SHO parameters to enable the target cell to be added to the
active set earlier as possible. For details, see guidebooks related to call drop analysis.
Solution
After engineers change the average HSDPA throughput per HS-PDSCH code to 300 kbps, 16
subscribers can access the network. The problem is solved.
According to Figure 1.2, the cause of security mode rejection is conflict with already
existing integrity protection and or ciphering information. This cause means that the latest
integrity protection or ciphering information is inconsistent with the configuration.
Figure 1.3 shows the ciphering mode information configured in previous security mode
command.
Figure 1.3 ciphering mode information configured in previous security mode command
According to Figure 1.3, the encryption algorithm configured in the security mode command
is no encryption, namely, no encryption is conducted on the message.
According to previous command, another security mode command is found, as shown in
Figure 1.4.
According to Figure 1.5, there are two encryption methods: UEA1 and no encryption.
According to protocols, there are two encryption methods: encryption and no encryption,
engineers need select the algorithm to be encrypted. Namely, UEA1 is needed here.
By comparison of these two encryption mode commands, they are from different CN domain:
The CN domain No. of the successful command is 4669, while that of the failed one is 4666.
Namely, the CS and PS domain configures different encryption methods.
After the RNC receives the security mode command shown in Figure 1.5, it selects the UEA1
as the encryption algorithm. It then receives the security mode command shown in Figure 1.3
and this command requires no encryption. Therefore, the RNC rejects this security mode
command.
Solution
After setting the CS and PS encryption algorithm to the same, engineers solve the problem
successfully.
RNC, the RNC receives the RRC Connection Request message and responds the RRC
Connection Setup message which the UE fails to receive.
Figure 1.1 shows the signaling of UE upon failure in receiving RRC Connection Setup
message.
Figure 1.1 Signaling of UE upon failure in receiving RRC Connection Setup message
Figure 1.3 shows the normal signal strength upon occurrence of problems.
The messages at IUB interface of NodeB and internal message are normal and without alarms
according to tracing. To further locate problem in the NodeB equipment room, by test, the UE
can sometimes access to the network by antenna where the Ec/Io is about –3 dB, but in rest
time it fails with the same phenomena. The RSCP is about –70 dBm. If the UE moves farther,
the Ec/Io is about –5 dB and the UE cannot connect to the network in a probability of 80%.
The phenomenon is that the UE fails to receive setup message in downlink. According to
detection of NodeB console, the output power of NodeB is 24 dBm, but the normal output
power is 36 dBm. Therefore the power amplifier is problematic.
Solution
After changing the power amplifier, engineers solve the problem successfully.
5.7.2 Abnormal UE
There are abundant phenomena about abnormal UE. The following paragraphs provide an
example.
In Figure 1.2,
Solution
This problem is related to UE performance. There are no more solutions except changing UE.
6 Summary
Compared with V2.0, V3.0 addresses operability. It is closer to on-site engineers and can
guide on-site engineers to solve actual problems during network optimization. Therefore this
guidebook has more updated parts compared with the V2.0 guidebook and addresses the flow
for analyzing problems. It guides engineers to solve problem step by step. The fundamental
knowledge serves as appendix for reference by on-site engineers.
Compared with V3.0, the V3.1 guidebook adds the following content:
RRC connection of HSPDA service
Analysis and cases of admission failure in RAB assignment process
HSDPA-related DT and traffic statistics indexes
Some traffic statistics indexes according to that of RNC 1.6C01B064.
In the following versions, the following content needs adding or updating:
Analyzing of HSUPA access problems
Updating the methods for analyzing traffic statistics and detailed indexes according to
RNC version
Updating the analysis of common problems
7.1.1 Paging by CN
The CN originates paging so that the CN can request UTRAN to connect to UE. The paging
process is the signaling process without connection at IU interface. The CN triggers paging by
sending paging messages. The UTRAN sends the paging message from CN to UE in the
paging process at UU interface so that the paged UE is connected to CN.
Paging messages are sent in non-connection message at IU interface. The RNC sends
PAGING TYPE 1 message on PCCH in the following two situations:
After the RNC receives the paging message from CN, the cell Non Searching
Indication is specified to non-searching (the RNC does not search whether the UE is in
the connected state) in the paging message.
After the RNC receives the paging message from CN, the cell Non Searching
Indication is specified to searching; but the UTRAN cannot find SRNTI (UE in idle
state) by IMSI.
If the paging message at IU interface contains LAI or RAI, the RNC will send the PAGING
TYPE 1 message to all cells in the specified location area or routing area.
