Re: [MMUSIC] I-D Action: draft-ietf-mmusic-msrp-usage-data-channel-11.txt

[Jose M Recio <[e-mail]>]

On 10/08/19 20:58, Christer Holmberg wrote:
> Hi,
>
>> This is a very valid point. The draft is silent (I believe intentionally so?) on this point.
>>
>> My assumption would be that application specification and policy configuration would define what the GW needs >to do to properly interwork. For example, for RCS it would essentially redirect all data channel connections from >RCS clients towards the messaging server, according to some configuration and policy parameters.
>>
>> I agree that anything beyond that would probably require changes in ietf-mmusic-data-channel-sdpneg.
> Let's assume it is ok to include dcsa attributes with path.

Section 5.1.1.2 ("Use of the dcsa Attribute") specifies path as one of 
the attributes that can be used

> We all agree that the path is not going to be used for routing on the data channel interface.
>
> But, are the data channel endpoints still going to insert the path(s) in the MSRP messages?

I think nothing prevents endpoints to format From-Path and To-Path in 
MSRP messages as required by the application, for example, as defined by 
RCS specifications.

If the other endpoint is not a GW or is a GW acting as B2BUA, those 
fields would be processed as defined by MSRP and the application using 
the protocol.

If the other endpoint is a GW providing transport interworking, there's 
no reason for look at MSRP messages. It should be restricted to 
transport interworking. How transport is established it's not defined by 
the draft, it just says that path attributes should not be used for that 
purpose.

There's no other scenario, the draft leaves relays out of the scope.
>
> Currently the draft doesn't say what the path attributes are used for - it only says that they are not used for routing.
[Already commented in other message, repeating here so we focus the 
conversation on a single thread]

How the path attribute is used is defined by RFC4975, section 8.2. ("URI 
Negotiations"), that states that the URI is used to determine transport 
parameters, protection level and "to identify the target when sending 
requests and responses.".

As this draft forbids the use of path parameter for transport 
establishment, then for data channel transport the URI can only be used 
"to identify the target when sending requests and responses.".

I believe an implementor can find this, I can add a short paragraph to 
clarify in next revision that would essentially replicate what it's 
already in RFC4975, what do you think?

>
> Also, section 6 says: "Path attributes SHALL NOT be used for transport level interworking." What is meant by that?
My understanding is that it means that the URI fields (IP, port, etc) 
should not be used for the purposes of establishing the transport 
towards the endpoint.

