Re: [Cats] E-impact [Was: Other "related" I-Ds]

Hi Adrian,

Agree to all that.  But the walking was not what i questioned,  to stay in your analogy, but the purpose of it, even what that "basic CATS scenario" is you refer to (we might have different views on even that).

You asked about views what metrics could and should be.  I wondered and suggested to consider some sort of energy as one such, that's all.

I don't think that doing so will affect the basic engineering tools.  Like any metric, it may impact aspects like update frequency but any metric does that question to answer.

Best

Dirk
P.S.: if you worked on cardinals without semantics exposed to CATS nodes,  you wouldn't even notice that my use case was running on an energy objective.  I'm working on showing this as an extension to our previous work where those cardinals were compute units.  I hope we can share results soon...

From:Adrian Farrel <adrian@olddog.co.uk<mailto:adrian@olddog.co.uk>>
To:Dirk Trossen <dirk.trossen@huawei.com<mailto:dirk.trossen@huawei.com>>;'cats' <cats@ietf.org<mailto:cats@ietf.org>>
Date:2024-03-25 19:22:31
Subject:E-impact [Was: Other "related" I-Ds]

Hi Dirk,

Just tweaking the subject line before responding with my chair hat half on
and half off!

I think thar energy *is* a factor for CATS, but the previous discussion
arrived at "not at day one." The challenge for CATS is clearly learning to
walk, and adding juggling with knives at this stage may make it "hard".

Normally, I describe this sort of problem as "engineering with the Force."
You need to get on with the basic engineering work and kind of be aware that
the more complicated problem exists and will need to be factored in one day.
The end result should be a core set of engineering tools that apply to the
basic CATS scenario, but which are easily extended to factor in other
things, such as energy consumption.

But you are right. A WG has a tendency to be inward looking. We should be
happy to share our work more widely so that derivatives can be built.

Cheers,
Adrian (wondering about DOGS, EATS, CATS)

-----Original Message-----
From: Dirk Trossen <dirk.trossen@huawei.com>
Sent: 25 March 2024 12:25
To: adrian@olddog.co.uk; 'cats' <cats@ietf.org>
Subject: RE: [Cats] Other "related" I-Ds

Adrian, all,

Not necessarily wanting to complicate the discussion but I wonder of the
relation of CATS to the e-impact discussions (e.g., in relation to the draft
at
https://datatracker.ietf.org/doc/draft-cparsk-eimpact-sustainability-conside
rations/)?

In the past, we had discussed the aspect of energy as a possible objective
of CATS, as also the abstract of the use case draft outlines that
"Distributed computing is a tool that service providers can use to achieve
better service response time and optimized energy consumption", although the
use cases themselves do not address more specifically that aspect (mainly
focusing on optimizing latency it seems).

Hence for the metric discussion, capturing an energy angle could be an
aspect of 'compute awareness' that can certainly  be linked to the e-impact
discussions. Here, would an exposure of CATS to the e-impact community be a
good step to get the discussion going as to the nature of such possible
metrics that could lead to an EATS derivate of CATS?

Best,

Dirk

-----Original Message-----
From: Cats <cats-bounces@ietf.org> On Behalf Of Adrian Farrel
Sent: 24 March 2024 12:49
To: 'cats' <cats@ietf.org>
Subject: [Cats] Other "related" I-Ds

Hi all,

In Brisbane, Jeff T raised a point about other I-Ds (specifically in RTGWG)
that related in some way to the work of CATS.

This caused me to trawl the Datatracker for non-expired I-Ds that are
somehow connected to "network computing function" but which don't show up on
the CATS page in the Datatracker because they don't have "-cats-" in the
name.

Many of these will be concerned with on-path computation, and others with
micro-computation (for example, to enable other on-path functionality such
as routing or security).

We would need to think carefully about whether any of this work fits in the
CATS charter. And, in fact, there may be no overlap at all in many cases.

However, they are worth skimming to see what can be learned especially as we
get further into our discussion of metrics.

The list is provided without any statement on my part as to whether these
could be useful.

Cheers,
Adrian

https://datatracker.ietf.org/doc/draft-rcr-opsawg-operational-compute-metric
s/
https://datatracker.ietf.org/doc/draft-lou-rtgwg-sinc/
https://datatracker.ietf.org/doc/draft-dcn-dmm-cats-mup/
https://datatracker.ietf.org/doc/draft-ietf-mops-ar-use-case/
https://datatracker.ietf.org/doc/draft-irtf-coinrg-use-cases/
https://datatracker.ietf.org/doc/draft-ietf-teep-usecase-for-cc-in-network/
https://datatracker.ietf.org/doc/draft-song-inc-transport-protocol-req/
https://datatracker.ietf.org/doc/draft-lou-rtgwg-sinc-deployment-considerati
ons/
https://datatracker.ietf.org/doc/draft-contreras-coinrg-clas-evolution/
https://datatracker.ietf.org/doc/draft-fu-idr-computing-info-notification-do
main/
https://datatracker.ietf.org/doc/draft-zhang-rtgwg-srv6-computing-connect-us
ecases/
https://datatracker.ietf.org/doc/draft-zhou-intarea-computing-segment-san/
https://datatracker.ietf.org/doc/draft-dreibholz-rserpool-applic-distcomp/

--
Cats mailing list
Cats@ietf.org
https://www.ietf.org/mailman/listinfo/cats