Sei sulla pagina 1di 31

Preface

In pursuit of the idea First-class service, Customer service, and in order to make more profit and provide better service for our customers, we have edited Maintenance Experience (Issue 252 for CDMA products of ZTE, which includes 14 articles about maintenance cases of CDMA products of ZTE (MSS and ZXC10 3GCN) and special technology topics on common problems troubleshooting. With the wide application of CDMA products of ZTE, routine maintenance of the products, a major job of the technical maintenance personnel of operators, has become more and more important. This issue, which collects articles written by CDMA technical personnel of ZTE, describes practical cases that are often found during the commissioning and maintenance, as well as the solutions. We hope it will be helpful to your routine maintenance work. For more articles of Maintenance Experience and related technical materials, please log on ZTE technical support website http://ensupport.zte.com.cn. If you have any requirement, question, or suggestion, please feel free to contact us. Your attention and support are greatly appreciated.

Maintenance Experience
Bimonthly for CDMA Products No.2 Issue 252, February 2011

Maintenance Experience Editorial Committee


Director: Qiu Weizhao Deputy Director: Huang Dabin Editors: Fang Xi, Wang Zhaozheng, Xu Xinyong, Xiao Shuqing, Zhang Jian, Zhang Jiebin, Zhou Guifeng, Zhao Cen, Zhao Haitao, Jiang Haijun, Xu Zhijun, Huang Ying, Ge Jun, Dong Yemin, Dong Wenbin Technical Senior Editors: Gao Xiaoxia, Chen Zhiliang, Xue Xiaoxing Executive Editor: Chen Lin

Maintenance Experience Editorial Committee ZTE Corporation February, 2011

Maintenance Experience Newsroom


Address: ZTE Plaza, No. 55, Hi-tech Road South, ShenZhen, P.R.China Postal code: 518057 Contact: Ning Jiating Email: doc@zte.com.cn Tel: +86-755-26771195 Fax: +86-755-26772236 Document support mail box: doc@zte.com.cn Technical support website: http://ensupport.zte. com.cn

Contents

Analysis of High Utilization of the IP Address Pool and Low Success Ratio of PPP Connection....... 2 Analysis of A Trunk Circuit Encountering Signaling Blocking After Disaster-Recovery Changeover .. 8 Analysis of Two Outgoing H3 Messages Being Generated During Interception on A Called Subscriber in One Office ..................................................................................................................10 Analysis of A Failure in Suspension and Resumption for Accounting...............................................12 Analysis of DNS Error During the xGW-PDSN Wap Service ............................................................15 Analysis of A Roaming Subscriber Failing to Send A Short Message ..............................................16 Analysis of the Time Recorded in CDR Suddenly About Jumping One Hour ...................................18 Analysis of the SDP Decoding Failure of Incoming SIP Calls from an Interconnected IMS Office .. 22 Analysis of AAA Sending Unrecognizable User Name to PPS .........................................................24 Analysis of the Problem of MSCe Being Directly Interconnected With SCP ....................................27

February 2011

Issue 252

Analysis of High Utilization of the IP Address Pool and Low Success Ratio of PPP Connection
Qian Wei / ZTE Corporation

Fault Phenomenon
Figure 1 shows the alarms appearing in the system.

Figure 1. Alarm Information

Problem Analysis
(1) Single subscriber After performing all-subscriber signaling trace and the failure observation, an engineer found that some subscribers had the signaling messages shown in Table 1.
Table 1. Failure Observation psPPPGetIPRepFromDBS() Err:DB returned IPAddress is: [0],release user. psPPPDown() Err:User's PCF[192.168.50.90],Nego Status[2],Fail Reason[0],blRecvLCPPkt[1],blRecvIPCPPkt [1],ucType[30]. IMSI = 262199506317659 4

IMSI = 262199506317659

From the first row of Table 1, the engineer inferred that the system failed to allocate an address to the terminal. He then observed the signaling trace performed for this subscriber with IMSI 262199506317659, and found that the IPCP process was prior to the process where the PDSN sent a LCP Termination Request message, as shown in Figure 2.

Maintenance Experience

www.zte.com.cn

Figure 2. PDSN initiating LCP Termination Request After the failure of PPP Connection

The engineer checked the last IPCP message. In a process of successfully establishing the PPP connection, the IPCP connection request from the terminal should contain the allocated IP address at the end of the IPCP negotiation process. The IP address, however, was empty in this flow, as shown in Figure 3.

