Re: [MMUSIC] I-D Action: draft-ietf-mmusic-media-path-middleboxes-06.txt

[Hadriel Kaplan <[e-mail]>]

On Jan 10, 2013, at 6:02 PM, Dan Wing <[e-mail]> wrote:

> I am glad this document was resurrected.  
> 
> Can the document say SBC instead of the awkward (and inaccurate) 
> "SIP ALG combined with MIDCOM agent"?  Or is saying SBC impolitic.
> 

I don't think it can or should say "SBC" - because this doc does not describe what an SBC does.  At least not the first half (section 4), which is really about an inline ALG in the router/firewall. (ie, it's not the target/source of the IP addresses of the packets, doesn't change SIP/SDP, etc.)

Anyway, I read this latest draft, and it looks exactly like the older ones, so I don't see it fixing the issues with the older ones:
1) It claims to describe the impact on ICE, but doesn't
2) Doesn't describe the issues with forked or moved calls on early in-band media signaling (especially if you actually comply with recommendations #2+#4)
3) Uses terminology from ancient history (ie, RFC 3303) that no one else uses
4) Recommendation #5 is backwards I think - it should recommend protocols not to require/expect to exchange media-plane signaling after the session has been terminated. (in other words, since you can't rely on middleboxes being lenient, don't design a protocol to expect them to be so) 

-hadriel