Hi, I think this is a good start, but I have some comments:
Media-plane stuff: ------------------ This looks rather ok, but I am not sure I understand the retransmission issue. The CSeq will take care of that, won't it? Or, do you mean that retransmission can delay the information to reach the recevier? If so, isn't it covered with the first bullet (low latency)? SUB-NOT vs INFO stuff: ---------------------- For INFO I think we should say that it can be used when the data may be sent is associated to the dialog sent within, and is sent only for the duration of that dialog (unless a media-plane session is more appropriate for that specific data type, of course). For INFO I think we should also add cases when there are dialog-path intermediates (without media access) interested in the data. For SUB-NOT I think we should add text about data which is not associated to an invite dialog, or not associated to the lifetime of an invite dialog. I think that the current SUB-NOT bullets could be applicable to INFO also. I don't think we can forget the multiple dialog issue with SUB-NOT (which is one of the reasons why people use INFO in the first place). Of course that doesn't automatically mean that people are allowed to use INFO for anything they could also use SUB-NOT for, but I see no reason why INFO can't be used e.g. in cases where it is "very likely that some data will be sent" (ISUP is one example, where in many cases nodes will know it will be sent). To me one important difference between INFO and SUB-NOT is, as described earlier, whether the data is associated with a specific invite dialog, and whether the data will only be sent during the lifetime of that invite dialog. If not, SUB-NOT is better. What do you mean by "multiplicity of events containing that data"? Regards, Christer > -----Original Message----- > From: Jonathan Rosenberg [mailto:[EMAIL PROTECTED] > Sent: 13. joulukuuta 2007 7:14 > To: IETF SIP List > Subject: [Sip] proposed litmus tests for INFO > > Based on Hadriels use cases, I'd like to suggest some litmus > tests that one can use to help determine if something is best > done by INFO or something else: > > Cases where a media-plane session is the best solution: > > * there are requirements for low latency delivery (camera > control for > example) > > * there are requirements that the data shouldn't be > retransmitted since latency trumps reliability (coordinate > updates for a pointer on a whiteboard or something like that) > > * the data consists of a large number of messages that are > likely to be sent during the duration of the session > > * the data is potentially very large > > Cases where SUB-NOT is the best solution: > > * the entity that wants to receive the data knows with > high probability that the other side is likely to send a > notification with the data, and it knows specifically what > type of data it needs to receive > > * the entity that wants to receive the data is likely to > receive a multiplicity of events containing that data > > Cases where INFO is the best solution: > > * the recipient of the data has no idea if the sender will > ever generate the data; the recipient is just saying that it > knows how to handle it, should the sender have something to send > > * the recipient could potentially handle many different > types of data with the above characteristic > > * the size of the data that is to be sent is small and > infrequent, at most a few times during the session > > * the data needs to be sent reliably and latency is not > much of an issue > > * the data has no impact on SIP dialog state, transaction > state, media session states or intermediate proxy or b2bua states > > > -Jonathan R. > > -- > Jonathan D. Rosenberg, Ph.D. 499 Thornall St. > Cisco Fellow Edison, NJ 08837 > Cisco, Voice Technology Group > [EMAIL PROTECTED] > http://www.jdrosen.net <http://www.jdrosen.net/> > PHONE: (408) 902-3084 > http://www.cisco.com <http://www.cisco.com/> > > > _______________________________________________ > Sip mailing list https://www1.ietf.org/mailman/listinfo/sip > This list is for NEW development of the core SIP Protocol > Use [EMAIL PROTECTED] for questions on current sip > Use [EMAIL PROTECTED] for new developments on the application of sip > _______________________________________________ Sip mailing list https://www1.ietf.org/mailman/listinfo/sip This list is for NEW development of the core SIP Protocol Use [EMAIL PROTECTED] for questions on current sip Use [EMAIL PROTECTED] for new developments on the application of sip
