[OPS-AREA]Re: [nmrg] Re: [NMOP] Re: [OPSAREA @IETF126] IRTF/IETF OPS Transfer: Status & Exploring Coll aboration Opportunities

"Joe Clarke \(jclarke\)" <[email protected]> Mon, 13 Jul 2026 07:32:38 +0000
Newsgroups gmane.ietf.ops,gmane.ietf.opsawg
Message-ID <CH2PR11MB8867C24122AC20D9089E1286B8FA2@CH2PR11MB8867.namprd11.prod.outlook.com>
--===============0736854182616703133==
Content-Language: en-US
Content-Type: multipart/alternative;
 boundary="_000_CH2PR11MB8867C24122AC20D9089E1286B8FA2CH2PR11MB8867namp_"

--_000_CH2PR11MB8867C24122AC20D9089E1286B8FA2CH2PR11MB8867namp_
Content-Type: text/plain; charset="Windows-1252"
Content-Transfer-Encoding: quoted-printable


Catching up on this conversation and thanks for sharing your thoughts and c=
ommenting. What you are describing making very much sense to me. Regarding:

Q4, How does an AI agent interact with Network Digital Twin, Intent-Based N=
etworking, SIMAP, Knowledge Graphs and YANG data in Message Brokers?
  A:   It depends on the specific situation, there should be no one-size-fi=
ts-all solution.

[JMC] See above on a thought on digital twin.  And, yeah, I don=92t think w=
e=92ll have a one-size fits all model.

[TG] My biased view is that a Network Digital Twin is helpful in context of=
 Planning & Reasoning where Intent-Based Networking, SIMAP and Knowledge Gr=
aphs in context of Learning & Self-Correction. I see the YANG schema and se=
mantics in the YANG schema registry (https://datatracker.ietf.org/doc/html/=
draft-ietf-nmop-yang-message-broker-integration-13#section-4.4) and in the =
Streaming Catalog (https://datatracker.ietf.org/doc/html/draft-ietf-nmop-ya=
ng-message-broker-integration-13#section-4.5) in Message Brokers as a sourc=
e of information for the RAG (Retrieval-Augmented Generation) and identifyi=
ng from which Message Broker topics which YANG data should be consumed from=
, which Chonfeng lists under Tool Use & Execution.

What are your thoughts? Does that make sense?

[JMC2] It makes sense, sure.  What is yet to be seen is if all of that sche=
ma and organization makes RAG any more efficient than loading in configs, i=
nventory, and topology raw data and then presenting network questions/probl=
ems.  I=92ve seen the latter be pretty good though not perfect.  I=92m not =
positive so much more effort is going to help, though.

Joe

Best wishes
Thomas

From: Joe Clarke (jclarke) <[email protected]>
Sent: Thursday, June 25, 2026 8:27 PM
To: Chongfeng Xie <[email protected]>; Graf Thomas, SCS-INI-NET-VNC-E2=
E <[email protected]>; ops-area <[email protected]>; opsawg <opsawg@=
ietf.org>; nmop <[email protected]>; [email protected]
Cc: [email protected]
Subject: Re: [nmrg] Re: [NMOP] Re: [OPSAREA@IETF126] IRTF/IETF OPS Transfer=
: Status & Exploring Collaboration Opportunities

Be aware: This is an external email.

Chongfeng, I mostly agree with everything you said.  That is, I have a simi=
lar understanding.  I do have a few comments, though, based on my implement=
ation experience.


Q1. Which AI agents capabilities are suitable for network management?
   A:   AI agents have the following common capabilities,

    - Perception: Processes multi-modal inputs (text, images, audio) and tr=
acks context and state over time.

[JMC] This depends on model.  Not all models can handle all types of input.=
  And the model capabilities will influence overall cost.

    - Planning & Reasoning: Breaks high-level goals into actionable sub-tas=
ks, dynamically re-plans when obstacles arise, and evaluates trade-offs bef=
ore acting.

    - Tool Use & Execution: Autonomously calls external tools=97search engi=
nes, APIs, calendars, code interpreters, and databases=97to take real actio=
n in the digital world (e.g., send emails, book flights, query data).

[JMC] More specifically, with network management, tools can access devices =
to get data and adjust configuration.  In the former, what the agent can th=
en do with the data it retrieves heavily depends on the model and its train=
ing data.  For example, many frontier models generally do well with basic n=
etwork operational data.  Adding on to tool use, RAG or purpose-trained or =
tuned models, and you can get even better results from such tools calls.