Figure 3. No Valid IP Address Contained in IPCP Connection Request

CDMA Network Products

February 2011

Issue 252

(2) Performance statistics The engineer viewed the previous records about the utilization of the IP address pool of the SMP from the performance statistics of the PDSN OMC, as shown in Figure 4. As shown in Figure 4, the IP-addresspool utilization (represented by the blue

line shown in Figure 4) of CPU 4 of the SMP module reached 100% from 10:00 till 18:00. The engineer viewed the statistical records about the success ratio of the PPP connection in the same day, as shown in Figure 5. As shown in Figure 5, the successful connection ratio (represented by the red line in Figure 5) of CPU 4 of the SMP module declined

Figure 4. Statistics of IP Address Pool Utilization of SMP on September 10, 2010

Figure 5. Statistics on Success Ratio of the PPP Connection on September 8, 2010

Maintenance Experience

www.zte.com.cn

greatly in the same period. The lowest one was about 15%. By comparing the utilization (blue line) of the IP address pool of CPU 4 shown in Figure 4 with the success ratio of the PPP connection of the same CPU shown in Figure 5, the engineer found that the success ratio was 100% when the utilization of the IP address pool was lower than 100% from 10:00 till 18:00. The success ratio was lower than 100% when the utilization of the IP address pool was 100%. The amplitude of the success ratio declining depended on the number of the subscribers attempting to dial a number. The engineer found the same phenomena after observing the similar performance statistical data in other days. Therefore, he determined that the system exception was due to exhaustion of the

IP address pool of CPU 4 of the SMP module. (3) Configuration As shown in Figure 6, the terminals were allocated to different CPUs for processing according to the last two digits of the IMSIs. The terminals with a number ranging from 00 to 49 were allocated to CPU 3, while the terminals with a number ranging from 50 to 99 were allocated to CPU 4. Statistically, the subscriber IMSIs can be regarded as distributed evenly. With the configuration shown in Figure 6, the subscribers could be allocated evenly to two CPUs, hence the configuration was correct.

Figure 6. Configuring the Relationship Between the IMSI and the Module

Figure 7. Configuration of the IP Address Pool

CDMA Network Products

February 2011

Issue 252

As shown in Figure 7, the IP addresses of the two CPUs of the SMP were different in numbers. CPU 3 could use 126 IP addresses while CPU 4 could use only 62 IP addresses. Because subscribers were evenly allocated to two CPUs for processing according to the IMSI, CPU 4 with less IP addresses first encountered the problem of exhaustion of IP addresses. This analysis result is also verified by Table 1 and Figure 2, where the CPU numbers are all 4.

IP address pool of CPU 3 had 60 more IP addresses than that of CPU 4. The purpose of adding new IP address pools was to make two CPUs have a similar number of IP addresses, hence maximizing the utilization of the IP addresses. Figure 8 shows the configuration of new IP address pools. IP address pools ippool3 and ippool4 were added, as shown in Figure 8. After the configuration was updated, CPU 3 and CPU 4 each had 157 IP addresses. After the configuration of the PDSN system was updated, the alarms about high utilization of the IP address pool and the low success ratio of the PPP connection never occurred. The engineer checked the performance statistical data of the system.

Troubleshooting
The engineer added new IP address pools. As mentioned previously, the

Figure 8. PDSN Configuration After Adding New IP Address Pools

Maintenance Experience

www.zte.com.cn

Figure 9. Statistics on Utilization of the IP Address Pool of SMP After Updating the IP-Address-Pool Configuration

As shown in Figure 9, the peak value of the IP address pool utilization was about 50% during the time span from 10:00 to 18:00 in which utilization of the IP address pool had used to reach 100%. As shown in Figure 10, the success ratio of the PPP connection was 100% all day long after

the IP-address-pool configuration of the system was updated. This proves that analysis of the problem was correct. The proposed solution settled the problem of high utilization of IP address pools and low success ratio of the PPP connection.

Figure 10. Statistics on Success Ratio of the PPP Connection After Updating the IP-Address-Pool Configuration

CDMA Network Products

February 2011

Issue 252

Analysis of A Trunk Circuit Encountering Signaling Blocking After DisasterRecovery Changeover


Zhou Fei / ZTE Corporation

Fault Phenomenon
An office adopted the 1+1 disaster recovery mode, but the disaster-recovery changeover was never performed. During a disaster recovery drill, an engineer changed over the services from the active office to the standby office, and then he found that many circuits were running improperly. He queried for these circuits in Daily Maintenance > Dynamic Management, and found that they were in signaling blocking status. He then queried the office direction and found that it was normally in connected status. At that time, all services were improper, and no call could be connected. After all services were changed over to the active office, all the trunk circuits and services recovered.

possible that the LSTP data was wrong? Figure 1 shows the networking diagram. According to the requirements for the disaster recovery drill, the LSTP should be configured as follows.

Figure 1. Networking Model

(1) SCCP layer

MSCeIN1 was directed to office MSCe 1, without backup. MSCeIN2 was directed to office MSCe 2, without backup. MSCeIN2 was directed to office MSCe 3, without backup.

