[nmrg] Re: Request for Review on NMRG Draft - Use Cases and Practices for Intent-Based Networking

["Mr. Jaehoon Paul Jeong" <[e-mail]>]

Hi Houda,
I have submitted the revision (05) that addresses your comments about
the governance and accountability for AI agents:
https://datatracker.ietf.org/doc/html/draft-irtf-nmrg-ibn-usecases-05

You can see the differences between -04 and -05 on the webpage above:
https://author-tools.ietf.org/iddiff?url1=draft-irtf-nmrg-ibn-usecases-04&url2=draft-irtf-nmrg-ibn-usecases-05&difftype=--html

Please let us know if you have further comments.

Thanks.

Best Regards,
Paul


On Thu, Oct 29, 2026 at 10:02 PM Houda CHIHI <[e-mail]> wrote:

> That's Great !
> ------------------------------
> *De :* Mr. Jaehoon Paul Jeong <[e-mail]>
> *Envoyé :* jeudi 29 octobre 2026 13:56
> *À :* Houda CHIHI <[e-mail]>
> *Cc :* nmrg <[e-mail]>; skku-iotlab-members <
> [e-mail]>; Mr. Jaehoon Paul Jeong <
> [e-mail]>
> *Objet :* Re: [nmrg] Re: Request for Review on NMRG Draft - Use Cases and
> Practices for Intent-Based Networking
>
> Houda,
> Thanks for your provided references and suggestion.
>
> I will revise this draft accordingly with the new references and your
> comments.
>
> Thanks.
>
> Best Regards,
> Paul
>
> 2026년 10월 29일 (목) 오후 9:52, Houda CHIHI <[e-mail]>님이 작성:
>
> Mr. Jaehoon Paul Jeong Hello,
>
>
> Thank you for the revision and for incorporating agent governance into
> Section 5.3.2 and Section 6. The additions address the concern I raised.
> As requested, here are suggested references for the Informative References
> section.
> *Byzantine fault tolerance (Section 5.3.2):*
> L. Zheng et al., "Rethinking the Reliability of Multi-Agent Systems: A
> Perspective from Byzantine Fault Tolerance," Proc. AAAI, vol. 40, no. 41,
> Jun. 2026.
> M. Castro and B. Liskov, "Practical Byzantine Fault Tolerance," Proc.
> OSDI, USENIX, 1999, pp. 173–186.
> *Trusted agent technologies / pre-verification (Section 5.3.2):*
> TM Forum, "Zero-Trust Agents for Autonomous Networks," Catalyst C26.0.933,
> 2026.
> S. Prakash, "Agent Identity Protocol (AIP): Verifiable Delegation for AI
> Agent Systems," Internet-Draft draft-prakash-aip-01, IETF, Aug. 2026.
> *Security considerations (Section 6):*
> "A Concise Review of Security Challenges and Protection for LLM-based
> Agents in Telecom Information Systems," Proc. IEEE Conf., Guangzhou, China,
> Mar. 2026.
> NGMN Alliance, "Network Automation and Autonomy Phase III: Agentic AI for
> Autonomous Mobile Networks," v1.0, 11 Aug. 2026.
> *Optional, to align with active NMRG work:*
> A. Clemm and T. Eckert, "Agentic AI for Intent-Based Networking,"
> Internet-Draft draft-cxxx-nmrg-ai4ibn-00, IRTF NMRG, Jul. 2026.
> The TM Forum Catalyst and the IEEE review are the strongest anchors: the
> former defines the IPOE model (Identity, Policy, Observability, Evidence)
> that matches your "pre-verification + manual approval" text, and the latter
> identifies prompt injection, excessive agency, and data exposure as the
> primary telecom agent risks, which maps directly to your Section 6 concerns.
> Happy to review the -05 revision once posted.
> Best regards,
>
> Houda CHIHI, PhD
>
> ------------------------------
> *De :* Mr. Jaehoon Paul Jeong <[e-mail]>
> *Envoyé :* jeudi 29 octobre 2026 13:38
> *À :* Houda CHIHI <[e-mail]>
> *Cc :* nmrg <[e-mail]>; skku-iotlab-members <
> [e-mail]>; Mr. Jaehoon Paul Jeong <
> [e-mail]>
> *Objet :* Re: [nmrg] Re: Request for Review on NMRG Draft - Use Cases and
> Practices for Intent-Based Networking
>
> Hi Houda,
> Thanks for your input and suggestion regarding the comment: "Agent
