Re: [Speermint] SIP Interconnect draft

This seems a reasonable approach.

John 

> -----Original Message-----
> From: speermint-bounces@ietf.org 
> [mailto:speermint-bounces@ietf.org] On Behalf Of David Hancock
> Sent: 04 December 2009 18:43
> To: speermint@ietf.org
> Subject: [Speermint] SIP Interconnect draft - requirement entity
> 
> I would like to start addressing the comments received during
> presentation of draft-hancock-sip-interconnect-guidelines-02 at the
> IETF#76 SPEERMINT meeting.  This email discusses the comment 
> received on
> the first issue. 
> 
> Here is the issue:
> What entity is used as the subject of normative statements in the
> document?
> 
> Comments recorded in the meeting minutes:
> JON PETERSON on Requirement Entity (slide 5): Make sure that the
> interconnect spec describes what the SIP traffic between domains looks
> like must comply with) WITHOUT RESPECT TO WHAT ELEMENT makes 
> this happen
> or how it is achieved.  IMPORTANT POINT / KEY CHANGE!! 
> 
> Proposed resolution:
> Actually version 02 of the draft already proposes a resolution, so I
> would like to start with that. This solution places 
> requirements on the
> SBE and DBE, but modifies the definition of these entities slightly
> (within the scope of the draft) to meet the spirit of the 
> above comment.
> 
> Here is the text from the -02 draft, inserted immediately after the
> architecture diagram in section-3:
> 
>    This document places requirements on the following two entities in
>    the Peering architecture:
> 
>    o  the SBE in support of the SIP/SDP interface between two peering
>       networks
> 
>    o  the DBE in support of the RTP/RTCP interface between peering
>       networks
> 
>    Based on the definition of the SBE in the peering architecture
>    document [I-D.ietf-speermint-architecture], it's not 
> entirely correct
>    to make the SBE responsible for meeting all the SIP signaling
>    requirements at the peering interface.  For example, the 
> SBE can play
>    the role of a firewall or a SIP proxy that is largely 
> transparent to
>    the SIP messages crossing the interface.  In reality, the
>    responsibility for supporting this interface belongs to all the SIP
>    entities in the SIP signaling chain, including the SBE plus the SIP
>    proxies and UAs inside the network.  Furthermore, when this network
>    is serving as a transit network, the SIP signaling chain can extend
>    into another network.  To resolve this issue, this document extends
>    the definition of the SBE slightly so that when it is specified as
>    the target entity of a normative requirement on the SIP/SDP peering
>    interface, the SBE represents all the SIP entities in the SIP
>    signaling chain that can effect the SIP/SDP peering interface.
> 
>    Likewise, the DBE can act as a media relay (performing no media
>    encode/decode) that is transparent to the RTP/RTCP packets 
> exchanged
>    with the peer network.  Therefore, this document extends the
>    definition of the DBE so that when it is specified as the target
>    entity of a normative requirement on the RTP/RTCP interface, it
>    represents the media endpoint in the peering network that 
> terminates
>    the RTP/RTCP stream.
> 
> Is this approach acceptable to everyone, or would you like to see
> something different?
> 
> Thanks
> David
> 
> 
> _______________________________________________
> Speermint mailing list
> Speermint@ietf.org
> https://www.ietf.org/mailman/listinfo/speermint
> 

Re: [Speermint] SIP Interconnect draft - requirement entity