Problem Analysis
The engineer analyzed the configuration, but found nothing unusual. Why did the trunk circuits encounter signaling blocking? The engineer later found that not all services became abnormal. The intraoffice calls could be put through. After viewing the on-the-site networking diagram and the test result, the engineer detected that all the trunk circuits over the LSTP encountered this problem. Was it

(2) MTP3 layer LSTP1MSCe1, MSCe2, and MSCe3

Active route: link A from LSTP1 to active MSCe 1, MSCe 2 and MSCe 3 Alternate route 1: link B from LSTP 1 to MSCe 4 (standby) Alternate route 2: link C from LSTP 1 to LSTP 2

Maintenance Experience

www.zte.com.cn

A signaling link was enabled between LSTP 1 and MSCe 4 (standby). LSTP 1 only sent signaling maintenance messages to the signaling point of MSCe 4 (standby). At any time, the DPC was the signaling point of each active MSCe when LSTP 1 was sending service messages. The local LSTP was not manufactured by ZTE. Before the drill, the engineer confirmed that this device was configured with links A, B, and C by asking the maintenance engineer of this vendor. The engineer then worked together with the engineer of this vendor to analyze the device configuration. Actually, these three links did not adopt the alternative route mode. The networking diagram is shown Figure 2.

If these two routes were obstructed, LSTP would fall all services on MSCe 2. In this case only signaling links A and B were configured, which did not fulfill the requirement. Signaling link B was not configured as the second alternate route to MSCe 1, as shown in Figure 2.

Signaling link A was enabled from LSTP 1 to MSCe 1, and signaling link A was enabled from LSTP 2 to MSCe 1.

Signaling link B was enabled from LSTP 1 to MSCe 2, and signaling link B was enabled from LSTP 2 to MSCe 2.

Figure 2. Networking Model

Troubleshooting
The engineer asked the engineer of the LSTP to configure links A, B, C according to the requirements of ZTE. After that, the engineer implemented the disaster recovery drill again. During the drill, all the trunk circuits over the LSTP and the services were proper.

Signaling link C was enabled between LSTP 1 and LSTP 2.

On the SCCP layer, take LSTP 1 as an example. When all services fell on MSCe 1, the first-selected route was signaling link A, and the second-selected route was from signaling link C to signaling link A.

CDMA Network Products

February 2011

Issue 252

Analysis of Two Outgoing H3 Messages Being Generated During Interception on A Called Subscriber in One Office
Xu Wei / ZTE Corporation

Fault Phenomenon
A lawful interception center reported that two identical interception records were received when the intercepted MS acted as a called party, while only one record was received when this MS acted as a calling party. Figure 1 shows the flow of an SIP WIN call.

Analyzing and Processing


After collecting the BCM and Proxy signaling messages, an engineer found that the SIP-related messages appeared during the call process. Apparently, the call with the called number being intercepted was an intraoffice call, but was a call containing an incoming sub-call and an outgoing subcall. The calling subscriber was a PPS subscriber. This project required that 3GCN was interconnected with PrepaidServer over the SIP protocol to implement the PPS service, which was fundamentally different from the traditional CDMA WIN implementation mode. It was necessary for the product to implement this function. The core network devices should support the SIP protocol for interworking with Prepaid Server, the application server, to implement the intelligent services and value-added services.
Figure 2. Networking Diagram Figure 1. Flow of A SIP WIN Call

Figure 2 shows the networking diagram.

By analyzing the problem, the engineer drew the following conclusion: This symptom is due that a PPS call is equivalent to an outgoing call plus an incoming call. In this case, interception was triggered twice after the called number had been intercepted, because double calls were involved.

10

Maintenance Experience

www.zte.com.cn

Two schemes are available to avoid this situation. (1) When the security department is setting surveillance, implement interception by IMSI instead of MDN. That is to say, set Target Type to 1-imsi rather than 0-mdn, as shown in Table 1.
Table 1. Description of Surveillance Setting Lawful interception number type:0-mdn 1-imsi 2-esn Target Integer: 3-isdn TT N/A Type 0254 4-min 254-all Lawful interception can be triggered normally by different number types.

0: Trace type 1 indicates that interception is triggered only when the intercepted subscriber is a local-office subscriber.

1: Trace type 2 indicates that interception is triggered no matter whether the intercepted subscriber is a local-office subscriber. This mode is used for interception at a gateway office.

Note: Under this solution, the system cannot intercept the called subscriber in an outgoing call and the calling subscriber in an incoming call. The system can only intercept this subscriber in his home office. In this case, the system cannot intercept subscriber B who is a nonlocal-office subscriber when subscriber A calls subscriber B. Subscriber B can only be intercepted in his home office. It is impossible to intercept subscriber B in office A.

Note: This solution requires that the MIN/IMSI number of the subscriber is known to the security department and the corresponding relationship between MDN and IMSI should be provided to the security department in real-time. (2) When the security department is setting surveillance, trace type 1 instead of trace type 2 is used, as shown in Table 2.
Table 2. Description of Surveillance Setting 0: Trace type one. Only under the condition that LI users are local ones, can LI be triggered. 1: Trace type two. 1 LI can be triggered in the situation that LI users are registered locally and remotely. Trace type two is used for LI at GMSCe.

TRACETYPE

Trace type

Integer: 01

CDMA Network Products

11

February 2011

Issue 252

Analysis of A Failure in Suspension and Resumption for Accounting


Li Hua / ZTE Corporation

Fault Phenomenon
After a number segment was added in an office, error code 200051 Fail to get HLR office attribute by user was returned during the suspension and resumption operations performed for a subscriber, indicating that the system failed to get the attribute of the HLR office according to the subscriber.

