And may I add that I agree on your "never trust" point of view !
Great, Thanks !
:-)
Den 22 jan. 2020 18:59 skrev Paul Kyzivat <pkyzi...@alum.mit.edu>:
You can find the definitive statement you are looking for in Section 8
of RFC3264:
... When issuing an offer that modifies the session,
the "o=" line of the new SDP MUST be identical to that in the
previous SDP, except that the version in the origin field MUST
increment by one from the previous SDP. If the version in the origin
line does not increment, the SDP MUST be identical to the SDP with
that version number. ...
Now you have a MUST you can point to.
(Nevertheless, in any code that I write I would never trust this!)
Good luck,
Paul
On 1/22/20 11:09 AM, Sundbaum Per-Johan (Telenor Sverige AB) wrote:
> Hi !
>
> I'm in IMS operations and sometimes I get stuck in the crossfire between
> different vendors !
>
> The two SDP's are identical except the version-number, I have checked.
> For the moment it seems that most of our nodes are accepting that version is
> incremented without session has changed in any other way, but
> we have a vendor that are claiming that it's a violation of RFC 4028 -
> Session Timer and also claims that that violation is causing problems, so far
> a bit unclear why and what, could be some kind of misunderstanding, but I
> have also colleagues that are arguing that RFC 4028 is only valid for UA
> with support for timer.
>
> I think your view on the matter makes very good sense !
>
> Thanks !
>
> BR/pj
>
>
> Sensitivity: Internal
>
Sensitivity: Internal
> -----Original Message-----
> From: sip-implementors-boun...@lists.cs.columbia.edu
> <sip-implementors-boun...@lists.cs.columbia.edu> On Behalf Of Paul Kyzivat
> Sent: den 22 januari 2020 16:25
> To: sip-implementors@lists.cs.columbia.edu
> Subject: Re: [Sip-implementors] SDP version
>
> Sorry, but I'm too lazy to compare those two messages byte by byte to see if
> they are identical other than the version number.
>
> I don't understand the point of your question. Are you looking for an excuse
> to increment the version number without checking if the SDP has changed?
>
> There is a fine line here between what is required behavior, what is good
> practice, what is considerate, ...
>
> It really is a savings for the recipient of the SDP if he can avoid parsing
> it. And generally it should be easier for the sender to "know"
> that the SDP is unchanged and so indicate that when sending it. It is a bit
> harder for the recipient to determine that. OTOH, as an implementor I'm not
> very trusting and so am inclined to not trust that it is unchanged without
> checking. Even so there are ways to take advantage of that indication without
> trusting it. (If the version # is the same as prior then do a string
> comparison of the two SDPs to verify. If the version #s are different, then
> simply start parsing the SDP without bothering with the comparison.)
>
> And we shouldn't second guess what optimizations a peer may or may not want
> to do.
>
> I guess the worst case is a situation where you would normally generate the
> SDP from scratch for every message and so not have an easy way to tell if it
> has changed. Then it becomes a question of whether you do the extra work to
> determine changed or not, or force extra work on the recipient. I think you
> should do it.
>
> The session timer connection is, IMO, a red herring. That draft is poorly
> written, in that it seems to imply that an invite for session timer is
> mutually exclusive with one for offer/answer. The two processes share
> messages and can coexist or not.
>
> Thanks,
> Paul
>
> On 1/22/20 6:50 AM, Sundbaum Per-Johan (Telenor Sverige AB) wrote:
>> Sorry for the no good picture !