If the paging message at IU interface contains no LAI or RAI, the RNC will send the
PAGING TYPE 1 message to all cells under the RNC.
Besides previous situations, the UTRAN sends the PAGING TYPE 2 message on DCCH,
which is called the cooperation paging.
Figure 1.1 shows the flow chart of PAGING TYPE 1 message.
PAGING
RANAP RANAP
According to Figure 1.1, the CN originates paging in a location area which is distributed
under two RNCs. After the RNC receives the paging message, it searches for the matching
cells and calculates the paging occasion. It sends the PAGING TYPE 1 message to the cell on
PCCH at the paging occasion.
CN SRNC UE
PAGING
RANAP RANAP
According to Figure 1.2, if the UTRAN judges that the paging is cooperation paging, the UE
must be in CELL_DCH or CELL_FACH state, so the UTRAN immediately sends the paged
UE the PAGING TYPE 2 message on DCCH.
If the UE is in CELL_PC or URA PCH state, the UTRAN sends the PAGING TYPE 1 to UE.
After the UE receives the PAGING TYPE 1 message, it originates the cell update process to
transit to CELL_FACH state.
Conclusion: If the UE is in CELL_DCH or CELL_FACH state, the network side sends the
PAGING TYPE 2 message. If the UE is in other state, the network side sends the PAGING
TYPE 1 message.
− If one pair matches, this means that UE accepts the paging and forwards the IE CN
domain identity, UE identity, and Paging cause to upper layer.
Otherwise, the UE neglects the paging record.
The UE must monitors the frames (paging occasions) indicated by red dots in each paging
period, and then decode the qth PI. For the calculation of q, see the formula (3).
12 bits (transmission
288 bits for paging indication off)
A 10ms (length) PICH consists of 300 bits (b0, b1… b299). Wherein, the first 288 bits (b0,
b1…b288) are to carry paging indicator. The rest 12 bits is for following use.
Each PICH frame carries NP paging. NP is the number of paging indications per frame. It
defines the maximum paging indicators supported by each frame on PICH. The UE obtains
the value of NP in cell system information. The NP is 18, 36, 72, and 144; namely, the 288
bits are divided by NP, so each division has 288/NP bits. Each division is a paging indicator.
Table 4.1 describes the mapping relationship between {PI0, .., PIN-1} and PICH bits
{b0,..,b287}.
The UE determines by calculating its paging indicator suffix p that p is relevant to the qth bits
of PICH frame.
Np
q PI 18 SFN SFN / 8 SFN / 64 SFN / 512 mod144 mod Np
144
(3)
Wherein,
PI = DRX index mod NP = (IMSI div 8192) mod NP.
SFN is the paging occasion for UE. It is the PCCPCH SFN when PICH is present.
From the formula (3), the UE can know the suffix of PI so that the UE can monitor relevant
bits on PICH only. Once the UE detects that the bits are set to 1, it knows that it is paged. It
starts receiving and decoding paging messages from 7680 chips after completion of PICH
radio frame.
Figure 1.5 shows the time sequence relationship between PICH and SCCPCH.
PICH
The end of PICH radio frame is 7680 chips earlier than associated S-CCPCH frame.
From previous data, each frame of the cell PICH carries 36 PIs. Each PI consists of 8 bits
(288/36). The UE must monitor the bit216 (27x 8) to bit223 of each PICH radio frame. If
these 8 bits changes to 1, the UE knows that it might be paged, so it receives paging message
on SCCPCH.
The UE has two basic operation modes: idle mode and connected mode. When the power is
on, the UE is in idle mode. It is identified by non-access layer identity, such as IMSI, TMSI,
or P-TMSI. The UTRAN does not save the information of UE in idle mode. It pages
respectively the UE that powers on and camps on a cell, or pages all UE in idle mode under
an RNC simultaneously. After UE completes RRC connection setup, UE transits from idle
mode to the CELL_FACH or CELL_DCH state of connected mode. After the RRC
connection is released, the UE transits from connected mode to idle mode.
According to access layer, the access process is the process of UE transiting from idle mode
to connected mode. It includes:
Cell search
Receiving system information broadcast in the cell
Cell selection and reselection
Random access
Once the UE is in connected mode, it can carry out the following non-access layer activities:
PLMN selection and reselection
Location registration
Service application
Authentication
frequency bands within UTRA band to find the proper cell to camp on in the selected
PLMN.
After the UE locks a frequency, it completes cell search by timeslot synchronization, frame
synchronization, and scramble synchronization.
