[multipathtcp] Future of the MPTCP Working group

[<[e-mail]>]

For people not in the WG meeting in Montreal, and also after a couple of private comments, we wanted to summarise our perspective about the future of the working group.
With the completion of the protocol bis, our (revised) charter is essentially complete. We've already published RFC 8041 (Use Cases and Operational Experience with Multipath TCP), as well as RFC 7430 (Analysis of Residual Threats and Possible Fixes for Multipath TCP (MPTCP)). The middlebox item turned into the 0-RTT Converters work in tcpm (which is in WG last call). Excellent work!
Having completed the charter, the options are to re-charter or close. Note that closing the WG can be considered the default option. Re-chartering is possible (subject to the usual approvals procedure), but needs an increase on the current energy levels, in particular we would need a few topics - on each of which a group of people want to contribute, in terms of design, implementation and review (as well as the usual consensus to do work). Note even if the WG is closed, at least some MPTCP work could continue - for instance, work that is a 'minor extension' of MPTCP, or its maintenance, could be done in TCPM (under its current charter).
As a data point, there are currently the following active individual drafts:

1.       Multipath TCP extension for Robust Session Establishment

2.       Initial-Path Selection for Connection Establishment in Multipath TCP

3.       Privacy threats and possible countermeasures for Multipath-TCP (MPTCP)

4.       TFO support for Multipath TCP

5.       5G Session Continuity Support in MPTCP

6.       Multipath TCP Inactivity Time Option

7.       Multipath TCP Subflow Rate Limit Option

8.       Extended Socket APIs to control subflow priority in Multipath TCP

9.       One Way Latency Considerations for MPTCP

10.   A Stochastic Optimal Scheduler for Multipath Transmission Control Protocol (MPTCP)
Also - again just as a data point - #3 has authors from multiple institutions, as do #8 & #9, whilst #1 & #2 look to be addressing a similar problem (I know the authors now plan to collaborate). In addition, we believe there are several people interested in a more secure version of MPTCP (and expired individual drafts such as draft-bonaventure-mptcp-tls). So these might be topics on which a group of people want to contribute, in terms of design, implementation and review. For efforts to generate extra energy and collaboration, please make sure they're visible to us and on the mailing list, well before the Singapore IETF.
On a related point, there was a discussion in tsvarea about "Where to do multipath work in the IETF?" https://datatracker.ietf.org/meeting/105/materials/minutes-105-tsvarea-00 which finished with Mirja saying "I think the community needs to propose something here, please work on that if you're interested and think you can scope it in a reasonable way."
Best wishes,
Phil and Yoshi