PRODUCT_LINE: BCM50 ============================================ TITLE ===== BCM050.R300.FEPS-150 ======================== Release: 3.0 Issue: N/A Category: GEN Superceded By: NONE Release Date: 20090324 Patch Conflict(s): N/A Special Instructions: NO Patch Version: N/A ======================== Software Update Name: BCM050.R300.FEPS-150 Applicable H/W Platforms: BCM50, BCM50a, BCM50e, BCM50b, BCM50ba, BCM50be, SRG50 Applicable S/W Platforms: 6.0.2.05, SRG50 R3 Category: GEN Installation Recommendations: This update should be applied to all new installs of BCM50 R3 and SRG50 3.0. Existing installations should have this update applied if they are experiencing any of the issues that are corrected. There is no need to schedule specific maintenance to apply this upgrade if there are no reported issues. In those cases, systems should be updated at the next regular service opportunity. Component & Version: feps 1.2.282.40.112 iptelprovideragent 1.2.290.37.111 Dependencies: Required Updates: -BCM050.R300.SOFTWARE-MANAGEMENT-32 or later OR -BCM050.R300.SU.System-53.200804 or later Updates should be applied in this order: - BCM050.R300.SOFTWARE-MANAGEMENT-32 or later OR BCM050.R300.SU.System-53.200804 or later - BCM050.R300.FEPS-150 Product Dependencies: None Size: 3.37 MB System Impact: Time to apply approximately 2 minutes Does update application force reboot: No Other Impacts: IP Trunking will stop while the update is being applied Limitations: None Update Removable: No Description: ------------ The following issues are addressed: 1. SIP calls not disconnecting properly The BCM often sends a SIP re-INVITE (inactive media) just as it is tearing down a call. This can cause some third party switches to leave trunks tied up for up to several minutes. Q01938547-01 2. No caller ID on remote IP site on calls transferred on PRI trunk In the case in which the calling name is not known, the BCM would inject 'anonymous' even when the presentation is not restricted. This causes the terminating BCM to display 'anonymous' rather than giving the number as a fallback for unspecified name display. Q01951623 3. Call forward to DID does not work intermittently Sometimes some MCDN payload was being incorrectly decoded, causing the failure. Q01926400-01 4. One VoIP line is always active If calls are being attempted while the VoIP (SIP or H.323) gateway is starting up, some lines are sometimes tied up until the service is restarted. Q01893655-01 5. SIP Re-direction does not work when Outbound Proxy table used If the Outbound Proxy table entries are used, a 302 Moved Temporarily response from the NRS is ignored, and the call cannot be routed. Q01975895 6. No Voice Path on SIP call forwarded to Call Pilot on CS1K BCM does not handle the SDP session version number correctly when the SDP is received in a reliable response during call setup, resulting in no voice path. Q01915068 This update includes the content of the following superseded updates: BCM050.R300.FEPS-143 ---------------------- 1. IP phone can't get through to mobile phone over H.323 because of locate timeout. The issue is resolved by increasing the H.323 timeout. Q01810385-03 2. RFC3407 support is now added to allow the compact form of SDP capabilities to be used for advertising T.38 between endpoints. BCM50 SIP Fax call does not switch to T.38 because that capability is not advertised. Q01920355 3. Spurious MCDN being sent out over SIP in 183 Progress messages in certain circumstances, causing call drops from some non-Nortel switches. Q01914609 4. No speech path when SIP call fwd to CS1K CallPilot When there are multiple answers returned for a single offer, and the version number in the answer is not incremented, it should be ignored. Q01915068 5. One way speech path after second SIP transfer when Secure Media: Best Effort is selected This is experienced because On CS1K the SIP stack does not discriminate and report to the application that have RTP/SAVP media descriptors. Q01943060 6. No speech path on tandem SIP calls to CS1K 5.0 The SIP messages sometimes come in with a different order from previously seen, and are not being handled properly. Q01928517-01 7. No speech path in blind transfer from SipPBX line to CS1K using ISUP This was caused by a timing window in the tandem call that was not handling a new SDP OFFER while setting up the previous media streams, resulting in stale addresses being used. Q01932882 BCM050.R300.FEPS-98 --------------------- 1. BCM sends wrong SIP response code to INVITE when no trunks are available Should send "503 Service Unavailable" instead of "480 Temporarily Unavailable". Q01854111-01 2. SIP MWI transactions don't honor "302 Moved Temporarily" responses Sending an MWI indication to the CS1K using the NRS, it never arrives because the redirect is ignored, and the request is repeatedly re-sent to the NRS instead of the Signaling Server. Q01882153-01 3. BCM calls torn down after 30 seconds if OPTIONS response is not received Certain tandem calls do not return far end capabilities which causes the OPTIONS transaction to hang, and after 30 seconds of no response, the far end is assumed to be unreachable, and the call is torn down. There is now a guard timer set so that an empty response can be sent to avoid this lockup. Q01891602 4. Call from BCM