Re: [MMUSIC] WG Poll on the middleboxes draft (draft-ietf-mmusic-media-path-middleboxes)

[Flemming Andreasen <[e-mail]>]

Some comments (as an individual)

A lot of the SIP and SDP work started out with a somewhat "purist" 
Internet view of the world, where things such as middleboxes were not 
considered. As other standards and industry organizations became 
interested in using SIP and SDP, different architectures were defined 
and some of those arcitectures did include the notion of middleboxes 
(e.g. 3GPP IMS and CableLabs PacketCable). From an IETF point of view, 
there was, at least initially, not a great deal of knowledge about these 
architectures and what the middleboxes defined by them were doing. 
Conversely, there was (is) also a concern that such middleboxes may 
break end-to-end transparency and hence there was a desire to try and 
alleviate that.

To that effect, it was seen as useful to have a document that could
a) Explain what/how middleboxes might be operating in these architectures
b) Provide guidelines as to how such middleboxes could minimize (ideally 
avoid) impacting end-to-end transparency and/or how protocols could be 
used to try and alleviate any impact such middleboxes might have.

The middleboxes draft is trying to address the above as it relates to 
the media path. The target audience is thus IETF participants that would 
like an overview of how these middleboxes may affect SIP/SDP-signaled 
media streams as well as what can be done to try and overcome that. 
Similarly, the document is targeted at people involved in these "other" 
architecture efforts, with a goal of making it clear how middleboxes may 
affect the operation of SIP/SDP-signaled media streams and hence provide 
some "design principles". This is not unlike some of the NAT work that 
was done in BEHAVE.

It could be argued that these architectures have now been around for so 
long that producing the above document will not make any difference at 
this point. While I have some sympathy for this, I also think we have to 
recognize the importance of documenting what we know and to make that 
readily available for new people that will be working in this space.

In other words, I still believe there is value in pursuing this 
document. As to whether it should be a BCP or Informational, I don't 
have any strong opinions.

Thanks

-- Flemming (as an individual)




On 6/8/12 6:37 PM, Flemming Andreasen wrote:
> Hi
>
> As part of the MMUSIC charter we have the following milestone
>
> Sep 2012 	Submit Considerations for using SDP offer/answer with 
> middleboxes for BCP
>
>
> and we have the middleboxes draft
>
> http://www.ietf.org/id/draft-ietf-mmusic-media-path-middleboxes-04.txt
>
> to address that milestone.
>
> As discussed at IETF 83 (Paris), Hadriel Kaplan raised some concerns 
> with the goal of the document and the potential target as further 
> explained in the following e-mail:
>
> http://www.ietf.org/mail-archive/web/mmusic/current/msg09278.html
>
> A set of (initial) technical comments were also provided by Hadriel in 
> the following e-mail:
>
> http://www.ietf.org/mail-archive/web/mmusic/current/msg08640.html
>
> For now, the chairs would like to focus on the first set of questions 
> above, i.e. what is the goal and potential target audience for the 
> document and is there still value in pursuing it ?
>
>
> The chairs would like to poll the group for opinions and interest in 
> this. Specific areas to consider:
>
> 1) What is the target audience for the document ?
>
> 2) Do people believe that the target audience will read and/or care 
> about this document at this point ?
>
> 3) Should the document be a BCP or Informational ?
>
>
> Thanks
>
> -- Miguel & Flemming (as chairs)
>
>
>
>
>
>
>
>
> _______________________________________________
> mmusic mailing list
> [e-mail]
> https://www.ietf.org/mailman/listinfo/mmusic