>>
>> SDP in 200OK for initial INVITE from PBX SDP PDU
>> v=0
>> o=- 195986 1 IN IP4 10.81.253.92
>> s=RFC3264OfferAnswerExchange
>> c=IN IP4 10.81.253.92
>> b=CT:10000000
>> t=0 0
>> m=audio 23812 RTP/AVP 8 110
>> c=IN IP4 10.81.253.92
>> a=rtpmap:8 PCMA/8000
>> a=rtpmap:110 telephone-event/8000
>> a=fmtp:110 0-15
>> a=sendrecv
>> a=ptime:20
>>
>>
>> SDP in re-INVITE from PBX
>> SDP PDU
>> v=0
>> o=- 195986 2 IN IP4 10.81.253.92
>> s=RFC3264OfferAnswerExchange
>> c=IN IP4 10.81.253.92
>> b=CT:10000000
>> t=0 0
>> m=audio 23812 RTP/AVP 8 110
>> c=IN IP4 10.81.253.92
>> a=rtpmap:8 PCMA/8000
>> a=rtpmap:110 telephone-event/8000
>> a=fmtp:110 0-15
>> a=sendrecv
>> a=ptime:20
>>
>> BR/pj
>>
>> From: Rohit Jain <jainrohit...@gmail.com>
>> Sent: den 22 januari 2020 12:36
>> To: Sundbaum Per-Johan (Telenor Sverige AB)
>> <per-johan.sundb...@telenor.se>
>> Cc: sip-implementors@lists.cs.columbia.edu
>> Subject: Re: [Sip-implementors] SDP version
>>
>> Hi,
>>
>> SIP entities (client/server) checks SDP version to determine whether there
>> is any change in SDP from previously received SDP. If the version is changed
>> then they go on to parse the entire SDP to determine the change.
>> Based on that SIP entity can initiate media processing(more relevant in case
>> of servers) or just relay the message without any media level processing. If
>> just the version number is incremented but there is no actual change in
>> SDP, then it will trigger unnecessary processing on the receiver, so it
>> should be avoided.
>>
>> Also can you specify what could be the requirement ("but is that true also
>> for UA without support for timer") in which it is required to increment the
>> SDP version number without any actual change in SDP.
>> PS: Not able to get the attached picture.
>>
>> Regards,
>> Rohit J
>>
>> On Wed, Jan 22, 2020 at 4:31 PM Sundbaum Per-Johan (Telenor Sverige AB)
>> <per-johan.sundb...@telenor.se<mailto:per-johan.sundb...@telenor.se>> wrote:
>> Hi !
>> What do you think about incrementing the version in SDP o-line, but
>> not changing anything else in SDP, should such a behavior be accepted or
>> should it be rejected ?
>>
>> RFC 4028 states that such behavior at least should be avoided, but is that
>> true also for UA without support for timer ?
>>
>> " RFC 4028 - Session Timer
>> In that case, the offer MUST indicate
>> that it has not changed. In the case of SDP, this is accomplished by
>> including the same value for the origin field as did previous SDP
>> messages to its peer. The same is true for an answer exchanged as a
>> result of a session refresh request; if it has not changed, that MUST
>> be indicated. "
>>
>> An example from reality in picture below, should it be accepted ?
>> I think it should be accepted.
>>
>> [cid:image001.png@01D5D11B.59A78290]
>>
>> BR/pj
>>
>>
>>
>> Sensitivity: Internal
>> _______________________________________________
>> Sip-implementors mailing list
>> Sip-implementors@lists.cs.columbia.edu<mailto:Sip-implementors@lists.c
>> s.columbia.edu>
>> https://lists.cs.columbia.edu/mailman/listinfo/sip-implementors
>>
>>
>> --
>> Regards,
>> Rohit Jain
>>
>>
>> Sensitivity: Internal
>> _______________________________________________
>> Sip-implementors mailing list
>> Sip-implementors@lists.cs.columbia.edu
>> https://lists.cs.columbia.edu/mailman/listinfo/sip-implementors
>>
>
> _______________________________________________
> Sip-implementors mailing list
> Sip-implementors@lists.cs.columbia.edu
> https://lists.cs.columbia.edu/mailman/listinfo/sip-implementors
>
_______________________________________________
Sip-implementors mailing list
Sip-implementors@lists.cs.columbia.edu
https://lists.cs.columbia.edu/mailman/listinfo/sip-implementors
_______________________________________________
Sip-implementors mailing list
Sip-implementors@lists.cs.columbia.edu
https://lists.cs.columbia.edu/mailman/listinfo/sip-implementors
