Lync video freezes using OCS Relay

[jarrod sousa]

Hello, i have a customer who has integrated Lync 2010 with thier VCSC running X7.2.  the original integration was setup using OCS relay method.  i know that B2BUA is the "correct" way to set this up, but i wasn't ther person who set it up.  my question is simply this:

when a VCS<>Lync call is made, video and audio gets negotiated correctly (h.263 video as there is no AMG) at 1st.  after a random period of time, the video from the VCSC to the lync client will freeze.  i have run packet captures during the call and failure and i see no change in media traffic (video or audio ports).  i have eliminated every vairable i can think of.  i made sure all the clients were on the same Lync cluster, which had a SIP trunk to the VCSC.  no traffic is going accross an edge server or the WAN.  it never fails to happen, but the timing is not consistent.

i have read in other support community posts that using OCS relay can cause freezing video and using B2BUA will solve the problem.  can anyone explain to me how using B2BUA would remedy the freezing video or what is going on in the Lync side to cause this? 

Last thing is all VCSC devices are H323 only, so all the calls are interworked, forcing the RTP stream to go through the VCSC.

thanks!

[Martin Koch]

Hi Jarrod!

I just did not find it in the documentation, but thats a common issue in between lync and the video

systems. That seems to be related to some packet loss and the video does not recover properly.

You should be able to trigger the issue by shortly unplugging and replugging the cable to

one of the involved endpoints.

Its not enough that the VCS is kept in the media path, its important that its the b2bua.

That one will inject some packets and that will help the video to recover.

You might also want to check why you have packet loss.

[jarrod sousa]

Thanks Martin!  i am not sure how easily it will be to track down the packet loss as all the traffic was LAN, i am also pretty sure the customer is not using QoS to prioritize traffic.  if you have any links to docs identifing this issue that would also be extremely helpful.  i have been searching Google for days and can't seem to find any doc/forum stating this.

[khoabui76]

You probably will not find any doc out there that explains how B2bUA addresses the interoperability issue with Lync.  But I can guarantee you that B2BUA does fix the problem.  Just look at B2BUA is the upgrade of OCS Relay.

[Martin Koch]

It is not really fixing the problem, as the problem is the packet loss. :-)

Also the symptom, the freezing might still occur, but its quite fast unfrozen.

Jarrod, to answer our question about a reference, look at the vcs lync deployment guide,

its mentioned in the known limitations section:

http://www.cisco.com/en/US/docs/telepresence/infrastructure/vcs/config_guide/Cisco_VCS_Microsoft_OCS_2007_R2_and_Lync_2010_Deployment_Guide_X7-1.pdf

Lync clients may lose or fail to receive video if B2BUA is not used 
Lync clients require a full video frame to be received before they display a video image. A full frame is sent either when requested by a fast picture update request, or when the video stream is paused and then re-started. Lync, however, does not request fast picture updates. 
A work around has been implemented for MOC/Lync clients where endpoints and the VCS interworking function perform fast picture update requests on behalf of the MOC/Lync – but this is not available for Lync clients; use the VCS B2BUA to work around this; B2BUA performs the Fast Picture Update requests on behalf of MOC/Lync clients. 

[Blackii76]

Hello everyone,

I permit myself to include me into your discussion. I hope you could help me to solve my case.

I have a trouble regarding the video call between Lync 2010 and Video Bridge (MCU). When the dial is establish, We can see 3-5 seconds freezes every couple of minutes and the quality is not really good. I have implemented the B2BUA application and as you mentioned, the problem is the packets loss. (I see the "packets loss" directly on the MCU system)

My enterprise consists of many site worldwide and several MPLS link exists.

My questions are :

Should we determine the bandwidth limitation between Lync and MCU from VCS system to prevent the freezes ? I tried but it's not possible to put a users lync into zone and limit the bandwidth. It's possible with endpoints but no room bridge. I guess the packet loss revolves around the limitation of bandwidth. Maybe the QoS will solve this point ?

The CAC in lync server can prevent the video freezes ? The VGA format is applied but the lync transmission is fixed to 180kbits/s on each bridge communication.

I hope I have expressed myself clearly

Kind Regards,

Hans