[Onsen] Re: [ainetops] Re: Semantic meaning across representations and boundaries

[Christopher Janz <[e-mail]>]

Thomas — thank you, I really appreciate your engagement and very helpful
comments and suggestions.

The slide adjustments I've taken as you propose: "operational and
analytical" instead of "raw," and "service-model worlds" for the ONSEN box.
And you're right that incident and postmortem work is an important and
concrete example of pragmatics; I've put it on the triad slide and I'm
citing your lifecycle and semantics drafts as a human-in-the-loop example.
On your multi-vendor question: yes, I think the ontology approach is
complementary to letting the network and service models carry the mapping.
The models handle it where a shared model already exists; ontology and
lexicon earn their keep exactly where one doesn't, which includes across
different ICT domains. That's the case where I'd expect it to matter most.
Updated slides are here:
https://github.com/ChrisFJanz/meaning-across-representations/blob/main/semantic_model_deck.pdf

The bigger matter you raise is the one I most want to pick up, and if I can
add my own bias to yours: what I am particularly interested in on a
research level is using cognitive machines to automate procedures and
working out which reference and accumulating supports make that automation
functional and trustworthy. Your two drafts are almost a worked example:
the lifecycle state machine is a procedure to automate, the symptom
semantics are a reference support it leans on, and the label store is an
accumulating one. So, the experiment almost writes itself: automate an
operational procedure - anomaly and incident handling is the obvious first
one - with a human in the loop and watch how a growing symptom lexicon and
an accumulating, reconciled commons of signatures make the automation
better over time, including reconciling labels across operators and
vendors. Research brings the ontology, the lexicon and the automation;
operators bring the pragmatics and the validation. I'd call anomaly a first
sphere of pragmatics deliberately: it's concrete and you've already got
running code; but the work is extensible in parallel or subsequently into
intent, lifecycle, and the commercial and provenance context.

One thing I might ask you: the knowledge-graph design team that you and
Brad were part of last year did a lot of groundwork - KG-for-netops, NORIA,
the Marcas work, the SIMAP/NORIA-O proof of concept - and as Brad indicated
that team has moved to carry on now with independent work. What’s the
hand-off or continuity from that effort?

Your joint-meeting idea is very welcome. Happy to help shape a first
experiment we could put in front of both groups.

Thanks again, Thomas.
Chris

On Sun, Aug 2, 2026 at 2:53 AM <[e-mail]> wrote:

> Dear Chris,
>
>
>
> Very well done. I like the slide deck a lot. You have clearly put a lot of
> thoughts into this.
>
>
>
> The problem space you are describing and how ontology and lexicon's can be
> applied to bridge data, service and network YANG models is an area which
> interests me and I believe will be very relevant in todays and future
> cognitive AI domain. A few biased suggestions and comments for the slide
> deck:
>
>
>
>
>
> Slide 2
>
> --
>
> replace "raw" with "operational and analytical"
>
> replace "bridge two schema worlds" with "update and bridge two service
> model worlds"
>
> comment: normalizing multi-vendor, interesting alternative approach, until
> know my thoughts where that network and service models would do that
> mapping instead. Do you agree that this would be complementary and
> especially useful to bridge different ICT domains? See also slide 9.
>
>
>
> Slide 5
>
> --
>
> comment: Network Incident Network Analytics Postmortems are a form of
> pragmatics. With (slide 13,
> https://datatracker.ietf.org/meeting/126/materials/slides-126-nmop-sessb-swisscom-network-analytics-bgp-routing-loop-postmortem
> <https://datatracker.ietf.org/meeting/126/materials/slides-126-nmop-sessb-swisscom-network-analytics-bgp-routing-loop-postmortem-00>,
> draft-ietf-nmop-network-anomaly-lifecycle) we want to describe symptoms and
> find their commonalities.
>
>
>
> Slide 8
>
> --
>
> comment: For human in the loop, you might want to consider
> draft-ietf-nmop-network-anomaly-lifecycle and
> draft-ietf-nmop-network-anomaly-semantics as another "supplied by a machine
> or human" and between human and machine example.
>
>
>
> Slide 15
>
> --
>
> comment: I miss the experiment aspect here. As an operator I would love to
> see what research is proposing on ontology and lexicon, being applied with
> a human in the loop workflow and pragmatics being mapped. This is for me an
> excellent example where I see NMRG and NMOP working jointly together.
>
>
>
>
>
> If you agree that pragmatics is something which network operators should
> contribute and ontology and lexicon's is proposed by research, I suggest
> that you propose how the two communities could work together (a workflow)
> in validating and feedback the research from a human in the loop systems
> side. This could be perhaps an example topic we could discuss in a joint
> meeting between NMOP and NMRG, thinking out loud here.
>
>
>
> Best wishes
>
> Thomas
>
>
>
> *From:* Christopher Janz <christopher.janz=[e-mail]>
> *Sent:* Friday, July 31, 2026 3:00 PM
> *To:* nmop <[e-mail]>; [e-mail]; [e-mail]; [e-mail];
> [e-mail]
> *Subject:* [netmod] Semantic meaning across representations and boundaries
>
>
>
> *Be aware:* This is an external email.
>
>
>
> All – as I think has been broadly recognized during meetings last week and
> through various activities and discussion threads both leading up to and
> following them, a common theme and general problem seems to have emerged,
> in various ways, within and across NMOP, NETMOD, ONSEN, NMRG, and now as
> one element of the AIOps discussion.
>
>
>
> These multiple discussions and problem contexts point at the same need:
> for semantic models for network information — meaning made explicit and
> independent of form – capable of supporting bridging of model and system
> boundaries and information consumption by machines.
>
>
>
> I have put some form, structure and justification around a problem
> statement and tied it all to the various takes/contexts mooted and
> discussions in play. In the end I think there is a relatively focused but
> true research problem to be considered, which I have outlined and would
> suggest properly belongs in NMRG. For further discussion there … of course
> such a program would not preclude related activities and outcomes in other
> WGs, perhaps working against narrower/more specific targets.
>
>
>
> Please have a look, and respond with your thoughts – if you wish, on the
> github repository here:
> https://github.com/ChrisFJanz/meaning-across-representations ... The deck
> is there, and also attached here for convenience.
>
>
>
> Best
>
>
>
> Chris
>
>
>
>
> _______________________________________________
> ainetops mailing list -- [e-mail]
> To unsubscribe send an email to [e-mail]
>