Voice: SIP vs H323 thoughts

<relevance>
I guess CCIE level needs understanding of concepts, OEQs seem to
be headed that way, and I always feel bad when something does
not "click" right
</relevance>

<history>
It used to be the case that H323 vs. SIP was kind of Tacacs vs Radius.
And like that case, market has decided the way.

But before the direction was settled, cisco (through documentation and
or course material) that SIP was bound to be more "heavy" on resources
because as it is a text based protocol, you need to do parsing at the
ends. This is from a programming point of view as sound as saying that
OSPF is "heavy" because you have to run the SPF algo. I don't really
buy the argument but it has some grounds.
</history>

<confrontation>
Now, cisco is pushing the SIP way, and in so doing, now SIP performs
better than H.323, even for native H.323 solutions like CVP.
It actually says that CVP servers perform more than 50% better when
using SIP instead of H.323, as denoted by a server supporting 500 ports
when using H.323 but 850 on SIP.
</confrontation>

Any idea why is that so ?
Let alone SIP things like requiring MTPs on trunks sometimes, and
other oddities like having to reinvent routing (CUP supports static
routes only)...

-Carlos

-- 
Carlos G Mendioroz  <tron_at_huapi.ba.ar>  LW7 EQI  Argentina
Blogs and organic groups at http://www.ccie.net