> governance and accountability are not addressed".
>
> This revision addresses agent governance concerning fault tolerance and
> cybersecurity attacks as follows.
> ---
> 5.3.2. AI Agent-Enabled IBN
> System reliability and security: This dimension considers the stability
> and cybersecurity. This not only includes the logical verification of
> instructions generated by AI Agents (e.g., Domain-Specific Language (DSL)
> syntax compliance) but also involves issues such as data privacy
> protection, malicious behavior identification, and Byzantine fault
> tolerance in agent interactions. Note that the Byzantine fault tolerance
> allows the IBN systems to keep operating correctly or reach right consensus
> even if some components of the IBN systmes are malicious or faulty.
>
> Trusted agent technologies: It includes a multi-layer verification
> mechanism of “agent pre-verification” and “manual approval” and also
> abnormal behavior detection algorithms based on traffic fingerprints.
>
> ---
> 6. Security Considerations
>
> A malicious intent can break down the whole IBN system if the analysis of
> each intent for the impact on the IBN system is not performed appropriately
> on time. Thus, the IBN defense system should execute both the security
> check for an intent during the Fulfillment Phase and the attack monitoring
> during the Assurance Phase. Such security check and attack monitoring need
> to be performed by the collaboration between AI agents and network
> administrators.
>
> The security for SDVs is tightly associated with the safety of drivers and
> passengers in the SDVs. The SDVs compromised by a hacker may malfunction
> and lead to road accidents. Thus, the IBN for SDVs needs more robust
> defense against various cybersecurity attacks than the IBN for other
> domains. An SDV security defense system includes authentication and
> authorization of an intent issuer (e.g., SDV User), the validation of an
> intent prior to policy enforcement, and the prediction of the possible
> results by the policy enforcement. These security considerations for the
> SDVs can also be beneficial to the IBN systems for other domains.
> ---
>
> Could you suggest some text with reference information for the Informative
> Reference Section?
>
> I will incorporate your text into the -05 version.
>
> Thanks.
>
> Best Regards,
> Paul
>
>
> On Thu, Oct 29, 2026 at 8:57 PM Houda CHIHI <[e-mail]> wrote:
>
> Dear All;
>
>
> Hope that you are doing well, I have read draft-irtf-nmrg-ibn-usecases-04
> with interest. I would like to raise one substantive comment regarding
> Section 5.3.2 ("AI Agent-Enabled IBN").
> *Comment: Agent governance and accountability are not addressed.*
> Section 5.3.2 describes AI agents as enablers of intent-based networking
> but treats the agent as a black box. When an AI agent translates an intent
> incorrectly, deploys a harmful policy, or executes an action that violates
> operator expectations, the draft does not specify who is accountable or how
> the failure is detected and remediated.
> This is not a minor omission. The draft's central claim is that AI agents
> operationalize the intent lifecycle. But an agent without identity, audit,
> and accountability is not an enabler — it is a risk. Without governance,
> AI-enabled IBN remains a demonstration rather than a deployable system.
> I suggest the draft specify minimum governance requirements, for example:
>
>    1. *Persistent, verifiable agent identity* across interactions and
>    domains;
>    2. *Immutable audit trails* of every agent decision and action;
>    3. *Explicit human-in-the-loop gates* for consequential actions;