>
> Regards,
>
> Christer
>
>
> On 18/07/19 17:18, Christer Holmberg wrote:
>> Hi,
>>
>> If think there is a bigger question here, that MAY have to be clarified ietf-mmusic-data-channel-sdpneg:
>>
>> To my understanding, the SDP dcsa attribute is used to provide information that is needed to operate the protocol (MSRP in this case) on the data channel - between the data channel endpoints.
>>
>> Now, if the data channel peer is a GW, and you need to provide some information (path attributes etc in the case of MSRP) for what happens beyond the GW, is the dcsa attribute really intended for that?
>>
>> So, if the data channel peer is an MSRP GW, using legacy MSRP on the "other side", is the dcsa attribute the correct mechanism to provide the GW with whatever information is are needed on the "other side"?
>>
>> Regards,
>>
>> Christer
>>
>>
>>
>>
>>
>>
>>
>>
>>
>>
>>
>> On 13/07/2019, 15.24, "mmusic on behalf of Makaraju, Raju (Nokia - US/Naperville)" <[e-mail] on behalf of [e-mail]> wrote:
>>
>>       >[JM] Draft covers either direct connection to an endpoint or direct connection to a gateway. As middleboxes are out of scope in the connection using >data channel, then CEMA is out of scope. This short explanation currently appears as justification for saying that CEMA is out of scope. I can remove >the explanation and just state that CEMA is out of scope.
>>       [Raju] I agree with that.
>>       
>>       > For example, if the GW is implemented on an SBC, it can use CEMA towars the inside network, using a non-dc transport. But not towards endpoints
>>       >using data channel transport.
>>       [Raju] Let's please make this clear in the draft.
>>       
>>       >[JM] I agree for RCS should be like that. Would mandating this restrict applicability for other applications?
>>       [Raju] I agree that such design limitation should only be for RCS (self-imposed), but not for other applications.
>>       
>>       Thanks
>>       Raju
>>       
>>       -----Original Message-----
>>       From: mmusic <[e-mail]> On Behalf Of Jose M Recio
>>       Sent: Tuesday, July 9, 2019 8:23 AM
>>       To: [e-mail]
>>       Subject: Re: [MMUSIC] I-D Action:
>> draft-ietf-mmusic-msrp-usage-data-channel-11.txt
>>       
>>       
>>       On 9/7/19 10:57 AM, Makaraju, Raju (Nokia - US/Naperville) wrote:
>>       > Do we need to make any further clarifications for the references to CEMA?
>>       > Section 5.1.1.2 has a sentence that is a little awkward, but seems to be saying the CEMA is out of scope.
>>       
>>       [JM] Draft covers either direct connection to an endpoint or direct connection to a gateway. As middleboxes are out of scope in the connection using data channel, then CEMA is out of scope. This short explanation currently appears as justification for saying that CEMA is out of scope. I can remove the explanation and just state that CEMA is out of scope.
>>       
>>       > But in section 6 for gateway function it mentions CEMA can be used.
>>       
>>       [JM] Draft supports gateways providing transport level interworking between MSRP endpoints using different transport protocols. CEMA can't be used towards the endpoints using data channel transport, but the GW could use CEMA to interwork with endpoints using other protocols.
>>       
>>       For example, if the GW is implemented on an SBC, it can use CEMA towars the inside network, using a non-dc transport. But not towards endpoints using data channel transport.
>>       
>>       > My understanding is I think section 6 suggests to use CEMA equivalent procedures without CEMA attributes!
>>       
>>       [JM] GW can use CEMA with endpoints using non-dc transports. This can be removed, but I think it would make the GW transport interwork option too restrictive. The B2BUA GW would not be affected.
>>       
>>       > Also, in section 6 should we add a further statement that the gateway assumes one MSRP data channel instance and one sub-protocol per SIP dialog as RCS mandates that a single SIP session/dialog has only one MSRP m= line?
>>       >
>>       [JM] I agree for RCS should be like that. Would mandating this restrict applicability for other applications?
>>       
>>       > Thanks
>>       > Raju
>>       >
>>       >
>>       > -----Original Message-----
>>       > From: mmusic <[e-mail]> On Behalf Of
>>       > [e-mail]
>>       > Sent: Saturday, June 1, 2019 11:22 PM
>>       > To: [e-mail]
>>       > Cc: [e-mail]
>>       > Subject: [MMUSIC] I-D Action:
>>       > draft-ietf-mmusic-msrp-usage-data-channel-11.txt
>>       >
>>       >
>>       > A New Internet-Draft is available from the on-line Internet-Drafts directories.
>>       > This draft is a work item of the Multiparty Multimedia Session Control WG of the IETF.
>>       >
>>       >          Title           : MSRP over Data Channels
>>       >          Authors         : Keith Drage
>>       >                            Maridi R. Makaraju (Raju)
>>       >                            Juergen Stoetzer-Bradler
>>       >                            Richard Ejzak
>>       >                            Jerome Marcon
>>       >                            Jose M. Recio
>>       > 	Filename        : draft-ietf-mmusic-msrp-usage-data-channel-11.txt
>>       > 	Pages           : 19
>>       > 	Date            : 2019-06-01
>>       >
>>       > Abstract:
>>       >     This document specifies how the Message Session Relay Protocol (MSRP)
>>       >     can be instantiated as a data channel sub-protocol, using the SDP
>>       >     offer/answer exchange-based generic data channel negotiation
>>       >     framework.  Two network configurations are documented: a WebRTC end-
>>       >     to-end configuration (connecting two MSRP over data channel
>>       >     endpoints), and a gateway configuration (connecting an MSRP over data
>>       >     channel endpoint with an MSRP over TCP or TLS endpoint).
>>       >
>>       >
>>       > The IETF datatracker status page for this draft is:
>>       > https://datatracker.ietf.org/doc/draft-ietf-mmusic-msrp-usage-data-cha
>>       > nnel/
>>       >
>>       > There are also htmlized versions available at:
>>       > https://tools.ietf.org/html/draft-ietf-mmusic-msrp-usage-data-channel-
>>       > 11
>>       > https://datatracker.ietf.org/doc/html/draft-ietf-mmusic-msrp-usage-dat
>>       > a-channel-11
>>       >
>>       > A diff from the previous version is available at:
>>       > https://www.ietf.org/rfcdiff?url2=draft-ietf-mmusic-msrp-usage-data-ch
>>       > annel-11
>>       >
>>       >
>>       > Please note that it may take a couple of minutes from the time of submission until the htmlized version and diff are available at tools.ietf.org.
>>       >
>>       > Internet-Drafts are also available by anonymous FTP at:
>>       > ftp://ftp.ietf.org/internet-drafts/
>>       >
>>       > _______________________________________________
>>       > mmusic mailing list
>>       > [e-mail]
>>       > https://www.ietf.org/mailman/listinfo/mmusic
>>       >
>>       > _______________________________________________
>>       > mmusic mailing list
>>       > [e-mail]
>>       > https://www.ietf.org/mailman/listinfo/mmusic
>>       
>>       _______________________________________________
>>       mmusic mailing list
>>       [e-mail]
>>       https://www.ietf.org/mailman/listinfo/mmusic
>>       
>>       _______________________________________________
>>       mmusic mailing list
>>       [e-mail]
>>       https://www.ietf.org/mailman/listinfo/mmusic
>>       
>>