Trigger Time
The UE starts cell selection in the following situations:
The UE powers on
PLMN Selection
The UE obtains the PCCPCH scramble according to 8.1.3. The channel code (SF (=256,1)) of
PCCPCH is known, and it is unique in the whole UTRAN. Therefore the UE can read the
information on the broadcast channel.
First, the UE obtains SFN from system information sent on BCH (PCCPCH). The first
domain of the message is SFNprime. Its value is the initial SFN of the transport block, with
its range (0, 2, 4, 6…4094). The rage of SFNprime is (0…2047) after PER coding. The BCH
TTI is 20ms. It includes two radio frames, so the step of SFNprime is 2.
The scheduling information is known; namely, SIB_POS = 0 and SIB_REP = 8. After the UE
obtains SFN, it can read MIB in the radio frame (SFN = 0, 8, 16…).
After reading MIB, the UE judges according to PLMN identity in MIB whether the current
PLMN is the needed PLMN.
If yes, the UE searches for other SIBs according to the scheduling information of other
SIBs contained in MIB, and obtain their content.
If no, the UE starts cell search from the next frequency.
Table 1.1 lists the parameters and their description in the criterion S.
If the cell meets the criterion S, the UE judges the cell as a suitable cell. Therefore it camps on
the cell, reads other needed system information, and originates location registration.
If the cell does not meet the criterion S, the UE searches for the cell meeting the criterion S in
the neighbor cells of the cell in the following procedures.
synchronization and frame synchronization are needed) according to primary scrambling code
and reference time difference to cell.
In the Cell Selection and Re-selection info for SIB11/12, the UE obtains the following
parameters of neighbor cell:
Maximum allowed UL TX power
Qqualmin
Qrxlevmin
After this, the UE can calculate the Squal and Srxlev of neighbor cells, and judges whether
the neighbor cell meets the previous criterion.
Trigger Time
UE reselects a cell in the following conditions:
Idle mode time trigger (measured quality value of the serving cell is lower than that of
intra-frequency measurement threshold).
In idle mode, the serving cell in continuous Nserv DRX cannot meet the criterion S
(however the system information is configured).
When the UE detects itself in non-service area.
Measurement Rules
The measurement rules when HCS is not used:
If the cell broadcast system information indicates not to use HCS, the UE decides to start the
corresponding measurement. In the CPICH Ec/Io measurement state, the Squal corresponds to
Sx. In CPICH RSCP measurement state, the Srxlev corresponds to Sx.
Intra-frequency measurement
− If Sx > Sintrasearch, the UE need not to start intra-frequency measurement.
− If Sx <= Sintrasearch, the UE starts intra-frequency measurement.
− If the system information does not provide Sintrasearch, the UE always starts intra-
frequency.
Inter-frequency measurement
− If Sx > Sintersearch, the UE need not to start inter-frequency measurement.
− If Sx <= Sintrasearch, the UE starts inter-frequency measurement.
− If the system information does not provide Sintrasearch, the UE always starts inter-
frequency.
Inter-RAT measurement
− If Sx > SsearchRATm, the UE needs not to measurement the system m.
− If Sx <= SsearchRATm, the UE starts inter-RAT measurement on the system m.
− If the system information does provide SsearchRATm, the UE always starts inter-
RAT measurement on the system m.
The measurement rules when HCS is used:
If the cell broadcast system information indicates not to use HCS, the UE decides to start the
corresponding measurement.
For intra-frequency and inter-frequency threshold-based measurement rules
− If Srxlevs <= SsearchHCS or if FDD and Sx <= Sintersearch, the UE measures all inter-
frequency and intra-frequency cells.
− Else the UE measures on all intra-frequency and inter-frequency cells, which have
higher HCS priority level than the serving cell unless measurement rules for fast-
moving UEs are triggered.
− Else, the UE measure on all intra-frequency and inter-frequency cells, which have
equal or higher HCS priority level than the serving cell unless measurement rules for
fast-moving UEs are triggered.
For intra-frequency and inter-frequency measurement rules for fast-moving UEs
If the number of cell reselections during time period TCRmax exceeds NCR, high-mobility
has been detected. In this high-mobility state, UE shall
− Measure intra-frequency and inter-frequency neighbor cells, which have equal or
lower HCS priority than serving cell.
Hs = Qmeas_LEV,s - Qhcss
If it is indicated in system information that HCS is not used, the quality level threshold
criterion H is not applied.
Rs = Qmap,s + Qhysts
Where:
Ln = 0 if HCS_PRIOn = HCS_PRIOs
Ln = 1 if HCS_PRIOn <> HCS_PRIOs
The timer Tn is implemented for each neighbor cell. Tn shall be started from zero when one of
the following conditions becomes true:
Tn for the associated neighbour cell shall be stopped as soon as any of the above conditions
are no longer fulfilled. Any value calculated for TOn is valid only if the associated timer Tn is
still running else TOn shall be set to zero.