Analyzing and Processing


An engineer searched for this error in app_20101109.log, the reported accounting log. This accounting log covered the cases of this error code being returned because of illegal MDN subscriber query and by legal subscribers during the suspension/resumption operation. The engineer extracted all the records about this error code being returned by legal subscribers. These records are as follows:

============================================================================== 2010-11-09 10:44:38:326, 11091044383262000<FROM BOSS-SOAP>Mod CUser:MDN=861895 6227086,MSStat=1 2010-11-09 10:44:38:326, <MI Prompt Info>Assigned WorkSpace_Index = 0! 2010-11-09 10:44:38:326, 11091044383262000<SEND DBIO>Mod CUser:MDN=86189562270 86,MSStat=1 2010-11-09 10:44:38:326, 11091044383262000<FROM DBIO-1>ACK:Mod CUser: RETN=200051, DESC=Fail to get hlr office attribute by user; 2010-11-09 10:44:38:326, 11091044383262000<SEND BOSS-SOAP>ACK:Mod CUser: RETN=200051, DESC=Fail to get hlr office attribute by user; 2010-11-09 10:44:38:326, <MI Prompt Info>Released WorkSpace_Index = 0! 2010-11-09 11:03:24:326, 11091103243267600<FROM BOSS-SOAP>Mod CUser:MDN=861895 6209892,MSStat=233 2010-11-09 11:03:24:326, <MI Prompt Info>Assigned WorkSpace_Index = 0! 2010-11-09 11:03:24:326, 11091103243267600<SEND DBIO>Mod CUser:MDN=86189562098 92,MSStat=233 2010-11-09 11:03:24:326, 11091103243267600<FROM DBIO-1>ACK:Mod CUser: RETN=200051, DESC=Fail to get hlr office attribute by user; 2010-11-09 11:03:24:326, 11091103243267600<SEND BOSS-SOAP>ACK:Mod CUser: RETN=200051, DESC=Fail to get hlr office attribute by user; 2010-11-09 11:03:24:326, <MI Prompt Info>Released WorkSpace_Index = 0! ==============================================================================

12

Maintenance Experience

www.zte.com.cn

The system started to record this error at 10:44 and stopped at 11:03, lasting about 20 minutes. The engineer did not find this error in other intervals. Why did this error occur? Based on this thread, the engineer checked whether any unexpected alarms appeared during this interval, but did not find any useful information because the system did not generate any alarms related to this error during this interval.
====================================== 2010-11-09 10:44:35 ======= Begin of DataSync ====== 2010-11-09 10:44:36 RxCommBeginReq:

storage of all data had not been completed until 11:03. This interval matched the interval in which the accounting system returned this error. According to the processing mechanism of the system, each node was not working during data synchronization, no matter synchronization of increment lists or synchronization of all tables. As a result, the DBIO did not perform service provision for both new and old subscribers. Another question is that why the influence of data synchronization for a newly-added number segment had persisted for 20 minutes. The engineer carefully compared the accounting provision log with the synchronization log recorded by DBIO, and found that this error did not last for so long. The start time and the end time of this error being recorded are as follows:
============================== 2010-11-09 10:44:38:326, 11091044383262000<FROM DBIO1>ACK:Mod CUser: RETN=200051, DESC=Fail to get hlr office attribute by user; 2010-11-09 10:44:50:763, 11091044507632400<SEND BOSS-SOAP>ACK:Mod CUser: RETN=200051, DESC=Fail to get hlr office attribute by user; The duration of this record is less than one minute. The start time and the end time of data synchronization being recorded are as follows: 2010-11-09 10:44:35 ====== Begin of DataSync =======

Sender's dbVersion is 256, dbVersionRef is 36! [DBS_P_SSyncRx. c](490) 2010-11-09 10:44:36 RxCommBeginReq:

Receiver's dbVersion is 256, dbVersionRef is 35! [DBS_P_SSyncRx. c](492) 2010-11-09 10:44:36 [DBS_P_SSyncRx.c](527) 2010-11-09 11:03:26 2010-11-09 11:03:26 Saving DataBase Saving DataBase to Path ..\DATA\DATA1 End. to Path ..\DATA\DATA1 Succeeded! ====================================== RxCommBeginReq:

_ODC_S10_MEMUSE changed to 1.

