Re: [core] FW: New Version Notification for draft-fz-core-coap-pm-00.txt

[Christian Amsüss <[e-mail]>]

Hello Giuseppe,

On Fri, Oct 22, 2021 at 04:17:02PM +0000, Giuseppe Fioccola wrote:
> We have just uploaded a new draft on CoAP Performance Measurement
> Option, which we hope to discuss at IETF 112.  This document aims to
> specify a method for the Performance Measurement of CoAP.

skimming through your slides and the document I'm left wondering what
the obtained details would be used for. Estimates for RTT and losses are
produced, for example, also in the FASOR draft[1] -- but these get
applied to retransmission parameters, which are per-hop (ie. up to the
next proxy), where good estimates are direly needed.

What would the derived values be used for?

What are the intermediaries this is intended to work with? CoAP proxies
hide the identity of the client, and moreover often apply caching (which
will at any rate need a bit of consideration if this is to be applied
through proxies). Thus, on the server side the data in the bits would
likely appear completely mixed in presence of more than one client, and
clients would receive mixed signals in presence of cache entries.

So, for both aspects, it'd be helpful if you could sketch out the
envisioned use cases.

Best regards
Christian

[1]: https://datatracker.ietf.org/doc/html/draft-ietf-core-fasor-01

-- 
To use raw power is to make yourself infinitely vulnerable to greater powers.
  -- Bene Gesserit axiom