Please see my answers below.
Thanks,
Anil
-----Original Message-----
From: [EMAIL PROTECTED]
[mailto:[EMAIL PROTECTED] On Behalf Of Dale
Worley
Sent: Thursday, May 19, 2005 8:44 AM
To: [email protected]; [email protected]
Subject: [Sip-implementors] Re: [Sip] About Non-Invite Server
Transaction
> From: Anil Bollineni [EMAIL PROTECTED]
>
> Assume a gateway sends two INVITE's say to two different UAS's with
the
same
> Via branch parameter to two different users at the same time.
It must not do that. See section 8.1.1.7 of RFC 3261: "The branch
parameter value MUST be unique across space and time for all requests
sent
by the UA."
"The sent-by value is used as part of the matching process because
there could be accidental or malicious duplication of branch
parameters from different clients."
I want to talk about NAT firewall. If two UA's behind NAT firewall
generate same branch, and NAT ALG will change the VIA sent-by field to
same value. In this case it could result matching same transaction.
> I understand for retransmitted non-INVITE requests it MUST result
sending
same
> response by UAS
What is the purpose behind that.
i. I understand for retransmitted non-INVITE requests it MUST
result sending same response by UAS.
ii. It is like that to reduce processing retransmitted request
AGAIN by UAS. I have a doubt on two call flows with non-INVITE
server transactions.
I want to clear on this. Which one is true on the above for putting 'J'
timers. Why can't retransmitted request non-INVITE request should be
processed by UAS..
> Assume 200 response to BYE sent by server transaction, and 200
response is
lost
> before it reaches UAC. If it send BYE again, does UAS has to
retransmit
200
> response?
If the second BYE is a retransmission of the first BYE, as you have
noted,
the UAS is required to send the same 200 response.
> During transfer if a REFER is sent to UAS and say server transaction
say
> transmit 503 (Declined), due to some temporary failures (e.g. user
can't
accept
> more calls), and assume this 503 is delayed. Aftr this failure cleared
out, and
> REFER is retransmitted, and in this case the server transaction sends
same
error
> message, even though the user is not busy. First of all, am I
mentioning
the
> correct scenario.? If yes, what could be solution.
If the second REFER is a retransmission of the first REFER, as you have
noted, the UAS is required to send the same 503 response. If the second
REFER is a *new* request (which can be determined by the increased CSeq
number and the different Via branch parameter), the UAS may examine its
situation and return a different response if it chooses.
Dale
_______________________________________________
Sip-implementors mailing list
[email protected]
http://lists.cs.columbia.edu/mailman/listinfo/sip-implementors
_______________________________________________
Sip-implementors mailing list
[email protected]
http://lists.cs.columbia.edu/mailman/listinfo/sip-implementors
