Ok, I've had it with one little blurb in rfc2543bis03.The line that gets quoted "a B2BUA appears as a proxy".
What's it really say?
from the draft
Back-To-Back User Agent: Also known as a B2BUA, this is a logical entity that receives an invitation,and acts as a UAS to process it. In order to determine how the request
should be answered, it acts as a UAC and initiates a call outwards. A B2BUA appears like a proxy, but differs in that it maintains complete call state and must remain in a call. Since it is nothing more than a concatenation of other logical functions, no explicit definitions are needed for its behavior.
THIS DOES NOT MEAN A B2BUA IS A PROXY. IT MEANS IT LOOKS LIKE A PROXY ON A NETWORK SCHEMATIC DRAWING. Calls come in. Calls go out. Looks kind of like a proxy, but it is NOT a proxy and not constrained by the rules of a proxy.
A B2BUA is application logic which associates one or more calls. Note that "one or more" is not "exactly two" -- I actually heard someone say recently that a B2BUA could only deal with two calls at a time, because there are only two "B"s in the acronym. One must be careful when using acronyms these days.
So, what does this application logic do? That's up to the application! See the line in the def that says "Since it is nothing more than a concatenation of other logical functions, no explicit definitions are needed for its behavior." The logical functions are not defined by the SIP speciication, no more than the specification of every program one could write in C++ is defined in the C++standard. One could make a B2BUA that acts like a proxy. Bill Marshall has argued that a proxy is just a special case of a the more general B2BUA. One could also use a B2BUA as a firewall controller, a signaling anonymizer, or many other things.
So why do we have proxies and not just use B2BUAs for everything? Becasuse the limitations on B2BUA that make it a "proxy" are there for a reason -- they provide for network and feature transparency. In other words, proxies provide a way to make it relatively easy to add functionality at the edge of the network and have a reasonable expectation that the new functionality will actually work. Any B2BUA that does not conform to the rules of proxy behavior needs to do so very carefully, and is extremely likely to break some sort of functionality that wasn't considered in the design of that B2BUA. This is DANGEROUS, and non-proxy B2BUAs must be used VERY CAREFULLY, with DISRETION, and only when ABSOLUTELY NECESSARY.
--
Dean
----- Original Message -----
From:
To:
Cc: "Paul Kyzivat"
Sent: Friday, July 13, 2001 8:56 AM
Subject: More B2BUA ranting (was RE: [Sip] Forking: a modest proposal.)
> > -----Original Message-----
> > From: Juan-Carlos.Rojas@alcatel.fr
> >
> > Maybe I'm missing something, but I have the feeling that you
> > are always putting together the session control and the
> > bearer control , which is very likely the case of real user
> > agents, that is end-users, but definitely not the case of
> > servers in the
> > network, and a B2BUA is defined as appearing as a proxy (RFC2543).
> > The splitting of the session layer and the bearer layer is
> > the current trend of Next Generation Networks in the Telecom
> > world, including obviously the UMTS environment.
> >
> > So, when I was talking about a B2BUA providing services
> > (which is still appearing as a proxy) I was not talking about
> > bearer manipulation, codec insertion, etc. Instead, I was
> > talking about services provided on top of the session layer,
> > as e.g. conference,
> > meet-me, follow-me services, etc.
>
> Forgive my agreeing with you, but maybe you *are* missing something.
>
> I'd be very interested in the scenarios under which you beleive
> a B2BUA can provide any value add in the conference and meet-me
> scenarios without any corresponding node in the media path.
>
> I'm also very interested in knowing how acting as a B2BUA instead
> of a proxy when providing a find-me/follow-me service helps in
> any way. Given that using a B2BUA destroys feature transparency,
> the unnecssary insertion in this case seems very short sighted.
>
> /a
>
>
> _______________________________________________
> Sip mailing list http://www.ietf.org/mailman/listinfo/sip
> This list is for NEW development of the core SIP Protocol
> Use sip-implementors@cs.columbia.edu for questions on current sip
> Use sipping@ietf.org for new developments on the application of sip
>
_______________________________________________
Sip mailing list http://www.ietf.org/mailman/listinfo/sip
This list is for NEW development of the core SIP Protocol
Use sip-implementors@cs.columbia.edu for questions on current sip
Use sipping@ietf.org for new developments on the application of sip
沒有留言:
張貼留言