Did the subscriber perform relevant operations? The engineer found the thread from the log file in the trace directory on server 134. The log file was returned from the field. The log file kept the record of data synchronization that was carried out on the background subsystem during this interval. Data synchronization started at 10:44, and

CDMA Network Products

13

February 2011

Issue 252

2010-11-09 10:44:53 End. 2010-11-09 10:44:53 Succeeded! 2010-11-09 10:44:53 RxSaveTblOver: _ODC_S10_MEMUSE changed to 0. [DBS_P_SSyncRx. c](305) 2010-11-09 10:44:53 RxSaveTblOver: Notifying Application of Configuration Changing! [DBS_P_SSyncRx. c](322) ============================== Saving DataBase to Path ..\DATA\DATA1 Saving DataBase to Path ..\DATA\DATA1

The engineer later confirmed that data synchronization had been carried out three times in this interval, resulting in this error being reported during each period of data synchronization. In the interface specification of the version, this error is explained as follows: (1)There are two cases for error code 200051: I. There is no corresponding configuration in the network management system, resulting in failure of acquiring the configuration. II. The network management system is synchronizing data with DBIO, leading to that the user fails to acquire the configuration. (2) For error codes 201310, 210278 and 210183, the instruction should be retransmitted every 10 seconds. (3) For error code 208200, the accounting system needs to retransmit the instruction after adjustment of the node capacity. This interface specification also explains other error codes, which provides helpful guidance on handling of the errors returned by the accounting instructions.

The start times and the end times recorded in these two logs are matched respectively. The accounting provision log shows that this error was generated in three different intervals. Provision was implemented properly from the time when this error was reported for the first time to the time when this error was reported for the second time. Accounting provision was affected in the interval from the time when synchronization started to the time when storage completed.

Summary
When the accounting system returns an error code, preliminarily handle it according to the interface specification, which may reduce the difficulty of troubleshooting and improves the efficiency.

14

Maintenance Experience

www.zte.com.cn

Analysis of DNS Error During the xGWPDSN Wap Service


Wu Zhibin / ZTE Corporation

Fault Phenomenon
A subscriber dialed a number with a wap account. After the xGW-PDSN allocated an IP address, the DNS was displayed abnormally. The subscriber was able to access the wap website. The DNSs of the net service for Internet access over the CDMA network were 202.101.224.68 and 202.101.224.69, and the DNSs of the wap service were 220.192.8.58 and 220.192.32.103. As shown in Figure 1, a public network DNS was allocated for the wap service, resulting in abnormality of accessing web pages through the wap service. Figure 2 shows the onsite networking diagram.
Figure 1. ipconfig/all Displaying after Dialing With a wap Account

Analyzing and Processing


According to the standards, AAA delivers the DNS of the wap service, which is contained in the Radius Access-Accept message. The engineer used LMT to trace the signaling messages of the PDSN during dialing with a wap account, but did not find the DNS parameter in the Radius Access Accept message. The Radius Access Accept message grabbed from the AAA is as shown in Figure 3. The engineer found that this message contained the DNS parameter, but Cisco-AVPair was the private DNS of Cisco. After the engineer modified the AAA configuration and delivered a standard DNS, the problem was solved. The engineer found out the AAA-delivered DNS attribute value in the Radius Access Accept message of the PDSN, as shown in Figure 4. The mobile phone could access Internet properly.
Figure 3. Radius Access Accept Message Figure 2. Networking Model

Figure 4. DNS Attribute Value

CDMA Network Products

15

February 2011

Issue 252

Analysis of A Roaming Subscriber Failing to Send A Short Message


Xue Xiaoxing / ZTE Corporation

Fault Phenomenon
After roaming to the ZTE service zone, a subscriber was unable to send short messages over a control channel. After tracing the signaling messages, an engineer found that the MSCe sent an authentication request message to the home HLRe of the subscriber, and then this HLRe returned a Gerror signaling message. The subscriber failed to send short messages because of authentication failure. The signaling trace result is shown in Figure 1.

Analyzing and Processing


Because this fault was due to rejection from the called HLRe, the engineer first analyzed the configuration of the called HLRe. After analyzing the home HLRe configuration of the subscriber, the engineer confirmed that TermType was a mandatory parameter in an authentication request. The authentication request from the MSCe, however, did not carry this parameter. As a result, the called HLRe considered that the request message did not contain all necessary parameters, and then directly returned a GERROR message. The analysis is as follows: P a r a m e t e r Te r m Ty p e i s e x t r a c t e d f r o m parameter Clsmark2 that the radio side sends to the switching side. If the radio side does not report it to the MSCe, the MSCe is unable to know the terminal type. As shown in Figure 2, the SmsTrans message did not contain parameter Clsmark2 during short message origination.

Figure 1. Signaling Message (1)

Figure 2. Signaling Message (2)

16

Maintenance Experience

www.zte.com.cn

The CmServReq message carried parameter Clsmark2 during voice call origination, as shown in Figure 3. Consequently, the engineer drew the following conclusion: (1) When the caAddsTransfer_T message without carrying parameter btClsmark2 is transmitted for short message origination, the MSCe sends the authentication message without containing parameter termtype. (2) When the caCMServReq_T message carrying parameter tClsmark2 is transmitted for voice call origination, the MSCe sends the authentication request carrying parameter termtype. In the case of whether parameter tClsmark2 is necessary, IOS5.0 protocol specifies that parameter CIT in the CM request message for voice call origination is a mandatory parameter. Parameter CIT in the ADDS Transfer message over a control channel, however, is an optional parameter. In a radio environment, there is no problem when parameter tClsmark2 is not carried during short message origination. Some network

operators explicitly define termtype as an optional parameter for authentication. MAP 41D judges parameter TermType when AC receives an AUTHREQ request message. If the AUTHREQ request message does not carry parameter TermType and AC does not support the TSB51 authentication flow, AC returns Error because of unsupported operation. If the operation is supported, processing continues. For the suggestion of the MSCe compatible with informal processing flows, no protocol requires the MSCe to construct TermType in the event that the radio does not report parameter tClsmark2. The constructed parameter TermType neither conforms to the requirements of the specification nor reflects the real type of the terminal currently accessed. The purpose of requiring a terminal to carry this parameter on each access is for the network side to check its validity thus to block illegal terminals. If the MSCe is allowed to construct TermType at random, checking this parameter is meaningless.

