[p2p-sip] Questions on draft-marocco-p2psip-interwork-00
spencer at mcsr-labs.org
Fri Sep 8 00:07:46 EDT 2006
Hi, Enrico and David,
It's probably premature to discuss this draft in detail, but I did have some
questions as I read it, and asking now might help you prepare for the next
The draft calls out I-D.willis-p2psip-concepts specifically, but the
This document describes how user agents registered in P2PSIP overlay
networks [I-D.willis-p2psip-concepts] can interwork with those in
conventional SIP networks [RFC3261]. In particular, no assumption is
made about the overlay but that it allows clients to insert and
retrieve routing information, possibly bound to URIs [RFC3986].
Does "clients" mean "clients" in the I-D.willis-p2psip-concepts sense, or in
some other sense?
I really like the "P2PSIP Proxy Peers" term and description. You may have
noticed some musing in my notes on I-D.willis-p2psip-concepts about a P2PSIP
Overlay Server - I think this is pretty close to what I was thinking of. The
direction of "proxying" is reversed, I think, but this might be
symmetrical... Is it possible that the same concept might also work for the
P2PSIP clients described in I-D.willis-p2psip-concepts?
The Security Considerations section correctly describes the tension between
"too easy to insert into DNS" and "too difficult to insert into DNS" - but
that's awfully late in the draft to bring this up. Although
security-related, it's pretty fundamental to what we're trying to do, I
I'll stop here for now - but thanks for starting this work.
More information about the P2p-sip