Cisco Support Community
cancel
Showing results for 
Search instead for 
Did you mean: 
Announcements

Welcome to Cisco Support Community. We would love to have your feedback.

For an introduction to the new site, click here. And see here for current known issues.

New Member

VCS QoS DSCP markings question.

I want to set the  audio/video/signalling traffic in our VCS (Video Communication Server) to the values recommended in the Medianet 4.0 QoS SRND.  The recommendations are to mark interactive video and audio the same at CS4, and signalling at CS3.

However when I web into the VCS management app, there are only two drop down boxes for QoS/DSCP markings. One box is to enable 'media' DSCP or not. The other box is to enter in the DSCP value.

I'm assuming that 'media' is the video and audio.

Three questions:

1.  Does 'media' mean both video and audio packets?

2.  Any reason why signalling isn't part of the DSCP marking functions of the VCS - or is that configuration point somewhere else in the GUI?

3.  If there is no way for the VCS to mark signalling packets, then is it up to the next layer (the switch) to mark the signalling packets?

1 ACCEPTED SOLUTION

Accepted Solutions
Gold

VCS QoS DSCP markings question.

jkeefee,

I understand where you are coming from, and feature requests already exist for the VCS for enhancing the QoS configuration possibilities, including different tagging for audio, video and signaling and interface-, zone- and subzone-specific tagging as well.

Since a wide range of feature requests exists for the VCS (as for most of our products), it is always beneficial to put as much weight as possible behind each requests, since the requests which have the most customer demand are the ones which are likely to be implemented and improved first.

I would therefore advise you to reach out to your Cisco account manager to flag your need for more enhanced QoS capabilities on the VCS, so that this can be forwarded on to the VCS product management.

Hope this helps,

Andreas

5 REPLIES
Gold

VCS QoS DSCP markings question.

Hi,

what VCS software are you running?

On X7.0.3 (Which was just released), the System > Quality of Service page on the VCS allows you to set a "QoS mode"  of either "None" or "DiffServ". When set to "DiffServ" you are allowed to enter a DiffServ tag value.

The informational text for the "QoS mode" setting states the following (Which you can see by pressing the informational "i" symbol just right of the drop-down menu itself:

"Specifies the type of QoS (Quality of Service) tags to apply to all signaling and media packets.


None

: no specific QoS tagging is applied.


DiffServ

: puts the specified Tag value in the TOS (Type Of Service) field of the IPv4 header or TC (Traffic Class) field of the IPv6 header."

As the informational text states, both signaling and media packets will be tagged with DiffServ. The VCS currently does not allow for separate tagging for signaling, audio or video.

Hope this helps,

Andreas

New Member

VCS QoS DSCP markings question.

We are running x7.0.2. What you wrote is exactly what I saw when I went into the GUI.  I'm very surprised that the VCS does not allow for separate markings for signaling, audio, and video. All VC endpoints allow for this and Cisco's QOS SRND recommends to separately mark signaling and video.

Therefor to correctly mark the voice/video traffic EF to be placed in a priority queue, then signaling will also be marked EF and placed in the prioity queue, unless I create an ACL to deny it.  That's going to be quite the hassle to do.

Any reason why Cisco has done it this way?

New Member

VCS QoS DSCP markings question.

If I'm reading this correctly - this would apply specifically to packets coming from the VCS itself...not endpoints.

To be consistent with the rest of Cisco voice (CUCM) - voice signaling is typically CS3/AF31 which is DSCP26

As for endpoints:

I have several EX90's & C20's on my network running TC5.0 code, and from the GUI of the endpoint I see separate fields for audio/data/signaling/media.

For testing:

If you wireshark an endpoint while in an active call (if my memory serves me correctly) you'll see 2 streams - one for audio and one for video + sip signaling to/from the VCS (or CUCM if that's where it is registered)

I've seen a QoS design guide for telepresence that states pretty clearly the qos markings should be the same for the audio & video - otherwise you run the risk of having different queing mechanisms and drop characteristics for the 2 different streams and could end up with unexpecting behavior when QoS policies kick in. 

at best - you might end up with a/v sync issues that would come and go depending on what was happening with QoS and whether or not there was saturation and QoS policies were kicking in.

I have also seen that endpoints are fine while in a call - but a soon as we drop a call - the endpoint goes unregistered, and I suspect that its because while in an HD call...the sip signaling (tcp 5060) cant get through and registration timers expire.  marking the signaling properly helps this.

Paul

New Member

VCS QoS DSCP markings question.

pdinapoli - I've done extensive wireshark testing to insure voice/video/signaling packets retain their markings throughout our enterprise. It is only with the addition of this VCS where there is conflict now with Cisco's recommendation for marking signaling as CS3 - differently from voice and video.

I understand that voice and video are to be marked the same - CS4 - and placed in the same priority queue for exactly the reason you mention above.  But not only does the VCS mark voice and video the same - but now also signaling! All the Cisco references concerning Telepresence state to mark signaling packets CS3 and video as CS4 and to put signaling into a CBWFQ; whereas the video/voice will be placed in the priority queue.

So we're back to the question as to why the VCS forces one to mark not only video and voice, but also signaling, the same.  It's an annomoly for sure.

Gold

VCS QoS DSCP markings question.

jkeefee,

I understand where you are coming from, and feature requests already exist for the VCS for enhancing the QoS configuration possibilities, including different tagging for audio, video and signaling and interface-, zone- and subzone-specific tagging as well.

Since a wide range of feature requests exists for the VCS (as for most of our products), it is always beneficial to put as much weight as possible behind each requests, since the requests which have the most customer demand are the ones which are likely to be implemented and improved first.

I would therefore advise you to reach out to your Cisco account manager to flag your need for more enhanced QoS capabilities on the VCS, so that this can be forwarded on to the VCS product management.

Hope this helps,

Andreas

2498
Views
0
Helpful
5
Replies
CreatePlease login to create content