Figure 3. Signaling Message (3)

CDMA Network Products

17

February 2011

Issue 252

Analysis of the Time Recorded in CDR Suddenly About Jumping One Hour
Sun Zuotao / ZTE Corporation

Fault Phenomenon
The time of some conversations in a CDR file generated by an MSCe suddenly jumped about one hour, as shown in Figure 1.

In this CDR file, the answer time and the release time of some conversations were both an hour later than other conversations. This CDR file was named UN10110911407116GCDR. The daylight saving time ended on October 31, 24:00.

Figure 1. CDR File (UN10Files110911407116GCDR)

18

Maintenance Experience

www.zte.com.cn

Analyzing and Processing


The engineer observed the answer time and the release time listed in the CDR files generated before and after CDR file UN10110911407116GCDR, as shown in Figure 2 and Figure 3. UN10110911307115GCDR: UN10110911507117GCDR: The engineer found that all conversation records in CDR file UN10110911307115GCDR were generated at about 10:00, while the majority of the conversation records in CDR file UN10110911507117GCDR were generated at about 11:00, only several records were generated at about 10:00. The engineer determined that the fore-end time jumping resulted in the time jumping of the records in CDR file UN10110911407116GCDR, which were one hour later than the actual conversation time. Why did time jump? The engineer first checked whether the system had used an external clock. If the system used an external clock source such as NTP clock or GPS synchronization clock, the system time varied with the time of the external clock source. After checking the NTP log, the engineer found that the NTP clock was not enabled in the field.
==================================== 2010-10-31 23:40:38 From server: 0.0.0.0, Time before adjust: 0.0, Time after adjust: 0.0, Adjust Fail 2010-10-31 23:50:38 From server: 0.0.0.0, Time before adjust: 0.0, Time after adjust: 0.0, Adjust Fail 2010-11-01 00:00:38 From server: 0.0.0.0, Time before adjust: 0.0, Time after adjust: 0.0, Adjust Fail 2010-11-01 00:10:38 From server: 0.0.0.0, Time before adjust: 0.0, Time after adjust: 0.0, Adjust Fail ==================================== Figure 3. CDR File (UN10110911507117GCDR) Figure 2. CDR File (UN10110911307115GCDR)

CDMA Network Products

19

February 2011

Issue 252

The engineer used the SHOW GPSTIME command on the MML Terminal window to check the GPS time.
============================== NO.129 : No. |Year |Minute 1 11 SHOW GPSTIME; |Result |Month |Day |Hour |Second 2010 36 ------ 17 MSCE -----------

The result shows that the GPS clock of the BSC was used in the field. By analyzing the collected notification messages, the engineer learned that the GPS time of the BSC was quite different from the field time. In this case, the system did not allow automatic time update. This notification message, however, did not appear after November 10. Probably manual update was performed on November 9, resulting in time jumping. Figure 4 shows the notification messages. The engineer checked the operation logs, confirming that manual GPS update was performed at 11:37:28, November 9. The GPS time was also updated manually at 10:04:34, October 29, as shown in Figure 6.

------------------------Operation Success 17 13 49

-----------------------------Rows: 1 ==============================

Figure 4. Notification Messages

Figure 5. Operation Log (1)

Figure 6. Operation Log (2)

20

Maintenance Experience

www.zte.com.cn

Because the local engineer said that the daylight saving time ended on October 31, at 24:00, the problem might be due to that the daylight saving time was not enabled on the BSC but enabled on the MSCe. Therefore, the MSCe was an hour earlier than the BSC. Because the GPS time was updated manually on October 29, 2010, at 10:11:19, the MSCe system time jumped and rolled back an hour, which was recorded in the fore-end log.
====================================== ###:2010-10-29 10:11:19 01008 SCH8, OSS_Config: The Error-value between set-time and OSS-time Larger than 300 sec, after system start 3310122 sec, OSS-currenttime(Greenwich) is 341633479 sec, settime(Greenwich) is 341629879 sec, time-zone is 420, summerTOffset is60. ======================================

This symptom also appeared when the corresponding GPS time was filled in with a standard time instead of a daylight saving time under the condition that the daylight saving time was enabled on the BSC. From the situation verified by the laboratory, this problem was due to that the daylight saving time was not enabled on the BSC.

