Re: [tcpm] Rechartering TCPM for alternative congestion control algorithms

["Scharf, Michael (Michael)" <[e-mail]>]

> How would the usage of TCP based CCs in SCTP be handled? Would it be
> covered in
> the documents in TCPM? Would there additional documents be needed? If
> yes, would
> they go also to TCPM or to TSVWG? Just wondering...

Good question. I can only respond as far as TCPM is concerned. The current TCPM charter states in two other paragraphs [1]:

The focus of the working group is TCP. In cases where small
changes are directly applicable to other transports (e.g., SCTP or
DCCP), the mappings to other transports may be specified alongside
that for TCP, but other significant additions and changes to other
transports are not in scope.

[...]

TCP's congestion control algorithms are the model followed by
alternate transports (e.g., SCTP or DCCP), which are standardized in
other working groups, such as the Transport Area WG (tsvwg). In the
past, the IETF has worked on several documents about algorithms that
are specified for multiple protocols (e.g., TCP and SCTP) in the
same document. Which WG shepherds such documents will be determined
on a case-by-case basis. In any case, the TCPM WG will remain in
close contact with other relevant WGs working on these protocols to
ensure openness and stringent review from all angles.


My current thinking is to leave these two paragraphs in the TCPM charter unchanged. As far as I recall, in the past TCPM has handled basically all documents specifying algorithms both for TCP and for SCTP. In contrast, SCTP-only documents are explicitly outside of the current charter of TCPM (i.e., they probably have to go to TSVWG). Thus, it would depend whether a CC algorithm would be specified for TCP and SCTP in the same document, or not. I could also think of TCPM documents describing the usage in SCTP in a non-normative section or an appendix.

The proposed TCPM charter is very explicitly limited to CC algorithms "known to be widely deployed". Adoption in TCPM would probably require that said algorithm has been tested very extensively in TCP stacks (e.g., experiments in which the algorithm has been enabled by default). I don't know whether there is a way to ensure comprehensive testing of an algorithm both for TCP and SCTP. Therefore, I don't think we should mandate that all CC algorithms submitted to TCPM have to be specified both for TCP and SCTP in the same document. The congestion control in RFC 5681 is also specified for TCP only.

Are there any suggestion for an alternative handling?

Thanks

Michael


[1] https://datatracker.ietf.org/wg/tcpm/charter/