[asterisk-users] VUC: RE: Friday 11th: Aswath Rao: "Trapezoidal VoIP is Evil" on VoIP Users Conference at Noon EDT

[SIP]

Kristian Kielhofner wrote:
> On Thu, Sep 10, 2009 at 2:24 PM, Dean Collins <Dean at cognation.net> wrote:
>   
>> It might just be me but what is trapezoidal voip?
>>
>>
>>
>>
>>
>> Regards,
>>
>> Dean Collins
>> Cognation Inc
>> dean at cognation.net
>> +1-212-203-4357   New York
>> +61-2-9016-5642   (Sydney in-dial).
>> +44-20-3129-6001 (London in-dial).
>>
>>     
>
>   I don't know if I'll be able to make the call but my guess is he's
> referring to the SIP trapezoid:
>
> http://www.iptel.org/sip_trapezoid
>
>   The SIP trapezoid is a concept/teaching tool usually used in
> situations involving a proxy (or multiple proxies) illustrating the
> SIP concept of distinct separation between signaling and media.
> Granted this isn't unique to SIP...  ISDN, SS7, H.323, MGCP, and I'm
> sure others also make this distinction.  IAX is the only protocol I
> know of that doesn't (which is where most of it's NAT advantages come
> from).
>
>   He's probably going to talk about the advantages and disadvantages
> of "the trapezoid" although from the title I'm guessing he's going to
> focus on the disadvantages ;).
>
>   Then again I could be completely wrong.  The SIP trapezoid is real
> but this speculation is purely my own.
>
>   
See... I would say the 'trapezoid' is one of the great strengths of SIP. 
Forcing RTP along the same path as SIP means you can't rely on all those 
incredibly powerful advantages that routers have in pushing packets 
along the best routes they have. You have to rely on the SIP proxy to do 
routing better than a router... except with additional mandatory hops in 
between to hit remote proxies.

Just a bad idea overall. Let the proxies authenticate, since that's what 
they're best at. And let the routers route, since that's what they're 
best at.


N.