The MSCe quit the daylight saving time on October 31 and hence the system time rolled back an hour. In this case, the MSCe was an hour later than the BSC. The GPS time was updated manually on November 9, resulting in time jumping of the MSCe again. The MSCe was an hour earlier than before. At this moment, the system time recovered to normal state. The fore-end log also recorded it.
============================== ###:2010-11-09 10:44:10 01008 SCH8, OSS_Config: The Error-value between set-time and OSS-time Larger than 300 sec, after system start 4269648 sec, OSS-currenttime(Greenwich) is 342589450 sec, settime(Greenwich) is 342593049 sec, time-zone is 420, summerTOffset is60. ==============================

Summary
The time jumping of CDR files is due to the time jumping of the fore end. To handle it, focus on the reason why the time jumping occurs on the fore end. In the event when the NTP external time or the GPS external time is enabled, it is necessary to analyze the changes in the external clock. In addition, analyze the problem by viewing the notification messages, fore-end operation logs and the back-end operation logs.

CDMA Network Products

21

February 2011

Issue 252

Analysis of the SDP Decoding Failure of Incoming SIP Calls from an Interconnected IMS Office
Zhu Xuefeng / ZTE Corporation

Fault Phenomenon
Figure 1 shows the networking diagram.

branch=z9hG4bK-4cd11898-84 To: <sip:15348460601@10.244.129.81:600 1; user=phone>;tag=02239895630022e22e2a57b3-247e6d49 From: <sip:85331460@5.0.128.86>;tag=76 8 Call-ID: 1288771736-86@local CSeq: 1 INVITE User-Agent: ZTE-MSCe Content-Length: 0 --------------------------------------

Figure 1. Traffic Model of the Fault

When a fixed network subscriber under IMS calls a CDMA subscriber, a fault occurs at the CDMA end office, prompting SDP decoding failure.

The message prompts sdp decode fail (indicating SDP decoding failure). The problem is due to a mismatch between the coding/decoding type carried by the invite message and that configured at the CDMA end office. (2) Analyze the received invite message with the following content:
-------------------------------------INVITE sip:15348460601@10.244.129.81:6 001;user=phone SIP/2.0 Via: SIP/2.0/UDP 5.0.128.86:5060;branch=z9hG4bK4cd11898-84

Analyzing and Processing


(1) Trace the SIP signaling messages at the CDMA end office. Figure 2 shows the signaling messages. On receipt of the INVITE message from the IMS, the CDMA end office returns a 400-message with the following content:
-----------------------------SIP/2.0 400 sdp decode fail Via: SIP/2.0/UDP 5.0.128.86:5060;

Figure 2. Call Signaling Messages

22

Maintenance Experience

www.zte.com.cn

To: <sip:15348460601@10.244.129.81:600 1;user=phone> From: <sip:85331460@5.0.128.86>;tag=76 8 Call-ID: 1288771736-86@local CSeq: 1 INVITE Max-Forwards: 70 Contact: <sip:85331460@5.0.128.86> Supported: timer Require: 100rel Expires: 330 Allow: INVITE, ACK, BYE, CANCEL, OPTIONS, INFO, PRACK, REFER, SUBSCRIBE, NOTIFY, UPDATE, REGISTER Session-Expires: 1800 Min-SE:90 P-Asserted-Identity: <sip:85331460@5.0 .128.86;user=phone> MIME-Version: 1.0 Content-Type: multipart/mixed;boundary =tekBoundary Content-Length: 445 --tekBoundary Content-Type: application/sdp v=0 o=MxSIP 0 1416567652 IN IP4 5.0.129.4 s=SIP Call c=IN IP4 5.0.129.4 t=0 0 m=audio 20534 RTP/AVP 8 0 18 101 a=rtpmap:8 PCMA/8000 a=rtpmap:0 PCMU/8000 a=rtpmap:18 G729/8000 a=fmtp:18 annexb=no a=rtpmap:101 telephone-event/8000 a=fmtp:101 0-15 a=silenceSupp:off - - - --tekBoundary Content-Type: application/isup; version=itu-t92+

01 00

00 00 00 02

08 81 10 51 43 48 06 06 01 07 81 13 58 33 41 06 0f 00 --tekBoundary-------------------------------

The message shows that there is no space separating the first encapsulated SDP message from the second encapsulated ISUP message. The format of the message encountering the problem is as follows:
-----------------------------a=silenceSupp:off - - - --tekBoundary Content-Type: application/isup; version=itu-t92+ -----------------------------The correct one should be as follows: -----------------------------a=silenceSupp:off - - - --tekBoundary Content-Type: application/isup; version=itu-t92+ ------------------------------

The system will misjudge that the encapsulated ISUP message was also the content of the first encapsulated SDP message, resulting in failure of decoding the SDP message. As a result, the system prompts sdp decode fail. (3) Analysis and verification: Ask the IMS maintenance personnel to cancel encapsulation of the ISUP in the Invite message, and then perform a dialing test. The service becomes proper.

