Re: [tcpm] Comments draft-ietf-tcpm-rtorestart-06.txt

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

> A few quick comments on this latest revision:
> 
> * I believe this is a sender-side only change, but the abstract doesn't
> explicitly say this, i.e. "This document specifies a sender-only
> modification to TCP and SCTP "

Good point - this could help with reviews
 
> * The introduction does not explain why the WG has requested this to be
> experimental (is it the maturity or options of the algorithm;
> deployment
> experience; etc), e.g. "The specification in this draft is classified
> as
> "Experimental" pending experience with deployed implementations of the
> methods."

This is a valid concern, but I wonder if there is a more future-proof wording.

For similar TCPM documents we explicitly noted what further experimentation would be useful. In case of draft-ietf-tcpm-rtorestart, some of these experiments directly follow from some of the content in Section 5.

IMHO, the areas in which further experimentation is needed could be stated more explicitly, e.g., in the introduction with a reference to Section 5. This would also explain the status while being valid even if deployment experience increases in future.

Michael

> * Section 4 specifies the experimental algorithm, it may be useful to
> add a sentence at the start of this section that explicitly says this
> "This section specifies the modifications required", to contrast with
> the informative material in section 3,5,6.
> 
> * I have reviewed section 7, and this seems to me to address the
> SCTP-related comments received at the Dallas IETF meeting.
> 
> * To avoid doubt by security reviewers, You could possibly also note
> this in the security considerations section:  i.e. "This document
> specifies an experimental sender-only modification to TCP and SCTP."
> 
> Gorry