Re: [siprec] Adoption of material fromdraft-ram-siprec-metadata-format-02

Actually we had lot of discussion about this issue (metadata over SIP or over dedicated web service/HTTP) both internally and with partners.

Main advantages  (what I remember ) for SIP approach are as follows:
1. Simplicity, only one interface to implement, troubleshoot, secure and configure between SRC and SRS
2. RS SIP request is self-contained, it includes all relevant info in order to SRS to decide how to do recording . For example participants identity, location. So no need for SRS to correlate different messages from different interfaces
3. Footprint on SRC. For some implementation it is not feasible to have both SIP and HTTP stacks

Leon

-----Original Message-----
From: siprec-bounces@ietf.org [mailto:siprec-bounces@ietf.org] On Behalf Of Paul Kyzivat
Sent: Monday, June 20, 2011 9:00 PM
To: siprec@ietf.org
Subject: Re: [siprec] Adoption of material fromdraft-ram-siprec-metadata-format-02

Charles,

IIRC, the major driving forces were timeliness and synchronicity of metadata with the actual signaling, especially when initially establishing a RS, so that suitable decisions can be made about whether to record a session, etc.

AFAIK the avoidance of an HTTP server in the SRS wasn't a concern. (Its assumed in general to be a very capable box, and probably has an HTTP interface for querying, though that is out of our scope for now.) Avoiding an http client in the src might be slightly more significant, but I don't think that was much of a concern either.

Could we have skinned this cat with an http approach? Probably.
Obviously we had a lot more sip expertise than http expertise in the group. That may have been an influence as well.

	Thanks,
	Paul

On 6/20/2011 11:40 AM, Charles Eckel (eckelcu) wrote:
> Hi John,
>
> One general comment/question I have received from web application 
> savvy coworkers with whom I have discussed the contents of this draft 
> is that the metadata is essentially a document and HTTP already has 
> well defined mechanisms for document sharing and updating, so why not 
> exchange an HTTP address for the document and leverage those 
> mechanisms instead of sending all the metadata within SIP messages. My 
> assumption is that the requirement is to have a SIP based solutions 
> that does not require an HTTP server. Is that accurate?
>
> Thanks,
> Charles
>
>> -----Original Message-----
>> From: siprec-bounces@ietf.org [mailto:siprec-bounces@ietf.org] On
> Behalf Of Elwell, John
>> Sent: Monday, June 20, 2011 4:02 AM
>> To: siprec@ietf.org
>> Subject: [siprec] Adoption of material
> fromdraft-ram-siprec-metadata-format-02
>>
>> Although there are some open issues, to be discussed during 
>> Thursday's
> call, I would like the WG to
>> consider whether material in draft-ram-siprec-metadata-format-02 is
> ready for adoption into draft-
>> ietf-siprec-metadata, i.e., bring it under WG control. In particular,
> between now and the call, please
>> consider whether:
>> - the material is heading roughly in an acceptable direction;
>> - there are any issues that really should be closed before adoption;
>> - there are any aspects of the draft that should not be adopted at
> this time.
>>
>> If, on the call or shortly afterwards, we can agree on adoption, the
> authors of the two drafts would
>> be asked to perform a merger and publish a new version of
> draft-ietf-siprec-metadata before the cut-
>> off for IETF#81 (2011-07-11).
>>
>> John
>> _______________________________________________
>> siprec mailing list
>> siprec@ietf.org
>> https://www.ietf.org/mailman/listinfo/siprec
> _______________________________________________
> siprec mailing list
> siprec@ietf.org
> https://www.ietf.org/mailman/listinfo/siprec
>
_______________________________________________
siprec mailing list
siprec@ietf.org
https://www.ietf.org/mailman/listinfo/siprec