    - Learning & Self-Correction: Detects errors, reflects on outcomes, ite=
rates on solutions, and utilizes both short-term and long-term memory to im=
prove over time.

[JMC] This is where Digital Twin comes into place.  You don=92t necessarily=
 want a model to =93learn=94 as it=92s going through your production networ=
k.  Testing its =93reasoning=94 in a twin environment makes it safer for do=
ing that self-correction and iteration.

All the capabilities above are useful for network management. With the unde=
rground support of LLM models, the following functionalities can be realize=
d by AI agents in the context of network management,

   - Intent-based auto-assignment: Automatically routes NOC tickets based o=
n detected intent.

   - Intelligent root-cause analysis: Pinpoints fault origins with precisio=
n.

[JMC] Eh, maybe.  See above.  I=92ve had models go off the rails on hyper-s=
pecific problems where it didn=92t have a lot of relevant training data.

   - Actionable troubleshooting guidance:  Generates step-by-step plans and=
 recommended resolutions.

   - AI-driven orchestration:  Enables expert-human collaboration and autom=
ated execution of task reassignments/suspensions.

   - Continuous network optimization:  Delivers ongoing tuning recommendati=
ons for wireless, core, and cloud-network resources across diverse scenario=
s.

   - Automated recovery verification:  Validates fault resolution by queryi=
ng network and perceptual metrics via LLM-orchestrated API calls.

Q2. What is the impact of an AI network management process to YANG informat=
ion, service, network and data modelling?
    A:   AI and legcay network management models belong to different domain=
s,  the specific manangement modelling is aganostic to the AI network manan=
gent process. In other words, AI system can process any YANG information, s=
ervice, network and data modelling.

[JMC] This has exactly been my experience.  I=92ve had good luck with a YAN=
G-centric skill.  Where things can get problematic is the token consumption=
 XML and JSON produce when using open weight models with tighter context wi=
ndows.

Q3.  Is MCP (https://en.wikipedia.org/wiki/Model_Context_Protocol) a suitab=
le protocol between AI agents and IETF network management data sources and =
system components?
     A:  Yes, as an open standard that connects AI models to external data =
and tools, MCP enables AI assistants to securely read files, query database=
s, and execute actions across different applications.  The likelihood of MC=
P being used between between AI agents and IETF network management data sou=
rces and system components is very high.
In addition, other approaches can be used, such as APIs exposed by network =
controller, in this case, AI agents can call the APIs directly for network =
device configuration or data collection.

[JMC] Agree completely.  This does work well.  And MCP has considerable fra=
mework backing.

Q4, How does an AI agent interact with Network Digital Twin, Intent-Based N=
etworking, SIMAP, Knowledge Graphs and YANG data in Message Brokers?
  A:   It depends on the specific situation, there should be no one-size-fi=
ts-all solution.

[JMC] See above on a thought on digital twin.  And, yeah, I don=92t think w=
e=92ll have a one-size fits all model.

Joe

--_000_CH2PR11MB8867C24122AC20D9089E1286B8FA2CH2PR11MB8867namp_
Content-Type: text/html; charset="Windows-1252"
Content-Transfer-Encoding: quoted-printable