At cell-reselection, a timer Tn is stopped only if the corresponding cell is not a neighbour cell
of the new serving cell, or if the criteria given above for starting timer Tn for the
corresponding cell is no longer fulfilled with the parameters of the new serving cell. On cell
re-selection, timer Tn shall be continued to be run for the corresponding cells but the criteria
given above shall be evaluated with parameters broadcast in the new serving cell if the
corresponding cells are neighbors of the new serving cell.
The UE shall perform ranking of all cells that fulfill the S criterion among:
All cells with highest HCS_PRIO meets criterion H, namely, the cells with H > = 0, Note
that this rule is not valid when UE high-mobility is detected.
All cells, not considering HCS priority levels, if no cell fulfil the criterion H >= 0. This
case is also valid when it is indicated in system information that HCS is not used, that is
when serving cell does not belong to a hierarchical cell structure.
In all cases, the UE shall reselect the new cell, only if the following conditions are met:
The new cell is better ranked than the serving cell during a time interval Treselection.
M than 1 second has elapsed since the UE camped on the current serving cell.
Table 1.2 Broadcast parameters and description of cell reselection in system information
Parameter name Parameter description
Qoffset1s,n This specifies the offset between the two cells. It is used for TDD and
GSM cells and for FDD cells in case the quality measure for cell
selection and re-selection is set to CPICH RSCP.
Qoffset2s,n This specifies the offset between the two cells. It is used for FDD cells
in case the quality measure for cell selection and re-selection is set to
CPICH Ec/No.
Qhyst1s This specifies the hysteresis value (Qhyst). It is used for TDD and
GSM cells and for FDD cells in case the quality measure for cell
selection and re-selection is set to CPICH RSCP.
Qhyst2 This specifies the hysteresis value (Qhyst). It is used for FDD cells if
the quality measure for cell selection and re-selection is set to CPICH
Ec/No.
HCS_PRIOs, This specifies the HCS priority level (0-7) for serving cell and
HCS_PRIO neighbor cells.
Qhcss, Qhcsn This specifies the quality threshold levels for applying prioritized
hierarchical cell re-selection.
Qqualmin This specifies the minimum required quality level in the cell in dB. It
is not applicable for TDD cells or GSM cells.
Qrxlevmin This specifies the minimum required RX level in the cell in dBm.
PENALTY_TIME This specifies the time duration for which the
n TEMPORARY_OFFSETn is applied for a neighbor cell.
TEMPORARY_O This specifies the offset applied to the H and R criteria for a neighbor
FFSET1n cell for the duration of PENALTY_TIMEn. It is used for TDD and
GSM cells and for FDD cells in case the quality measure for cell
selection and re-selection is set to CPICH RSCP.
TEMPORARY_O This specifies the offset applied to the H and R criteria for a neighbor
FFSET2n cell for the duration of PENALTY_TIMEn. It is used for FDD cells in
case the quality measure for cell selection and re-selection is set to
CPICH Ec/No.
TCRmax This specifies the duration for evaluating allowed amount of cell
reselection(s).
NCR This specifies the maximum number of cell reselections.
TCRmaxHyst This specifies the additional time period before the UE can revert to
low-mobility measurements.
Treselections This specifies the cell reselection timer value.
SsearchHCS This threshold is used in the measurement rules for cell re-selection
when HCS is used. It specifies the limit for Srxlev in the serving cell
below which the UE shall initiate measurements of all neighbor cells
of the serving cell.
SsearchRAT 1 - This specifies the RAT specific threshold in the serving cell used in
SsearchRAT k the inter-RAT measurement rules.
SHCS,RATm This threshold is used in the measurement rules for cell re-selection
when HCS is used. It specifies the RAT specific threshold in the
serving cell used in the inter-RAT measurement rules.
Sintrasearch This specifies the threshold (in dB) for intra frequency measurements
and for the HCS measurement rules.
Sintersearch This specifies the threshold (in dB) for intra frequency measurements
and for the HCS measurement rules.
Slimit,SearchRATm This threshold is used in the measurement rules for cell re-selection
when HCS is used. It specifies the RAT specific threshold (in dB) in
the serving UTRA cell above which the UE need not perform any
inter-RAT measurements in RAT "m".
Figure 1.2 shows the timing information and acquisition indicator of access timeslot, and the
interval between access timeslots and timeslot number. Whether the information of an access
timeslot in the serving cell is available is decided by upper-layer signaling.
