Hi there,
   My name is Pete and I heard about P2P SIP at a (SIP Specific, not
vendor-specific) "Advancements in SIP" session at Cisco live this past
week. As someone who has done ridiculous amounts of CallManager
configuration and troubleshooting, coming from a routing and switching
background, the fact that I have to deal with troubleshooting call
routing loops has always bothered me.
   Why should I have to go to clusters of callmanagers or other h.323
or SIP entities, all integrated with each other, and configure what
are essentially static routes on them, pointing number patterns at the
appropriate entities? I asked this question of some people while I was
at TAC and they looked at me like I'd had too much NyQuil. Devices
know what DNs are on them, and call processing entities know what
patterns are configured for access  to local gateways... Why do I need
to go to other entities and configure these patterns etc repeatedly?
Why can't they just tell each other? This has always seemed
sub-optimal and inefficient to me.
   This is my understanding of what P2P SIP is meant to help with or
do, and I would like to contribute thoughts and ideas from the
perspective of someone whose done lots of both voice and routing and
switching design and integration =) ...I'd like to not have to
configure route patterns more than once. In my humble opinion,
afterwards, some sort of routing protocol should take care of telling
other clusters and entities about patterns and devices and RP-ish
things like split-horizon should be used to make sure that we don't
create call routing loops by sending calls back to where they
originated from. (with minor exceptions for certain callflows.)

Nice to meet you all,
   Peter
_______________________________________________
P2PSIP mailing list
[email protected]
https://www.ietf.org/mailman/listinfo/p2psip

Reply via email to