This document was generated from CDN thread
Created by: Joel Stanfield on 16-09-2010 11:46:35 PM
Our network monitoring has been reporting cases where MXP endpoints, while in calls, are bursting to much higher data rates for video and audio packets. Are there cases were data rates can jump higher than the setup call rate? The monitoring has reported up to 10x the setup call rate and even though of short duration (< 1 minute) it has caused havic with other COS1 traffic. Downspeeding alerts also are reported, apparently due to exceeding our COS1 bandwidth. This may be occuring when a H.239 stream is started.
Subject: RE: MXP call bandwidth bursts
Replied by: John Regan on 24-11-2010 02:22:19 PM
Hi Joel,
Have you looked at this command, and set this to on?
AllowLatency: <On/Off>
If set to Off (default), the IP call rate will sometimes exceed the maximum call rate specified for the call to reduce latency.
In some cases with poor network quality this might however result in increased packet loss. To avoid this, AllowLatency should be set to On.
Example: xconfiguration AllowLatency: On
Regards
John Regan
Product Specialist (Endpoints)
CDO Escalations
Cisco
Subject: RE: MXP call bandwidth bursts
Replied by: Marko Laurits on 13-05-2011 08:41:07 AM
I would always recommend to set AllowLatency: On
With older firmware, when you tried to send attached image as H.239, it caused always problems with slower links. For example, you could have 2 Mb/s DSL-line for Tandberg and your call speed was 384 kb/s then sending the attached images caused so high bursts that the DSL line got overloaded and it resulted in >30% packet loss of dual stream. Of course, the picture was unreadable from other side.