<html>
<head>
<meta http-equiv=3D"Content-Type" content=3D"text/html; charset=3DWindows-1=
252">
</head>
<body>
<div style=3D"direction: ltr; font-family: Aptos, Arial, Helvetica, sans-se=
rif; font-size: 12pt; color: rgb(0, 0, 0);">
<br>
</div>
<div id=3D"mail-editor-reference-message-container">
<p class=3D"MsoNormal" style=3D"margin: 0in; font-family: Aptos, sans-serif=
; font-size: 12pt;">
<span style=3D"font-family: &quot;Trebuchet MS&quot;, sans-serif; font-size=
: 10pt; color: rgb(14, 40, 65);">Catching up on this conversation and thank=
s for sharing your thoughts and commenting. What you are describing making =
very much sense to me. Regarding:</span></p>
<p class=3D"MsoNormal" style=3D"margin: 0in; font-family: Aptos, sans-serif=
; font-size: 12pt;">
<span style=3D"font-family: &quot;Trebuchet MS&quot;, sans-serif; font-size=
: 10pt; color: rgb(14, 40, 65);">&nbsp;</span></p>
<p class=3D"MsoNormal" style=3D"margin: 0in; font-family: Aptos, sans-serif=
; font-size: 12pt;">
<span style=3D"font-family: &quot;Trebuchet MS&quot;, sans-serif; font-size=
: 10pt; color: rgb(14, 40, 65);">Q4, How does an AI agent interact with Net=
work Digital Twin, Intent-Based Networking, SIMAP, Knowledge Graphs and YAN=
G data in Message Brokers?</span></p>
<p class=3D"MsoNormal" style=3D"margin: 0in; font-family: Aptos, sans-serif=
; font-size: 12pt;">
<span style=3D"font-family: &quot;Trebuchet MS&quot;, sans-serif; font-size=
: 10pt; color: rgb(14, 40, 65);">&nbsp; A:&nbsp; &nbsp;It depends on the sp=
ecific situation, there should be no one-size-fits-all solution.</span></p>
<p class=3D"MsoNormal" style=3D"margin: 0in; font-family: Aptos, sans-serif=
; font-size: 12pt;">
<span style=3D"font-family: &quot;Trebuchet MS&quot;, sans-serif; font-size=
: 10pt; color: rgb(14, 40, 65);">&nbsp;</span></p>
<p class=3D"MsoNormal" style=3D"margin: 0in; font-family: Aptos, sans-serif=
; font-size: 12pt;">
<span style=3D"font-family: &quot;Trebuchet MS&quot;, sans-serif; font-size=
: 10pt; color: rgb(14, 40, 65);">[JMC] See above on a thought on digital tw=
in. &nbsp;And, yeah, I don=92t think we=92ll have a one-size fits all model=
.</span></p>
<p class=3D"MsoNormal" style=3D"margin: 0in; font-family: Aptos, sans-serif=
; font-size: 12pt;">
<span style=3D"font-family: &quot;Trebuchet MS&quot;, sans-serif; font-size=
: 10pt; color: rgb(14, 40, 65);">&nbsp;</span></p>
<p class=3D"MsoNormal" style=3D"margin: 0in; font-family: Aptos, sans-serif=
; font-size: 12pt;">
<span style=3D"font-family: &quot;Trebuchet MS&quot;, sans-serif; font-size=
: 10pt; color: rgb(14, 40, 65);">[TG] My biased view is that a Network Digi=
tal Twin is helpful in context of Planning &amp; Reasoning where Intent-Bas=
ed Networking, SIMAP and Knowledge Graphs in context
 of Learning &amp; Self-Correction. I see the YANG schema and semantics in =
the YANG schema registry (</span><span style=3D"font-family: &quot;Trebuche=
t MS&quot;, sans-serif; font-size: 10pt; color: rgb(70, 120, 134);"><a href=
=3D"https://datatracker.ietf.org/doc/html/draft-ietf-nmop-yang-message-brok=
er-integration-13#section-4.4" data-outlook-id=3D"c3587784-5a44-4b51-9529-0=
4a5ac3513f2" style=3D"color: rgb(70, 120, 134); margin-top: 0px; margin-bot=
tom: 0px;"><u>https://datatracker.ietf.org/doc/html/draft-ietf-nmop-yang-me=
ssage-broker-integration-13#section-4.4</u></a></span><span style=3D"font-f=
amily: &quot;Trebuchet MS&quot;, sans-serif; font-size: 10pt; color: rgb(14=
, 40, 65);">)