to MCC (Cell) is released after established VoIP gateway (feps) was not handling vdi_alert after a progress was received, and the call was dropped as soon as answered because it was in the wrong state to accept it. Q01889182 5. Interop issue between BCM50R3 SIP and MSFT Exchange for set on SCS with DND set No Speechpath with MSFT exchange server because of redundant re-INVITE immediately after answer. BCM was sending a re-INVITE after the 200 OK (INVITE) with the same parameters as the negotiated parameters. Q01819465 6. Interop issue between BCM50R3 SIP and MSFT Exchange for call routing to Auto Attendant Same root cause as issue 5 above, except with Auto Attendant on the MSFT server. Q01819483 7. T38MaxBitRate was being erroneously sent out as T38FaxMaxBitRate Incoming faxes configured as a VOIP T.38 Fax fail. An incorrect parameter name caused the auto-negotiation to fail. Q01910759 BCM050.R300.FEPS-94 --------------------------------- 1. SIP private calls from MCS to BCM require a public received number BCM did not parse the private dialing numbers if the NPI/TON came only in the INVITE's Request URI. So BCM incorrectly used public received numbers to terminate the calls. Q01447030-01 2. BCM SIP calls forwarded to CS1000 5.5 CallPilot with dead air SIP inbound calls were forwarded (CFNA) to CS1000 5.5 CallPilot over SIP trunks, but the connection to the CallPilot was not established properly. This was due to BCM's implementation limitation, i.e. BCM did not support SIP Update. As a result, BCM was unable to forward the CallPilot media back to the call originators. Q01861580-01 3. SIP 503 Service Unavailable vs 480 Temporarily Unavailable When all VoIP trunks became busy, SIP inbound calls would be rejected by BCM with return code 480 Temporarily Unavailable. The correct return code should be 503 Service Unavailable, which allowed CS1000 to route the SIP calls fallback to PSTN over analog trunks. Q01854111-01 4. BCM tandem calls from VOIP H323 to T1 E&M trunks no ringback heard After H323 inbound calls were successfully routed to T1 E&M answer supervision trunks, CORE sent the call progress vdi messages back to FEPS with progress indicator 2. However, this type of call progress messages was ignored by FEPS. As result, the connection to the inband tone was never established. Q01860845 BCM050.R300.FEPS-46 ------------------- 1. No speech path on the calls from SRG50 to CS1000 MobileX over SIP trunks SRG IP sets called CS1000 MobileX over SIP trunks, the calls were answered by the cell phones of the twin sets, but no speech paths were established. The cause of the problem was that SRG mistakenly attached a duplicated SDP with its PRACK messages in response to CS1000's 180 alerting messages. Q01834016 2. Calls from BCM to SIPPBX over SIP trunks RNA and SIPPBX SIP trunks are not released BCM initiated SIP outbound calls to SIPPBX. The calls were ringing but not answered (RNA). Three minutes later, the Radvision stack timer timed out, the calls were released inside the BCM but no CANCEL messages were sent out to notify the SIPPBX. Therefore, the SIP trunks of the SIPPBX were locked up. Q01813159 3. Calls transferred from analog trunks to SIP trunks experience one-way speech A call from a BCM analog trunk to a BCM analog set was forwarded to a CS1000 IP set, then it was transferred to the second CS1000 IP set which experienced One-way speech path. The cause of the problem was that BCM initially selected codec G.729 for its outbound calls to CS1000, however, BCM revoked its previous decision and selected G.711 when CS1000 attempted to transfer the calls. Therefore, a codec mismatch occurred. Q01829899-01 4. FEPS stops and restarts intermittently This problem occurred when a call was set up with G.729 from a 3.1k Audio port. If a Re-INVITE was sent to the BCM on that port, then in trying to produce an answer to the OFFER, there was a logic error which caused the non-G.711 current codec to be re-offered in a continuous loop. This locked up the gateway, and generally caused a crash. Q01842956 BCM050.R300.FEPS-22 ------------------- 1. Call transfer fails BCM SIP gateway would occasionally fail to process an Attended REFER request correctly. The failure was caused by incorrectly locating and tearing down the call instance to be replaced. Q01771012 2. BCM failed to send BYE to CS1000 A SPS was configured to communicate through UDP with BCM, while TCP with CS1000. Calls were made between BCM and CS1000 via SPS. When a call was released from a BCM set, BCM failed to send the SIP BYE message to CS1000. This was a Radvision SIP stack bug, and the fix was provided by Radvision tech support. Q01800465 3. T.38 Fax calls across SIP trunks could cause trunks to be locked up T.38 Fax calls over SIP trunks could be established successfully. But if the call release was initiated by the far end, FEPS was not able to properly handle the interaction with other components, which caused the SIP trunk to be locked up. This was a timing issue. Q01758242-03 BCM050.R300.FEPS-2 ------------------ 1. T.38 fax parameters not acceptable by CISCO Call Servers FAX calls originated to CISCO Call Servers will fail because T.38 FAX parameters in Nortel T.38 SDP ANSWER are not compatible with the T.38 FAX parameters in the Call Server OFFER. Q01767675