5120 chips
The subscriber can originate random access transmission at the beginning of each access
timeslot. Figure 1.3 shows the structure of random access transmission. The structure includes
message part of 10ms or 20ms.
4096 chips
10 ms (one radio frame)
The preamble length of random access is 4096 chips. It includes a SIGNATURE. The
SIGNATURE is 16 chips and is repeated 256 times. In total there are 16 different
SIGNATURE.
The 10 ms message part radio frame is split into 15 slots, each of length T slot = 2560 chips.
Each slot consists of two parts, a data part to which the RACH transport channel is mapped
and a control part that carries Layer 1 control information. The data and control parts are
transmitted in parallel. A 10 ms message part consists of one message part radio frame, while
a 20 ms message part consists of two consecutive 10 ms message part radio frames. The
message part length is equal to the Transmission Time Interval of the RACH Transport
channel in use. This TTI length is configured by higher layers.
The data part consists of 10*2k bits, where k=0,1,2,3. This corresponds to a spreading factor
of 256, 128, 64, and 32 respectively for the message data part.
The control part consists of 8 known pilot bits to support channel estimation for coherent
detection and 2 TFCI bits. This corresponds to a spreading factor of 256 for the message
control part. The pilot bit pattern is described in 3GPP TS 25.211 table 8. The total number of
TFCI bits in the random-access message is 15*2 = 30. The TFCI of a radio frame indicates the
transport format of the RACH transport channel mapped to the simultaneously transmitted
message part radio frame. In case of a 20 ms PRACH message part, the TFCI is repeated in
the second radio frame.
The downlink AICH is divided into downlink access slots, each access slot is of length 5120
chips. The downlink access slots are time aligned with the P-CCPCH.
The uplink PRACH is divided into uplink access slots, each access slot is of length 5120
chips. Uplink access slot number n is transmitted from the UE p-a chips prior to the reception
of downlink access slot number n, n = 0, 1, …, 14.
Transmission of downlink acquisition indicators may only start at the beginning of a downlink
access slot. Similarly, transmission of uplink RACH preambles and RACH message parts may
only start at the beginning of an uplink access slot.
Figure 1.4 shows the PRACH/AICH timing relation.
Figure 1.4 Timing relation between PRACH and AICH as seen at the UE
The preamble-to-preamble distance p-p shall be larger than or equal to the minimum
preamble-to-preamble distance
p-p,min, i.e. p-p p-p,min.
In addition to p-p,min, the preamble-to-AI distance p-a and preamble-to-message distance p-m
are defined as follows:
When AICH_Transmission_Timing is set to 0, then
− p-p,min = 15360 chips (3 access slots)
− p-a = 7680 chips
− p-m = 15360 chips (3 access slots)
When AICH_Transmission_Timing is set to 1, then
− p-p,min = 20480 chips (4 access slots)
− p-a = 12800 chips
5. If the parameter Commanded Preamble Power exceeds the maximum allowed value, set
the transmit power of preamble to maximum allowed transmit power. If it is lower than
the needed minimum value (prescribed by 3GPP TS 25.101), set the transmit power of
preamble to the current value to be calculated. This value might be larger or smaller than,
equal to Commanded Preamble Power. Otherwise, set the transmit power of preamble to
Commanded Preamble Power. Send the preamble with the selected uplink access
timeslot, signature, and preamble transmit power.
6. The UE waits for confirmation signals corresponding to the used signature from NodeB.
If the UE fails not detect the +1 or –1 acquisition indicator in the downlink access
timeslot which has the same number of uplink access timeslot used by transit preamble
code, it randomly select the next available access timeslot. According to power ramp
step, it increases the Command Preamble Power, deduct the preamble code reset counter
by 1. If the Command Preamble Power is 6 dB larger than the maximum allowed power,
the UE report layer 1 state ("No ack on AICH") to MAC layer, and then quits the
physical random access process. If the retransmission counter value is larger than 0,
repeat the sixth step; otherwise, the UE reports layer 1 state ("No ack on AICH") to
MAC layer, and then quits physical random access process.
7. If the UE receives an –1 acquisition indicator, it reports layer 1 state ("Nack on AICH
received ") to MAC layer, and then quits physical random access process.
8. If the UE receives a +1 acquisition indicator, it sends the random access message part
after 3 or 4 uplink access timeslots before last transmission according to the value of
AICH_Transmission_Timing. The power for sending control part of random access
message is Pp-m higher than the power for sending preamble the last time. For the
transmit power of data part, see protocols.