 and in the Streaming Catalog (</span><span style=3D"font-family: &quot;Tre=
buchet MS&quot;, sans-serif; font-size: 10pt; color: rgb(70, 120, 134);"><a=
 href=3D"https://datatracker.ietf.org/doc/html/draft-ietf-nmop-yang-message=
-broker-integration-13#section-4.5" data-outlook-id=3D"8f571d13-f638-4eef-9=
f60-739752f2a8e6" style=3D"color: rgb(70, 120, 134); margin-top: 0px; margi=
n-bottom: 0px;"><u>https://datatracker.ietf.org/doc/html/draft-ietf-nmop-ya=
ng-message-broker-integration-13#section-4.5</u></a></span><span style=3D"f=
ont-family: &quot;Trebuchet MS&quot;, sans-serif; font-size: 10pt; color: r=
gb(14, 40, 65);">)
 in Message Brokers as a source of information for the RAG (Retrieval-Augme=
nted Generation) and identifying from which Message Broker topics which YAN=
G data should be consumed from, which Chonfeng lists under Tool Use &amp; E=
xecution.</span></p>
<p class=3D"MsoNormal" style=3D"margin: 0in; font-family: Aptos, sans-serif=
; font-size: 12pt;">
<span style=3D"font-family: &quot;Trebuchet MS&quot;, sans-serif; font-size=
: 10pt; color: rgb(14, 40, 65);">&nbsp;</span></p>
<p class=3D"MsoNormal" style=3D"margin: 0in; font-family: Aptos, sans-serif=
; font-size: 12pt;">
<span style=3D"font-family: &quot;Trebuchet MS&quot;, sans-serif; font-size=
: 10pt; color: rgb(14, 40, 65);">What are your thoughts? Does that make sen=
se?</span></p>
<p class=3D"MsoNormal" style=3D"margin: 0in; font-family: Aptos, sans-serif=
; font-size: 12pt;">
<span style=3D"font-family: &quot;Trebuchet MS&quot;, sans-serif; font-size=
: 10pt; color: rgb(14, 40, 65);"><br>
</span></p>
<p class=3D"MsoNormal" style=3D"margin: 0px 0in;"><span style=3D"font-famil=
y: Aptos, sans-serif; font-size: 13px; color: rgb(14, 40, 65);">[JMC2] It m=
akes sense, sure. &nbsp;What is yet to be seen is if all of that schema and=
 organization makes RAG any more efficient
 than loading in configs, inventory, and topology raw data and then present=
ing network questions/problems. &nbsp;I=92ve seen the latter be
<i>pretty good</i>&nbsp;though not perfect. &nbsp;I=92m not positive so muc=
h more effort is going to help, though.</span></p>
<p class=3D"MsoNormal" style=3D"margin: 0px 0in;"><span style=3D"font-famil=
y: Aptos, sans-serif; font-size: 13px; color: rgb(14, 40, 65);"><br>
</span></p>
<p class=3D"MsoNormal" style=3D"margin: 0px 0in;"><span style=3D"font-famil=
y: Aptos, sans-serif; font-size: 13px; color: rgb(14, 40, 65);">Joe</span><=
/p>
<p class=3D"MsoNormal" style=3D"margin: 0in; font-family: Aptos, sans-serif=
; font-size: 12pt;">
<span style=3D"font-family: &quot;Trebuchet MS&quot;, sans-serif; font-size=
: 10pt; color: rgb(14, 40, 65);">&nbsp;</span></p>
<p class=3D"MsoNormal" style=3D"margin: 0in; font-family: Aptos, sans-serif=
; font-size: 12pt;">
<span style=3D"font-family: &quot;Trebuchet MS&quot;, sans-serif; font-size=
: 10pt; color: rgb(14, 40, 65);">Best wishes</span></p>
<p class=3D"MsoNormal" style=3D"margin: 0in; font-family: Aptos, sans-serif=
; font-size: 12pt;">
<span style=3D"font-family: &quot;Trebuchet MS&quot;, sans-serif; font-size=
: 10pt; color: rgb(14, 40, 65);">Thomas</span></p>
<p class=3D"MsoNormal" style=3D"margin: 0in; font-family: Aptos, sans-serif=
; font-size: 12pt;">
<span style=3D"font-family: &quot;Trebuchet MS&quot;, sans-serif; font-size=
: 10pt; color: rgb(14, 40, 65);">&nbsp;</span></p>
<div style=3D"padding: 3pt 0in 0in; border-width: 1pt medium medium; border=
-style: solid none none; border-color: rgb(225, 225, 225) currentcolor curr=
