[Asterisk-Users] RTP timestamps (was: AArgh, * and the 7960)

[Vic Cross]

On Wed, 19 May 2004, Rich Adamson wrote:

> As you already know, the timestamp problem with cisco phones dropping
> packets has been associated with at least two channels.
> <snip> 
> Although iax & capi are getting hammered, the real issue seems to be 
> that no one has opened a high sev TAC case to fix the root problem.

If I was the one to open the TAC case, the transcript would read:

VC:  My phone is broken: I get no audio.
TAC: Show us a network trace.
VC:  (presents my ethereal traces, with the non-counting RTP timestamps)
TAC: (laughing) NEXT!!!

I have not read RFC1889 (RTP) in detail, but I am positive that the
timestamp field was put there for a reason.  Sure, maybe Cisco is a little
overzealous in the way their code handles non-conformance, but to try and 
put the blame entirely on them is misdirection.  My ATA-186 has problems 
with the same RTP stream.  GIGO.

* needs to generate RTP streams with valid timestamp progression -- surely
we're not happy to say "the Cisco 79x0 is the only phone that cares about
timestamps, so there's the problem".

(Eek, I'm defending Cisco...  Too many beers at lunchtime, maybe.)

Right, I'm fed up with looking like a whinger on this list, and will
whinge no more.  I'm off to try and get to the bottom of this.  Maybe I'll
stay up until 2am (GMT+10) to try and get kram on IRC; maybe I'll decide
that I need the sleep.


Cheers,
Vic Cross