>    4. *Rollback and remediation procedures* for agent error.
>
> Relevant ongoing work that the draft could reference includes the IETF
> Agent Identity Protocol (AIP) draft, TM Forum's Zero-Trust Agents work, and
> the NGMN Phase III report on Agentic AI, all of which address these
> requirements directly.
> I am happy to provide further detail or a short companion draft on agent
> governance requirements for IBN if the authors find it useful.
> Best regards,
> Houda CHIHI, PhD
> ------------------------------
> *De :* Hong Yong-Geun <[e-mail]>
> *Envoyé :* jeudi 29 octobre 2026 12:33
> *À :* Mr. Jaehoon Paul Jeong <[e-mail]>
> *Cc :* nmrg <[e-mail]>; skku-iotlab-members <
> [e-mail]>
> *Objet :* [nmrg] Re: Request for Review on NMRG Draft - Use Cases and
> Practices for Intent-Based Networking
>
> Dear Paul.
>
> I can confirm that the new version takes my comments into account.
>
> Thank you for your hard work.
>
> Best regards,
>
> Yong-Geun.
>
> *보낸 사람: *Mr. Jaehoon Paul Jeong <[e-mail]>
> *날짜: *수요일, 2026년 10월 28일 오전 12:14
> *받는 사람: *Yong-Geun Hong <[e-mail]>
> *참조: *nmrg <[e-mail]>; skku-iotlab-members <
> [e-mail]>; Mr. Jaehoon Paul Jeong <
> [e-mail]>
> *주제: *Re: [nmrg] Request for Review on NMRG Draft - Use Cases and
> Practices for Intent-Based Networking
>
> Hi Yong-Geun,
> I have addressed all your comments and submitted the revised draft:
> https://datatracker.ietf.org/doc/html/draft-irtf-nmrg-ibn-usecases-04
>
> I attached the revision letter.
>
> Could you check whether my revision satisfies your comments?
>
> If you have more comments, please let me know.
>
> Thanks.
>
> Best Regards,
> Paul
>
> On Tue, Jul 14, 2026 at 7:55 AM Yong-Geun Hong <[e-mail]>
> wrote:
>
> Dear Paul,
>
> Having read the document, I thought it was well written and ready for the
> next stage.
>
> Please find attached my review, which I hope will help you to improve it
> further.
>
> Best regards,
>
> Yong-Geun.
>
> *보낸 사람: *Mr. Jaehoon Paul Jeong <[e-mail]>
> *날짜: *화요일, 2026년 7월 7일 오전 12:41
> *받는 사람: *King, Daniel <[e-mail]>; Aijun Wang <
> [e-mail]>; Dr. Byoungman Robert An <[e-mail]>;
> Abdelkader Mekrache <[e-mail]>; Alexander Clemm <
> [e-mail]>; Prof. Yong-Geun Hong <[e-mail]>; Prof. Joo-Sang
> Youn <[e-mail]>; Bien Aime Mugabarigira <[e-mail]>
> *참조: *nmrg <[e-mail]>; Jérôme FRANCOIS <[e-mail]>;
> skku-iotlab-members <[e-mail]>
> *주제: *[nmrg] Request for Review on NMRG Draft - Use Cases and Practices
> for Intent-Based Networking
>
> Hi Daniel, Aijun, Robert, Abdel, Alex, Yong-Geun, Joo-Sang, Bien Aime, and
> NMRG,
> Could you review the NMRG draft titled "Use Cases and Practices for
> Intent-Based Networking" by July 13, 2026?
> https://datatracker.ietf.org/doc/html/draft-irtf-nmrg-ibn-usecases-03
>
> In the last IETF-125 NMRG session, I presented the revised draft after
> addressing all of Jerome's comments.
> - Slides in IETF-125 NMRG Session
>
> https://datatracker.ietf.org/doc/slides-125-nmrg-use-cases-and-practices-for-intent-based-networking/
>
> I believe this draft is mature for the RG last call.
>
> Before requesting the RG last call in the IETF-126 NMRG session,
> I would like you to review this draft for further improvement and share
> your valuable comments with us.
>
> Thanks in advance.
>
> Best Regards,
> Paul
> --
> ===========================
> Mr. Jaehoon (Paul) Jeong
> Professor
> Department of Computer Science and Engineering
> Sungkyunkwan University
> Mobile: [phone] <+82%2010-4758-1765>
> Phone: [phone] <+82%2031-299-4957>
> Email: [e-mail], [e-mail]
> URI: http://iotlab.skku.edu/people-jaehoon-jeong.php
> LinkedIn: https://www.linkedin.com/in/jaehoonjeong/
>
>