entcolor;">
<p class=3D"MsoNormal" style=3D"margin: 0in; font-family: Aptos, sans-serif=
; font-size: 12pt;">
<span style=3D"font-family: Calibri, sans-serif; font-size: 11pt;"><b>From:=
</b>&nbsp;Joe Clarke (jclarke) &lt;[email protected]&gt;
<br>
<b>Sent:</b>&nbsp;Thursday, June 25, 2026 8:27 PM<br>
<b>To:</b>&nbsp;Chongfeng Xie &lt;[email protected]&gt;; Graf Thomas, =
SCS-INI-NET-VNC-E2E &lt;[email protected]&gt;; ops-area &lt;ops-area=
@ietf.org&gt;; opsawg &lt;[email protected]&gt;; nmop &lt;[email protected]&gt;; =
[email protected]<br>
<b>Cc:</b>&nbsp;[email protected]<br>
<b>Subject:</b>&nbsp;Re: [nmrg] Re: [NMOP] Re: [OPSAREA@IETF126] IRTF/IETF =
OPS Transfer: Status &amp; Exploring Collaboration Opportunities</span></p>
</div>
<p class=3D"MsoNormal" style=3D"margin: 0in; font-family: Aptos, sans-serif=
; font-size: 12pt;">
&nbsp;</p>
<table class=3D"MsoNormalTable" align=3D"left" cellspacing=3D"0" cellpaddin=
g=3D"0" border=3D"0" style=3D"width: 100%;">
<tbody>
<tr>
<td class=3D"MsoNormalTable" style=3D"background-color: rgb(207, 74, 12); p=
adding: 5pt 2pt;">
</td>
<td class=3D"MsoNormalTable" style=3D"background-color: rgb(255, 248, 229);=
 padding: 5pt 4pt 5pt 12pt; width: 100%;">
<p class=3D"MsoNormal" style=3D"margin: 0in; font-family: Aptos, sans-serif=
; font-size: 12pt;">
<span style=3D"color: rgb(34, 34, 34);"><b>Be aware:</b>&nbsp;This is an ex=
ternal email.</span></p>
</td>
</tr>
</tbody>
</table>
<p class=3D"MsoNormal" style=3D"margin: 0in; font-family: Aptos, sans-serif=
; font-size: 12pt;">
&nbsp;</p>
<p class=3D"MsoNormal" style=3D"margin: 0in; font-family: Aptos, sans-serif=
; font-size: 12pt;">
<span style=3D"color: black;">Chongfeng, I mostly agree with everything you=
 said. &nbsp;That is, I have a similar understanding. &nbsp;I do have a few=
 comments, though, based on my implementation experience.</span></p>
<p class=3D"MsoNormal" style=3D"margin: 0in; font-family: Aptos, sans-serif=
; font-size: 12pt;">
<span style=3D"color: black;">&nbsp;</span></p>
<div id=3D"mail-editor-reference-message-container">
<div style=3D"padding: 3pt 0in 0in; border-width: 1pt medium medium; border=
-style: solid none none; border-color: currentcolor;">
<p class=3D"MsoNormal" style=3D"margin: 0in; font-family: Aptos, sans-serif=
; font-size: 12pt;">
<span style=3D"color: black;">&nbsp;</span></p>
</div>
<p class=3D"MsoNormal" style=3D"margin: 0in; font-family: Aptos, sans-serif=
; font-size: 12pt;">
Q1. Which AI agents capabilities are suitable for network management?&nbsp;=
 &nbsp;</p>
<p class=3D"MsoNormal" style=3D"margin: 0in; font-family: Aptos, sans-serif=
; font-size: 12pt;">
&nbsp; &nbsp;A:&nbsp; &nbsp;AI agents have the following common capabilitie=
s,&nbsp;</p>
<p class=3D"MsoNormal" style=3D"margin: 0in; font-family: Aptos, sans-serif=
; font-size: 12pt;">
&nbsp;</p>
<p class=3D"MsoNormal" style=3D"margin: 0in; font-family: Aptos, sans-serif=
; font-size: 12pt;">
&nbsp; &nbsp; - Perception: Processes multi-modal inputs (text, images, aud=
io) and tracks context and state over time.</p>
<p class=3D"MsoNormal" style=3D"margin: 0in; font-family: Aptos, sans-serif=
; font-size: 12pt;">
&nbsp;</p>
<p class=3D"MsoNormal" style=3D"margin: 0in; font-family: Aptos, sans-serif=
; font-size: 12pt;">
<span style=3D"font-family: Arial, sans-serif; color: black;">[JMC] This de=
pends on model. &nbsp;Not all models can handle all types of input. &nbsp;A=
