I honestly don't know why anybody would want to do both UDP and TCP.
(In other words, if you set up a TCP flow, I don't see a benefit in also
setting up a UDP flow which you will have to maintain for keep alives).
________________________________
From: sip-boun...@ietf.org [mailto:sip-boun...@ietf.org] On Behalf Of
Vavilapalli Srikanth-A19563
Sent: Friday, August 28, 2009 07:46
To: Audet, Francois (SC100:3055); s...@ietf.org; s...@core3.amsl.com;
sip-implementors@lists.cs.columbia.edu
Subject: Re: [Sip] Question on draft-ietf-sip-outbound-20 draft:
multipleflowcreation
Hi
Thanks for the reply..One more of clarification required..
So If UA wishes to create multiple transport flows with
"sip:primary.example.com"(As specified in draft-ietf-sip-outbound-20#section-13
to honor RFC3261 requirement for message size greater than MTU), would it be OK
it one FLOW lands on Host1 and other FLOW lands on Host2?
i.e for example if "sip:primary.example.com" got resolved into (IP1,
Port1, TCP) -> (IP2, Port2, TCP) -> (IP2, Port2, UDP) -> (IP1, Port1, UDP) in
that order, and if UA wises to create both TCP and UDP FLOWs with
"sip:primary.example.com", then it starts creating TCP FLOW with (IP1, Port1,
TCP) and can create a UDP FLOW with (IP2, Port2, UDP) based on RFC 3263
procedures. Is my understanding correct here?
In the above example, if (IP1) is down/crashed for some reasons, then
"sip:primary.example.com", can never use the TCP FLOW for transporting inbound
messages to that UA (Even though the IP1 and IP2 are state synchronized) until
User Agent detects the FLOW failure with (IP1) and re-create the TCP FLOW with
(IP1 or IP2). Is my understanding correct here?
Please clarify..
Regard
Srikanth
________________________________
From: Francois Audet [mailto:au...@nortel.com]
Sent: Thursday, August 27, 2009 9:22 PM
To: Vavilapalli Srikanth-A19563; s...@ietf.org
Subject: RE: [Sip] Question on draft-ietf-sip-outbound-20 draft:
multiple flowcreation
You would have 2. One for primary, and one for secondary. For each one
of those, you would use normal RFC 3263 procedures for determining which SRV
record to use.
________________________________
From: sip-boun...@ietf.org [mailto:sip-boun...@ietf.org] On
Behalf Of Vavilapalli Srikanth-A19563
Sent: Thursday, August 27, 2009 07:57
To: s...@ietf.org
Subject: [Sip] Question on draft-ietf-sip-outbound-20 draft:
multiple flowcreation
Hi
As specified in section 3.3/3.4 of draft-ietf-sip-outbound-20
draft, If a UA is configured with a outbound-proxy-set with two proxy uris:
"sip:primary.example.com" and "sip:secondary.example.com" And for example,
"sip:primary.example.com" got DNS resolved in to two SRV records "Host1" and
"Host2" and "sip:secondary.example.com" got DNS resolved in to "Host3" and
"Host4", Should UA create four FLOWs, One with each host OR just two FLOWs, one
with each proxy URI(i.e one flow with either Host1 or Host2 and second flow
with Host3 or Host4)?
Please clarify.
Regards
Srikanth
_______________________________________________
Sip-implementors mailing list
Sip-implementors@lists.cs.columbia.edu
https://lists.cs.columbia.edu/cucslists/listinfo/sip-implementors