According to previous operation flow for random access, when the UE accesses the network,
it first sends preamble, and then waits for the confirmation signals from NodeB in the fixed
downlink access timeslot. If the NodeB detects a preamble signal transmitted by UE, the
NodeB responds an acquisition indicator signal on downlink AICH. After sending preamble,
it detects acquisition indicator (AI) signal in the fixed downlink access timeslot. If the UE is
permitted, it keeps sending message part and completes a physical random access. If the UE
fails to receive AI, it keeps repeating the handshaking process of "sending preamble to
detecting AI" for preset times until permitted. Then it sends the message part and completes a
physical random access process. If the UE receives the signal indicating that access is
prohibited, it quits this random access process and reports the state. The message part of
random access carries the sign information of UE, the type of applicated service, and so on.
Table 1.1 describes the relationship among the access subchannel, access timeslot, and SFN.
Table 1.1 Relationship among the access subchannel, access timeslot, and SFN
SFN modulo 8 of Sub-channel number
corresponding P-
CCPCH frame 0 1 2 3 4 5 6 7 8 9 10 11
0 0 1 2 3 4 5 6 7
1 12 13 14 8 9 10 11
2 0 1 2 3 4 5 6 7
3 9 10 11 12 13 14 8
4 6 7 0 1 2 3 4 5
5 8 9 10 11 12 13 14
6 3 4 5 6 7 0 1 2
7 8 9 10 11 12 13 14
Figure 8.2 shows the definition of access timeslot set (taking the uplink and downlink access
timeslot fixed difference p-a = 7680 chips as example).
Figure 8.2 Definition of access timeslot set (taking the uplink and downlink access timeslot fixed
difference p-a 7680 chips as example)
AICH access
slots SFN mod 2 = 0 SFN mod 2 = 1
PRACH
access slots Access slot set 1 Access slot set 2
10 ms 10 ms
After the network side receives the authentication response message, it compares the XRES of
the message with the XRES of authentication quintuple parameters saved in VLR database.
This confirms whether the authentication succeeds. If yes, the flow proceeds. If no, an
abnormal processing flow starts. The flow releases the connection between network side and
UEs, and releases the occupied network resources and radio resources.
After successful authentication, the UE saves CK and IK to USIM card.
Sometimes, after the UE receives the authentication request message, it reports that the
authentication fails. Typical causes of authentication failure include:
When the UE authenticates the network, it checks the AUTN in authentication request
message sent by network side. If the MAC is faulty, the UE sends the authentication
failure message with the cause MAC Failure, as shown in Figure 8.4.
The network side decides according to subscriber identity reported by UE whether to originate
identification process. If the current identity is TMSI (or P-TMSI), the network side originates
identification process which requests UE of reporting IMSI. Then the network side restarts
authentication flow.
When UE detects that the SQN of AUTN message is faulty, the authentication fails with
the cause Synch failure.
Now the VLR removes all authentication quintuple parameters and starts the process of
synchronization with HLR. This process requires HLR to reuse authentication quintuple
parameters and to start authentication process.
The HSUPA load control algorithm described in this chapter is based on UMTS 6.0.
− For the DCH service, the consumption of lub resources equals to service bandwidth
multiplied by the configured activation factor.
− For the HSPA service, the consumption of lub resources equals to GBR of the service
multiplied by the configured activation factor.
The admission principles for lub resources are as follows:
A path of any type can be configured to the total bandwidth of the physical port where
the path is. As a result, the sum of bandwidth of all paths on the same physical port might
exceed the physical bandwidth. For this reason, two levels of admission are needed:
Path-level admission and physical port level admission.
The lub congestion control must apply to both service congestion and bearer congestion.
As for admission, just consider whether lub resources for the related service are
adequate.
The primary path of the service is admitted whenever possible. If this admission fails, try
admission of the secondary path.
The admission threshold varies with admission requests.
The admission requested by the handover has the highest priority, followed by that of new
services and that of reconfiguration in the case of rate increase. The admission threshold is set
to the reserved bandwidth of the lub interface.
In the case of handover, the admission uses 100% of the bandwidth by default. The
reserved bandwidth is 0 kbps. No new parameter is needed.
New services are configured on the basis of total bandwidth minus the reserved
bandwidth for handover.
The admission of a reconfiguration request in the case of rate increase is based on the
congestion threshold.
The preceding admission strategy of lub resources shows that the setting of activation factor
of the BE service greatly impact the number of admitted subscribers. If the activation factor is
set to 1.0, the transmission quality of the subscriber can be guaranteed, but the bandwidth of
lub interface is greatly wasted. If this parameter is set to a small value, the subscriber suffers
packet loss and the bandwidth utilization of the lub interface decreases.
