If I recall correctly, it was an optimization. Section 16 discusses the switch from TCP to UDP. It basically requires the proxy to become transaction stateful; see last sentence of the following snippet.
In some circumstances, a proxy MAY forward requests using stateful transports (such as TCP) without being transaction-stateful. For instance, a proxy MAY forward a request from one TCP connection to another transaction statelessly as long as it places enough information in the message to be able to forward the response down the same connection the request arrived on. Requests forwarded between different types of transports where the proxy's TU must take an active role in ensuring reliable delivery on one of the transports MUST be forwarded transaction statefully. > -----Original Message----- > From: [email protected] [mailto:sip- > [email protected]] On Behalf Of Aaron Clauson > Sent: Monday, June 21, 2010 9:27 AM > To: [email protected] > Subject: Re: [Sip-implementors] Retransmits on reliable transports? > > > -----Original Message----- > > From: Brett Tate [mailto:[email protected]] > > > > If I recall correctly... requests and 3xx-6xx responses are not > > retried; however 1xx-2xx responses expecting ACK or PRACK are > retried. > > Ok understood. Next questions. > > What's the logic behind that? Why would some responses be retransmitted > and > not others since getting an error response through is as important as > getting a 2xx response through not to mention requests? What mechanism > should be used to ensure reliability when a TCP UA knows its outbound > proxy > is stateless and uses UDP for the next hop? _______________________________________________ Sip-implementors mailing list [email protected] https://lists.cs.columbia.edu/cucslists/listinfo/sip-implementors
