[Live-devel] CPU usage abnormalities

[Cyrille Colin]

If I remember well RTCP network resource usage is RECOMMENDED to be 5%
of session bandwidth (RFC 3550), and is generally less than that if
bandwidth delimiter are present. Session bandwidth can be advertised in
SDP (b=.. in the session subsection), otherwise the client may do its
own calculation, typically when the medium bandwidth is deterministic.

So no idea what this implementation does, but there are usually
algorithms. Maybe you could have a look at them ?

In any case 80% is weird.  Would you have an ethereal trace you could
share ?

Cyrille



-----Original Message-----
From: live-devel-bounces at ns.live555.com
[mailto:live-devel-bounces at ns.live555.com] On Behalf Of Michael Kostukov
Sent: Monday, February 20, 2006 5:16 PM
To: live-devel at ns.live555.com
Subject: [Live-devel] CPU usage abnormalities


Yes, it was definitely RTCP implementation that was causing the 
problems. I disabled RTCP in both client and server (created subsessions

with NULL passed for RTCP instance). The result- 0% CPU on server side 
and 60% CPU usage drop on client side. PLUS - 80% less network traffic 
(!). Most of CPU usage and bandwidth usage was due to RTCP 
implementation doing something really wrong... Am I missing something? 
Is RTCP EVER supposed to use that many resources?

Cheers,

Michael
_______________________________________________
live-devel mailing list
live-devel at lists.live555.com
http://lists.live555.com/mailman/listinfo/live-devel