Suppose the following problem occurs in the test: The bandwidth of the lub interface is
adequate to bear a 384 Kbps service. If two 384 kbps PS BE services access the lub interface
after the activation factor is modified, the traffic rate of both services can reach only 64 kbps
when the data source rate is sufficient. The efficient bandwidth usage is only 128 kbps. In this
case, the bandwidth of the lub interface is greatly wasted.
This problem occurs because the RLC retransmits the PDU due to random packet loss in the
lub transmission. In this case, the delay of transmission of the TCP data packet increases, the
TCP flow control is enabled, and the rate decreases. The Overbooking function of the lub
resources is adopted to improve bandwidth utilization of the lub interface. To be specific, this
function improves the bandwidth utilization by setting the activation factor and avoiding
packet loss in the transmission layer.
The admission decision should be made for the Local Cell, the Local Cell Group, and the
NodeB at the same time. The admission is accepted only if all these admissions succeed. The
admission threshold varies with admission requests.
For handover, the admission is accepted only if the CE resources are allocated
successfully.
For a new service, the admission is rejected if the SF of the minimum codes supported
by the rest CE resources after the subscriber accesses exceeds the threshold of handover
reserved SF of the downlink CE resources. If the SF does not exceed the threshold, the
admission succeeds.
In the case of rate increase, the admission is rejected if the SF of the minimum codes
supported by the rest CE resources after the subscriber accesses exceeds the threshold of
congestion reserved SF of the downlink CE resources. If the SF does not exceed the
threshold, the admission succeeds.
11.1.5 LDR
Different types of channels choose LDR actions as listed in the following table:
BE Rate Reduction √
Inter-RAT Handover in CS √
Domain
Inter-RAT Handover in PS √ √ √
Domain
Iu QoS Negotiation √
Code reshuffling √
LDR actions supported by the HSUPA subscribers include inter-frequency load handover and
inter-RAT handover in the PS domain. If primary congestion occurs in uplink of a HSUPA
cell, the HSUPA subscribers can be selected as candidate subscribers together with the DCH
subscribers when the inter-frequency load handover or inter-RAT handover in the PS domain
is executed. The specific principles for subscriber selection remain unchanged.
If the HSUPA cells adopt the ENU based admission, the uplink LDR is also based on the ENU.
11.1.6 OLC
The HSUPA subscribers support only one OLC action: subscriber release. If overload
congestion occurs in uplink of a HSUPA cell, the HSUPA subscribers, together with the DCH
subscribers, are sorted by priority when the subscriber release is executed. Several subscribers
with the lowest priority are chosen for OLC processing.
MAXHSUPAUSERNUM
This parameter refers to the maximum number of subscribers supported by the HSUPA
channel of a cell. In the case of admission of a HSUPA subscriber, the number of subscribers
is check first. If the number of current accessed HSUPA subscribers is smaller than the value
of this parameter, the admission proceeds to the next step. If the number of current HSUPA
subscribers is larger than the value of this parameter, the admission is rejected. If the value of
this parameter is too high, the product is not able to support the accessed subscribers. If the
value of this parameter is too small, the HSUPA subscribers are rejected, though the related
resources are available.
DLHSUPARSVDFACTOR
This parameter reserves resources for the HSUPA downlink control channel. The larger the
value of this parameter is, the more resources are reserved for the HSUPA control channel.
This might cause waste of resources. If the value of this parameter is too small, the QoS of the
HSUPA subscribers is impacted in the case of resource shortage.
UlConvAMRThd
This parameter refers to the uplink threshold of the AMR voice service in conversational
services. It is used for uplink admission of AMR voice subscribers in conversational services.
The uplink admission control algorithm forecasts the load factor of the system after a new call
is accessed based on the load factor of the current system and the service features of the new
call. The algorithm then uses the sum of the forecasted load factor and the uplink load factor
on the common channel as the new forecasted load factor. At last, the algorithm compares the
forecasted load factor with the load factor threshold. If the forecasted load factor does not
exceed the load factor threshold, the call is accepted. If the forecasted load factor exceeds the
threshold, the call is rejected.
The uplink load threshold includes this parameter, UlConvNonAMRThd, UlOtherThd, and
ULHOThd. The relationship between these four parameters can be used to limit the ratio of
sessions to other services in the cell and to ensure priority of access of handover subscribers
and conversational services.
If the value of this parameter is too large, excessive load might exist over the system after
admission of the conversational service. In this case, congestions occur in the system. If the
value of this parameter is too small, subscribers might be rejected though the related resources
are available.