nd the model capabilities will influence overall cost.</span></p>
<p class=3D"MsoNormal" style=3D"margin: 0in; font-family: Aptos, sans-serif=
; font-size: 12pt;">
&nbsp;</p>
<p class=3D"MsoNormal" style=3D"margin: 0in; font-family: Aptos, sans-serif=
; font-size: 12pt;">
&nbsp; &nbsp; - Planning &amp; Reasoning: Breaks high-level goals into acti=
onable sub-tasks, dynamically re-plans when obstacles arise, and evaluates =
trade-offs before acting.</p>
<p class=3D"MsoNormal" style=3D"margin: 0in; font-family: Aptos, sans-serif=
; font-size: 12pt;">
&nbsp;</p>
<p class=3D"MsoNormal" style=3D"margin: 0in; font-family: Aptos, sans-serif=
; font-size: 12pt;">
&nbsp; &nbsp; - Tool Use &amp; Execution: Autonomously calls external tools=
=97search engines, APIs, calendars, code interpreters, and databases=97to t=
ake real action in the digital world (e.g., send emails, book flights, quer=
y data).</p>
<p class=3D"MsoNormal" style=3D"margin: 0in; font-family: Aptos, sans-serif=
; font-size: 12pt;">
&nbsp;</p>
<p class=3D"MsoNormal" style=3D"margin: 0in; font-family: Aptos, sans-serif=
; font-size: 12pt;">
<span style=3D"font-family: Arial, sans-serif; color: black;">[JMC] More sp=
ecifically, with network management, tools can access devices to get data a=
nd adjust configuration. &nbsp;In the former, what the agent can then do wi=
th the data it retrieves heavily depends
 on the model and its training data. &nbsp;For example, many frontier model=
s generally do well with basic network operational data. &nbsp;Adding on to=
 tool use, RAG or purpose-trained or tuned models, and you can get even bet=
ter results from such tools calls.</span></p>
<p class=3D"MsoNormal" style=3D"margin: 0in; font-family: Aptos, sans-serif=
; font-size: 12pt;">
&nbsp;</p>
<p class=3D"MsoNormal" style=3D"margin: 0in; font-family: Aptos, sans-serif=
; font-size: 12pt;">
&nbsp; &nbsp; - Learning &amp; Self-Correction: Detects errors, reflects on=
 outcomes, iterates on solutions, and utilizes both short-term and long-ter=
m memory to improve over time.</p>
<p class=3D"MsoNormal" style=3D"margin: 0in; font-family: Aptos, sans-serif=
; font-size: 12pt;">
&nbsp;</p>
<p class=3D"MsoNormal" style=3D"margin: 0in; font-family: Aptos, sans-serif=
; font-size: 12pt;">
<span style=3D"font-family: Arial, sans-serif; color: black;">[JMC] This is=
 where Digital Twin comes into place. &nbsp;You don=92t necessarily want a =
model to =93learn=94 as it=92s going through your production network. &nbsp=
;Testing its =93reasoning=94 in a twin environment makes
 it safer for doing that self-correction and iteration.</span></p>
<p class=3D"MsoNormal" style=3D"margin: 0in; font-family: Aptos, sans-serif=
; font-size: 12pt;">
&nbsp;</p>
<p class=3D"MsoNormal" style=3D"margin: 0in; font-family: Aptos, sans-serif=
; font-size: 12pt;">
All the capabilities above are useful for network management. With the unde=
rground support of LLM models, the following functionalities can be realize=
d by AI agents in the context of network management,</p>
<p class=3D"MsoNormal" style=3D"margin: 0in; font-family: Aptos, sans-serif=
; font-size: 12pt;">
&nbsp; &nbsp; &nbsp;&nbsp;</p>
<p class=3D"MsoNormal" style=3D"margin: 0in; font-family: Aptos, sans-serif=
; font-size: 12pt;">
&nbsp; &nbsp;- Intent-based auto-assignment: Automatically routes NOC ticke=
ts based on detected intent.</p>
<p class=3D"MsoNormal" style=3D"margin: 0in; font-family: Aptos, sans-serif=
; font-size: 12pt;">
&nbsp;</p>
