Jeroen van Bemmel wrote:
Vijay,

Jeroen: Thanks for reading the draft.  More inline.

An initial comment: it seems wasteful to keep all proxies that are
used to discover a path towards the UAS in the path of all signaling,
acting as pass-throughs. Rather, there should be some mechanism to
keep only the proxies that are needed to traverse NAT boundaries
(e.g. those with an outbound flow towards the UAS) I imagine a
two-phase approach: CONNECT as you have, let proxies that "know" they
need to stay in Record-Route, then send a second CONNECT across the
discovered Route

That could be an optimization to consider.  As I see it, there
are three network topologies where this would be useful:

 1) IETF rfc3261-type ecosystems: I suspect that here we'll have
  two, at most three proxies in such ecosystems; so such an
  optimization can be avoided.

 2) IMS-type ecosystems: Given the high number of intermediaries
  that will field a request, the optmization may be helpful
  here.

 3) P2P-type ecosystems: If recursive routing is being used to
  reach the peer, the optimization may be helpful here as well.

Also, I would suggest to make CONNECT a well formed SIP request (possibly with dummy values).

The current draft does not limit the headers that can appear
in a CONNECT request.  From intermediary transaction idenitification
point of view, only a Via is needed.  Dialog usage that requires
To, From and Call-ID is an endpoint function, and in this case,
the endpoint has other means to establish that dialog (i.e., set
up the TLS tunnel and then send a dialog-establishing request over
it.)

That would allow current proxies to participate in phase 1 of CONNECT
without change (although they might Record-Route since CONNECT would
be an unknown method - would need to strip such proxies from the
discovered Route, e.g. by mandating a special tag to use for R-R for
proxies that know what they're doing, removing all other entries)

Would not the current option tag suffice?

Thanks,

- vijay
--
Vijay K. Gurbani, Bell Laboratories, Alcatel-Lucent
2701 Lucent Lane, Rm. 9F-546, Lisle, Illinois 60532 (USA)
Email: [EMAIL PROTECTED],bell-labs.com,acm.org}
WWW:   http://www.alcatel-lucent.com/bell-labs


_______________________________________________
Sip mailing list  https://www1.ietf.org/mailman/listinfo/sip
This list is for NEW development of the core SIP Protocol
Use [EMAIL PROTECTED] for questions on current sip
Use [EMAIL PROTECTED] for new developments on the application of sip

Reply via email to