Pay attention to the network planning when setting this parameter, UlConvNonAMRThd,
UlOtherThd, and ULHOThd. If the value of this parameter is too large, the target coverage
is impacted. If the value of this parameter is too small, the target capacity cannot be ensured.
UlConvNonAMRThd
This parameter refers to the uplink threshold of non-AMR voice service in conversational
services. It is used for uplink admission of non-AMR voice subscribers in conversational
services.
The uplink admission control algorithm forecasts the load factor of the system after a new call
is accessed based on the load factor of the current system and the service features of the new
call. The algorithm then uses the sum of the forecasted load factor and the uplink load factor
on the common channel as the new forecasted load factor. At last, the algorithm compares the
forecasted load factor with the load factor threshold. If the forecasted load factor does not
exceed the load factor threshold, the call is accepted. If the forecasted load factor exceeds the
threshold, the call is rejected.
The uplink load threshold includes this parameter, UlConvAMRThd, UlOtherThd, and
ULHOThd. The relationship between these four parameters can be used to limit the ratio of
sessions to other services in the cell and to ensure priority of access of handover subscribers
and conversational services.
If the value of this parameter is too large, excessive load might exist over the system after
admission of the service. In this case, congestions occur in the system. If the value of this
parameter is too small, subscribers might be rejected though the related resources are
available.
Pay attention to the network planning when setting this parameter, UlConvAMRThd,
UlOtherThd, and ULHOThd. If the value of this parameter is too large, the target coverage
is impacted. If the value of this parameter is too small, the target capacity cannot be ensured.
UlOtherThd
This parameter refers to the uplink threshold of services except the conversational services.
This parameter is used for uplink admission of other services.
If the value of this parameter is too large, excessive load might exist over the system after
admission of the service. In this case, congestions occur in the system. If the value of this
parameter is too small, subscribers might be rejected though the related resources are
available.
Pay attention to the network planning when setting this parameter, UlConvAMRThd,
UlOtherThd, and ULHOThd. If the value of this parameter is too large, the target coverage
is impacted. If the value of this parameter is too small, the target capacity cannot be ensured.
ULHOThd
This parameter refers to uplink threshold of handover for uplink admission of handover
subscribers. This parameter applies only to uplink inter-frequency handover.
The uplink admission control algorithm forecasts the load factor of the system after a new call
is accessed based on the load factor of the current system and the service features of the new
call. The algorithm then uses the sum of the forecasted load factor and the uplink load factor
on the common channel as the new forecasted load factor. At last, the algorithm compares the
forecasted load factor with the load factor threshold. If the forecasted load factor does not
exceed the load factor threshold, the call is accepted. If the forecasted load factor exceeds the
threshold, the call is rejected.
The uplink load threshold includes this parameter, UlConvAMRThd, UlConvNonAMRThd,
and UlOtherThd. The relationship between these four parameters can be used to limit the
ratio of sessions to other services in the cell and to ensure priority of access of handover
subscribers and conversational services.
The value of this parameter should be lower than the uplink OLC trigger threshold of the
intelligent load control.
This parameter is used to reserve resources for handover and ensure quality of the handover.
The value of this parameter must exceed the values of UlConvAMRThd and
UlConvNonAMRThd. This parameter applies only to inter-frequency handover.
If the value of this parameter is too large, excessive load might exist over the system after
admission of the service. In this case, congestions occur in the system. If the value of this
parameter is too small, subscribers might be rejected though the related resources are
available. Pay attention to UlConvAMRThd, UlConvNonAMRThd, and UlOtherThd when
setting this parameter.
List of Reference
1. 3GPP R99 TS 24.008 V3.7.0. 2001-03
2. 3GPP R99 25_series. 2002-09
3. URNP-SANA. W-RNO Access Procedure Analysis Guidance 20041101-A-1.0.doc.
2003-05
4. URNP-SANA. W-Paging Procedure Analysis Guidance 20041101-A-1.0.doc. 2003-12
5. URNP-SANA. W-Paging Problem Analysis Guidance 20041101-A-1.0.doc. 2003-12
6. Joint-research team on RNO project. Call Delay Test Report. 2005-12
7. RAN Radio Performance Dept.. WCDMA RAN Radio Performance Call Access Delay
Optimization Test Report. 2004-02
8. RAN Radio Performance Dept.. W-RAN Traffic Statistics analysis and Problem
Location Guidance-20050926-A-1.0.doc. 2005-09
9. RAN Radio Performance Dept.. UMTS Radio Network KPI baseline (V3.3). 2006-01
10. RRNP. WCDMA RNO Network Event Definition Baseline 1.0. 2005-11