[Tsv-art] Tsvart last call review of draft-ietf-rtgwg-qos-model-11

[Colin Perkins via Datatracker <[e-mail]>]

Reviewer: Colin Perkins
Review result: Ready with Nits

This document has been reviewed as part of the transport area review team's
ongoing effort to review key IETF documents. These comments were written
primarily for the transport area directors, but are copied to the document's
authors and WG to allow them to address any issues raised and also to the IETF
discussion list for information.

When done at the time of IETF Last Call, the authors should consider this
review as part of the last-call comments they receive. Please always CC
[e-mail] if you reply to or forward this review.

I am not an expert on YANG or DiffServ, and I have not followed the development
and discussion related to this draft. This review is hence necessarily written
from a generalist transport perspective. Please accept my apologies if I touch
on topics that have been considered before in the working group.

The draft looks to be defining mechanisms to configure the use of existing QoS
mechanisms and to report on their effects. As such, any new transport protocol
impact would seem limited. The mechanisms described may make it easier to
deploy QoS, but the QoS techniques exist and can be used irrespective of
whether this draft is published.

For AQM, this draft specifies configuration parameters for RED and WRED. These
AQM algorithms have certainly been widely implemented and used, but there are
more modern alternatives that have been defined in IETF and that are also
starting to see use (e.g., PIE – RFC 8033, and several variants on CoDel – RFC
8289/8290). Has consideration been given to whether any other AQM algorithms
should be included? Is the mechanism extensible to support these and other
future AQM approaches? From a transport perspective I would not recommend use
of RED or WRED today, since the alternatives perform better and are harder to
misconfigure. Some discussion about extensibility and alternatives would be
helpful.

Similarly there are only two traffic classifiers specified, which may warrant
an extension point.

Otherwise, this seems broadly ready.

Colin