02-10-2003 04:36 AM
In order to measure delay in a VOIP architecture, I use Cisco-Voice-Dial-Control-Mib. Does anyone knows how the VoIPCallActiveRoundTripDelay Mib is calculated?
When I use a small client-server program (udpSpeed) to simulate an audio communication on the IP network, the RTT is in accordance with previsions, even it the network is loaded (I use crtp and ppp multilink). But when I make the same test with an audio communication between two gateways (using dial-peer), the RTT is really too high. The audio packet are well marked with dscp ef flags. I guess there is something wrong in the dial-peer, but I do not know what (one interface is FXO, the other is FXS).
Thanks
02-14-2003 10:59 AM
The description for the MIB VoIPCallActiveRoundTripDelay is given in the following URL:
ftp://ftp.cisco.com/pub/mibs/v1/CISCO-VOICE-DIAL-CONTROL-MIB-V1SMI.my
Use Ctlr-F for searching the particular term VoIPCallActiveRoundTripDelay and you get the description for it.
As far as the RTT values are considered, I would suggest you to check through the dial-peer configurations properly.
Hope this helps.
02-21-2003 02:36 AM
Thanks.
Seems to me this RTT is not measured from timestamps of EF packets, but is measured from control & signalization packets. Thoses are AF31 and are not placed in priority queues. One way to obtain right RTT is to mark thoses packets as EF.
Discover and save your favorite ideas. Come back to expert answers, step-by-step guides, recent topics, and more.
New here? Get started with these tips. How to use Community New member guide