<p class=3D"MsoNormal" style=3D"margin: 0in; font-family: Aptos, sans-serif=
; font-size: 12pt;">
&nbsp; &nbsp;- Intelligent root-cause analysis: Pinpoints fault origins wit=
h precision.</p>
<p class=3D"MsoNormal" style=3D"margin: 0in; font-family: Aptos, sans-serif=
; font-size: 12pt;">
&nbsp;</p>
<p class=3D"MsoNormal" style=3D"margin: 0in; font-family: Aptos, sans-serif=
; font-size: 12pt;">
<span style=3D"font-family: Arial, sans-serif; color: black;">[JMC] Eh, <i>=
maybe</i>. &nbsp;See above. &nbsp;I=92ve had models go off the rails on hyp=
er-specific problems where it didn=92t have a lot of relevant training data=
.</span></p>
<p class=3D"MsoNormal" style=3D"margin: 0in; font-family: Aptos, sans-serif=
; font-size: 12pt;">
&nbsp;</p>
<p class=3D"MsoNormal" style=3D"margin: 0in; font-family: Aptos, sans-serif=
; font-size: 12pt;">
&nbsp; &nbsp;- Actionable troubleshooting guidance:&nbsp; Generates step-by=
-step plans and recommended resolutions.</p>
<p class=3D"MsoNormal" style=3D"margin: 0in; font-family: Aptos, sans-serif=
; font-size: 12pt;">
&nbsp;</p>
<p class=3D"MsoNormal" style=3D"margin: 0in; font-family: Aptos, sans-serif=
; font-size: 12pt;">
&nbsp; &nbsp;- AI-driven orchestration:&nbsp; Enables expert-human collabor=
ation and automated execution of task reassignments/suspensions.</p>
<p class=3D"MsoNormal" style=3D"margin: 0in; font-family: Aptos, sans-serif=
; font-size: 12pt;">
&nbsp;</p>
<p class=3D"MsoNormal" style=3D"margin: 0in; font-family: Aptos, sans-serif=
; font-size: 12pt;">
&nbsp; &nbsp;- Continuous network optimization:&nbsp; Delivers ongoing tuni=
ng recommendations for wireless, core, and cloud-network resources across d=
iverse scenarios.</p>
<p class=3D"MsoNormal" style=3D"margin: 0in; font-family: Aptos, sans-serif=
; font-size: 12pt;">
&nbsp;</p>
<p class=3D"MsoNormal" style=3D"margin: 0in; font-family: Aptos, sans-serif=
; font-size: 12pt;">
&nbsp; &nbsp;- Automated recovery verification:&nbsp; Validates fault resol=
ution by querying network and perceptual metrics via LLM-orchestrated API c=
alls.</p>
<p class=3D"MsoNormal" style=3D"margin: 0in; font-family: Aptos, sans-serif=
; font-size: 12pt;">
&nbsp; &nbsp;&nbsp;</p>
<p class=3D"MsoNormal" style=3D"margin: 0in; font-family: Aptos, sans-serif=
; font-size: 12pt;">
Q2. What is the impact of an AI network management process to YANG informat=
ion, service, network and data modelling?</p>
<p class=3D"MsoNormal" style=3D"margin: 0in; font-family: Aptos, sans-serif=
; font-size: 12pt;">
&nbsp; &nbsp; A:&nbsp; &nbsp;AI and legcay network management models belong=
 to different domains,&nbsp; the specific manangement modelling is aganosti=
c to the AI network manangent process. In other words, AI system can proces=
s any YANG information, service, network and data modelling.&nbsp;&nbsp;</p=
>
<p class=3D"MsoNormal" style=3D"margin: 0in; font-family: Aptos, sans-serif=
; font-size: 12pt;">
&nbsp;</p>
<p class=3D"MsoNormal" style=3D"margin: 0in; font-family: Aptos, sans-serif=
; font-size: 12pt;">
<span style=3D"font-family: Arial, sans-serif; color: black;">[JMC] This ha=
s exactly been my experience. &nbsp;I=92ve had good luck with a YANG-centri=
c skill. &nbsp;Where things can get problematic is the token consumption XM=
L and JSON produce when using open weight models
 with tighter context windows.</span></p>
<p class=3D"MsoNormal" style=3D"margin: 0in; font-family: Aptos, sans-serif=
; font-size: 12pt;">
&nbsp;</p>
<p class=3D"MsoNormal" style=3D"margin: 0in; font-family: Aptos, sans-serif=
