Re: [rtcweb] Network times . was SDP Security Descriptions (RFC 4568) and RTCWeb

["Karl Stahl" <[e-mail]>]

 

 

Från: Daniel-Constantin Mierla [mailto:[e-mail]] 
Skickat: den 2 maj 2013 16:48
Till: Karl Stahl; 'Justin Uberti'
Kopia: 'Cullen Jennings (fluffy)'; [e-mail]
Ämne: Re: SV: [rtcweb] Network times . was SDP Security Descriptions (RFC
4568) and RTCWeb

 

 

On 5/2/13 2:45 PM, Karl Stahl wrote:

Rtp relays are fast routing devices that should not add even 1 ms to the
delay. 


By the 'routing device' do you mean some hardware based packet forwarding
(e.g, done in the ethernet card) or an application doing the forwarding in
user space?
[KS] Even without hw acceleration, and in user space, delays in packet
forwarding devices are not significant here. Our small embedded (264 MHz
ARM)  SIP E-SBC can carry 50 simultaneous G.711 calls with 20 ms voice
packets. That is 2500 packets each way, giving us 0.4 ms to handle each
packet – we do not buffer up packets, they are prioritized to go out first -
so that is the magnitude of delay introduced. And our big 3 GHz quad core
x86 machines handles 8000 simultaneous calls. We are talking 10 us per
packet (even in user space, which indeed is slower).

 

So, it not routers and RTP-relays that introduces delays. It is the distance
between them, slow access channels and the jitter buffers needed that
together make up the delays.

/Karl



At least our (Ingates) SBCs add no significant delay when carrying SIP
media.

 

But if there is a TURN server on the other side of the earth, you can get
additional delays of course.

The point I wanted to make was that any new technology should not be
designed for ideal cases, also not for worst cases, but to satisfy the
majority of the possible users.

Daniel

 
-- 
Daniel-Constantin Mierla - http://www.asipto.com
http://twitter.com/#!/miconda - http://www.linkedin.com/in/miconda
Kamailio Advanced Training, San Francisco, USA - June 24-27, 2013
  * http://asipto.com/u/katu *