Re: [tsvwg] Reminder: Advancing MP-DCCP towards a working group last call : Intended publication status

[[e-mail]]

Hi Martin,

Regarding deployment I can't speak for others, but I can say that I assume that MP-DCCP is not associated with the sparse use of DCCP. MP-DCCP has special characteristics which makes it suitable for providing multi-path transport for latency sensitive traffic, especially in combination with encapsulation. From that I see this de-coupled from any DCCP deployment.

DT as main editor of the MP-DCCP draft is confronted with requests from companies exactly looking into the encapsulation case, for example for drone or V2X communication. We also now that MP-DCCP is interesting for (future) operators of Hybrid Access. Btw. Xiaomi will present their Android 13 implementation effort and interoperability results on Monday in the TSVWG session.

Br

Markus


From: Martin Duke <[e-mail]>
Sent: Dienstag, 1. November 2022 19:39
To: Amend, Markus <[e-mail]>
Cc: [e-mail]; [e-mail]; [e-mail]
Subject: Re: [tsvwg] Reminder: Advancing MP-DCCP towards a working group last call : Intended publication status

(With no hats on)

It's true that similar efforts are happening in TCP and QUIC, but those are likely to get much more deployment experience, as the installed base is much wider than DCCP.

How much deployment is this going to see in the near term (in production, not research projects or proofs of concept)?

Martin

On Tue, Nov 1, 2022 at 8:37 AM <[e-mail]<mailto:[e-mail]>> wrote:
Dear Gorry, all,

after we have seen feedback from Xiaomi, BT and Vodafone who are in favour of PS, I formally request as one of the MP-DCCP draft authors to consider the move from EXP to PS.

The authors of the MP-DCCP draft sharing your view Gorry and the one from Tengmingxuan (Xiaomi), that MP-DCCP is close to existing MP efforts in IETF like MPTCP and MP-QUIC, which both follow PS track.

We further know that Tengmingxuan and BT have experience in implementing MP-DCCP and MPTCP, which further strengthens their support for PS.

That’s why we don’t see any reason to stay with EXP.

Br

Markus

From: tsvwg <[e-mail]<mailto:[e-mail]>> On Behalf Of Gorry Fairhurst
Sent: Donnerstag, 13. Oktober 2022 11:55
To: [e-mail]<mailto:[e-mail]>
Cc: [e-mail]<mailto:[e-mail]>
Subject: [tsvwg] Reminder: Advancing MP-DCCP towards a working group last call : Intended publication status

We'd like to close this topic at the next meeting.

If you would like to express thoughts on what experimental work would be needed if draft-ietf-tsvwg-multipath-dccp were too be published as an EXP spec rather than a Proposed Standard, please do so either on-list or privately to the chairs (at [e-mail]<mailto:[e-mail]> <[e-mail]><mailto:[e-mail]>).

Best wishes
Gorry, David and Marten
(TSV WG Co-Chairs)
On 28/09/2022 10:26, Gorry Fairhurst wrote:

TSVWG,
    When MP-DCCP was adopted in 2021, it proposed publication was EXP,
    based on this needing experience to complete the design. This work
    item has now progressed in TSVWG and seems to be approaching a WGLC.

    The IETF has completed related work on multipath protocol extensions for
    TCP (published in 2020 as PS in RFC8824), and more recently new work was
    adopted in the QUIC WG on multipath (which currently also targets PS).
    Bearing this in mind, and the additional implementation feedback from
    developing MP-DCCP, the chairs would welcome thoughts on what
    experimental work would still be needed - and whether this ID should
    target EXP or PS for publication.

Notes on this topic will help the chairs to prepare for the  anticpated WG last call of the MP-DCCP Spec.

Best wishes,

Gorry, David and Marten
(TSV WG Co-Chairs)