[CUWiN-Dev] IIT VoIP lab - requests?

[Aaron Huslage]

I agree with this. The flows from port 5060 are only signalling and  
QoS on them is much less important than the RTP streams. The only  
"one port" QoS that would be useful in the VoIP world is marking IAX  
packets, since the media and signaling are carried over the same ports.

--
Aaron Huslage
www.inveneo.org
Mobile: 1.503.860.1634
Office: 1.415.901.1969 x1245


On Feb 26, 2006, at 12:09 AM, William Waites wrote:

> On Fri, Feb 24, 2006 at 01:11:21PM -0600, Paul Smith wrote:
>>
>> We spoke yesterday and decided to focus on two variables for  
>> testing: one is
>> varying TOS in the IP header to prioritize SIP protocol packets  
>> over say
>> HTTP, and the other is setting up a multi-hop network in the lab  
>> and seeing
>> what happens to a SIP session when a node goes down.
>
> Hi Paul. Greetings from Jo'burg!
>
> Just a note about diffserv and VoIP.
>
> Strictly speaking, any type of diffserv should be applied to the
> RTP packets, not the SIP packets. In an environment where you don't
> have particular control over the client software (i.e. outside the  
> lab)
> it might be necessary to re-write the headers to set a known TOS value
> at the ingress gateway. But you can't just, say, match on port 5060
> since the RTP streams will use random UDP ports and it's these flows
> that we really care about.
>
> -w
> _______________________________________________
> CU-Wireless-Dev mailing list
> CU-Wireless-Dev at lists.cuwireless.net
> http://lists.chambana.net/cgi-bin/listinfo/cu-wireless-dev