[EMAIL PROTECTED] wrote:
>
> Hi,
>
> The bis drafts discusses stateless proxies that can
> relay incoming messages on TCP to another TCP channel.
> The requirement on the proxy being that it should place
> enough information in the message to be able to forward
> the response on the same TCP connection on which the
> request came in.
Right.
>
> Is the Via header a good choice to push this state
> information?
Yes.
> If I push the state using a proprietary
> parameter, how should I choose the parameter name to
> ensure that it does not conflict with any new Via
> parameters that get defined later? Should I use a
> prefix for the new parameter? Are clients guaranteed
> to retain this parameter in my Via header in the
> response?
No need to go proprietary. This is actually nothing more than the rport
parameter as defined in:
http://www.ietf.org/internet-drafts/draft-ietf-sip-nat-01.txt
along with the received parameter, it provides the IP address and port
of the peer on the other side of the socket. So long as you are not
maintaining parallel connections, thats sufficient to identify which tcp
connection to send the request on.
>
> How strict is the requirement for ensuring that
> responses are dispatched on the same connection?
It is not strict. In order to deal with things like transaction
failover, you need to be prepared to receive responses on a new TCP
connection. We've really tried to separate the transport layer stuff
from the rest of the sip operation.
-Jonathan R.
--
Jonathan D. Rosenberg, Ph.D. 72 Eagle Rock Avenue
Chief Scientist First Floor
dynamicsoft East Hanover, NJ 07936
[EMAIL PROTECTED] FAX: (973) 952-5050
http://www.jdrosen.net PH: (973) 952-5000
http://www.dynamicsoft.com
_______________________________________________
Sip-implementors mailing list
[EMAIL PROTECTED]
http://lists.cs.columbia.edu/mailman/listinfo/sip-implementors