; font-size: 12pt;">
Q3.&nbsp; Is MCP (<span style=3D"color: rgb(70, 120, 134);"><a href=3D"http=
s://en.wikipedia.org/wiki/Model_Context_Protocol" data-outlook-id=3D"b0e5b1=
08-7d2c-47b4-8566-570773dc4e96" style=3D"color: rgb(70, 120, 134); margin-t=
op: 0px; margin-bottom: 0px;"><u>https://en.wikipedia.org/wiki/Model_Contex=
t_Protocol</u></a></span>)
 a suitable protocol between AI agents and IETF network management data sou=
rces and system components?</p>
<p class=3D"MsoNormal" style=3D"margin: 0in; font-family: Aptos, sans-serif=
; font-size: 12pt;">
&nbsp; &nbsp; &nbsp;A:&nbsp; Yes, as an open standard that connects AI mode=
ls to external data and tools, MCP enables AI assistants to securely read f=
iles, query databases, and execute actions across different applications.&n=
bsp; The likelihood of MCP being used between between AI
 agents and IETF network management data sources and system components is v=
ery high.&nbsp;</p>
<p class=3D"MsoNormal" style=3D"margin: 0in; font-family: Aptos, sans-serif=
; font-size: 12pt;">
In addition, other approaches can be used, such as APIs exposed by network =
controller, in this case, AI agents can call the APIs directly for network =
device configuration or data collection.&nbsp;</p>
<p class=3D"MsoNormal" style=3D"margin: 0in; font-family: Aptos, sans-serif=
; font-size: 12pt;">
&nbsp;</p>
<p class=3D"MsoNormal" style=3D"margin: 0in; font-family: Aptos, sans-serif=
; font-size: 12pt;">
<span style=3D"font-family: Arial, sans-serif; color: black;">[JMC] Agree c=
ompletely. &nbsp;This does work well. &nbsp;And MCP has considerable framew=
ork backing.</span></p>
<p class=3D"MsoNormal" style=3D"margin: 0in; font-family: Aptos, sans-serif=
; font-size: 12pt;">
&nbsp;</p>
<p class=3D"MsoNormal" style=3D"margin: 0in; font-family: Aptos, sans-serif=
; font-size: 12pt;">
Q4, How does an AI agent interact with Network Digital Twin, Intent-Based N=
etworking, SIMAP, Knowledge Graphs and YANG data in Message Brokers?</p>
<p class=3D"MsoNormal" style=3D"margin: 0in; font-family: Aptos, sans-serif=
; font-size: 12pt;">
&nbsp; A:&nbsp; &nbsp;It depends on the specific situation, there should be=
 no one-size-fits-all solution.</p>
<p class=3D"MsoNormal" style=3D"margin: 0in; font-family: Aptos, sans-serif=
; font-size: 12pt;">
&nbsp;</p>
<p class=3D"MsoNormal" style=3D"margin: 0in; font-family: Aptos, sans-serif=
; font-size: 12pt;">
<span style=3D"font-family: Arial, sans-serif; color: black;">[JMC] See abo=
ve on a thought on digital twin. &nbsp;And, yeah, I don=92t think we=92ll h=
ave a one-size fits all model.</span></p>
<p class=3D"MsoNormal" style=3D"margin: 0in; font-family: Aptos, sans-serif=
; font-size: 12pt;">
<span style=3D"font-family: Arial, sans-serif; color: black;">&nbsp;</span>=
</p>
<p class=3D"MsoNormal" style=3D"margin: 0in; font-family: Aptos, sans-serif=
; font-size: 12pt;">
<span style=3D"font-family: Arial, sans-serif; color: black;">Joe</span></p=
>
</div>
</div>
</body>
</html>

--_000_CH2PR11MB8867C24122AC20D9089E1286B8FA2CH2PR11MB8867namp_--


--===============0736854182616703133==
Content-Type: text/plain; charset="utf-8"
MIME-Version: 1.0
Content-Transfer-Encoding: base64
Content-Disposition: inline

X19fX19fX19fX19fX19fX19fX19fX19fX19fX19fX19fX19fX19fX19fX19fX18KT1BTLUFSRUEg
bWFpbGluZyBsaXN0IC0tIG9wcy1hcmVhQGlldGYub3JnClRvIHVuc3Vic2NyaWJlIHNlbmQgYW4g
ZW1haWwgdG8gb3BzLWFyZWEtbGVhdmVAaWV0Zi5vcmcK

--===============0736854182616703133==--