CDMA Network Products

23

February 2011

Issue 252

Analysis of AAA Sending Unrecognizable User Name to PPS


Zhu Wei / ZTE Corporation

Fault Phenomenon
When AAA was interconnected with PPS, the engineer at the PPS side reflected that something was wrong with the user name that AAA sent to PPS. Only the first Request message carried the correct username. The subsequent Request messages carried unrecognizable username.

Analyzing and Processing


An engineer checked the interaction messages between AAA and PPS to locate the problem. The Request message sent to the PPS for the first time carried the correct user name, 0012CFE777BF, as shown in Figure 1. When the subscriber quota reached the threshold, the Request message sent to PPS second time carried unrecognizable user name, as shown in Figure 2.

Figure 1. Interaction Message (1)

Figure 2. Interaction Message (2)

24

Maintenance Experience

www.zte.com.cn

With packet capture, the engineer detected that something was wrong with the user name that AAA sent to PPS. AAA matches user names according to Wimax-Session-ID. If a user name is not matched, AAA generates a random value for attribute User-Name and fills in this value, and then sends it out. The unrecognizable user name probably was due to Wimax-Session-ID. The engineer then analyzed Wimax-Session-ID.

He viewed the first message, that is, the first Request message that AAA sent to PPS, as shown in Figure 3. Figure 3 shows the Wimax-Session-ID parameter that AAA sent to PPS. With packet capture, the engineer found that the parameter that PPS returned to AAA was generated by PPS itself, instead of the parameter that AAA delivered to PPS, as shown in Figure 4.

Figure 3. Interaction Message (3)

Figure 4. Interaction Message (4)

CDMA Network Products

25

February 2011

Issue 252

As shown in the captured packets, PPS delivered a wrong Wimax-Session-ID to AAA that would directly forward it to AGW. When the subscriber quota reached the threshold, AGW sent a Request message containing this wrong Wimax-Session-ID for more quotas. Because AAA did not generate this Wimax-Session-ID, it was unable to match the user name by it. When the user name was not matched, AAA generated a random value, and then sent this value to PPS. This value was the unrecognizable user name seen at the PPS side.

containing the Wimax-Session-ID generated by AAA or or containing this parameter. The problem was solved after the response message did not contain Wimax-Session-ID.

Summary
According to the NGW protocol, parameter Wimax-Session-ID is generated by AAA instead of PPS. In this case, PPS delivered the WimaxSession-ID generated by itself to AAA, leading to that AAA sent an unrecognizable user name to PPS. AAA can handle this case when the response message that PPS sends to AAA either carries the Wimax-Session-ID that AAA sends to PPS or not carry this parameter.

Troubleshooting
PPS should return a message either

26

Maintenance Experience

www.zte.com.cn

Analysis of the Problem of MSCe Being Directly Interconnected With SCP


Chen Jiansheng / ZTE Corporation

Fault Phenomenon
After the MSCe was interconnected with the SCP by using the DPC+SSN addressing mode, the office was accessible, but the MSCe failed to send signaling messages to the SCP during call origination and termination.

The engineer checked the failure observation, as shown in Figure 2. He found that the lower layer discarded the AnlyzdReq message. (2)The engineer continued to trace the SCCP-layer signaling messages and perform the failure observation. An ERROR message was received at the MTP3 layer, as shown in Figure 3. (3)The engineer checked the signaling point connecting the adjacent office, and verified with the opposite end that it was correct. He checked other configurations of the adjacent office, including the

Problem Analysis
(1) The engineer viewed the MAP signaling trace result of the MSCe and the failure observation records of the SPS. Figure 1 shows a call-termination flow, where the MSCe sent an AnlyzdReq message to the SCP, but received an Abort message. The call failed.

Figure 1. Signaling Trace Messages

Figure 2. Failure Observation

Figure 3. Signaling Messages

CDMA Network Products

27

February 2011

Issue 252

network type, adjacent office attributes and protocol type, but did not find any problems. The opposite-end office was accessible. (4)The engineer checked the failure observation records. The first record showed that the failure cause was SPC inaccessible, and the detailed information was SPC :0-4-250,net 1 inaccessible. Why did the system prompt "net 1 inaccessible" since the SPC was correct. The adjacent office configuration showed that the MSCe was interconnected with the SCP by using Net 5. (5)The MTP layer, however, did not find the signaling point route of the SCP in Net 1 when sending the message. As a result, it discarded the message. The adjacent office of the SCP, however, was configured in Net

5. The MSCe did not perform addressing according to the MML configuration. (6)The engineer doubted that whether a security variable affected this flow, because security variables had a higher priority over the MML configuration during the internal processing of the system. Therefore, the engineer searched for the possible security variables. The engineer found security variable SCP network type under Other Function. Its default value is 1, indicating that the network type of the SCP has been set to 1. Therefore, Net=5 in the MML configuration was invalid.

Troubleshooting
The engineer set this security variable to 5, as shown in Figure 4, and the transferred the configuration. After that, the MSCe could properly send the AnlyzdReq message to the SCP. The problem was solved.

Figure 4. Security Variable Setting

28

Maintenance Experience

Potrebbero piacerti anche