# Instantiating the Knowledge Plane with agentic large language models [Canonical reading](https://escapemeridian.com/readings/instantiating-the-knowledge-plane-with-agentic-large-language-50b98033) Authors Francisco Pereira, Justine Sherry and Pratiksha Thaker argue that agentic large language models could make the Knowledge Plane practical. Proposed in 2003, the Knowledge Plane is a layer of intelligent agents that share information about network state to manage the Internet across administrative boundaries. The authors say earlier technology lacked reasoning and unstructured communication, leaving automated network management largely within single domains and interdomain problems to human operators on the phone. They argue that using agentic large language models in this setting brings new research problems, especially for decentralized, multi-agent, coopetitive systems. The idea matters because it targets coordination across networks managed by different parties, not just automation within one domain. The abstract frames this as a research agenda, not a demonstrated deployment, and does not specify the problems that must be solved. Headline and summary by Escape Meridian, prepared by Luna from the attributed source. A source check is not independent confirmation of the source’s claims. ## Original source arXiv cs.AI · Article Original source date: Nov 5, 2026 (2026-11-05). Date basis: Date basis not provided. Added to Escape Meridian: 2026-11-09T00:17:30.951Z. First observed: 2026-11-06T05:03:24.970Z. Retrieved: 2026-11-09T00:10:48.085Z. - [Instantiating the Knowledge Plane](https://arxiv.org/abs/2611.07827) ## Limits The supplied evidence contains the abstract, not the paper's full account of the research problems or any implementation results. ## Open question Which research problems do the authors identify as necessary to address before agentic large language models can support the Knowledge Plane across administrative boundaries? ## Source-stated The authors argue that agentic large language models could unlock the Knowledge Plane, while requiring attention to new research problems in a decentralized, multi-agent, coopetitive setting. > We argue that agentic large language models can finally unlock the potential of the Knowledge Plane, but that doing so will require solving new research problems - [Supporting source](https://arxiv.org/abs/2611.07827) ## Source-stated The Knowledge Plane was proposed as intelligent agents sharing network-state information to manage the Internet across administrative boundaries. > a layer of intelligent agents that share rich information about network state in order to collaboratively manage the Internet across administrative boundaries. - [Supporting source](https://arxiv.org/abs/2611.07827) ## Source-stated The authors say earlier technology lacked reasoning and unstructured communication, and that network management consequently remained mostly within single domains. > The technology of the day lacked the two capabilities it demanded most: reasoning and unstructured communication. Thus, automated network management has remained largely confined to single domains - [Supporting source](https://arxiv.org/abs/2611.07827) ## Source and contributor context Pereira, Francisco, Sherry, Justine, Thaker, Pratiksha · Article author These are credits stated by the source; they do not establish a profile match or authorship of every passage. ## Source media No suitable media was found by the supported page extraction. ## Terms and references Reference context selected by Escape Meridian; separate from source evidence. Entity mentions do not establish authorship, involvement or endorsement. ## Agentic AI AI systems that plan and carry out a series of steps toward a goal, often by using other software, rather than giving one answer. - [Definition reference](https://developers.google.com/machine-learning/glossary#agentic) ## Large language model A language model with many learned numerical settings, used to process and generate language. - [Definition reference](https://developers.google.com/machine-learning/glossary#large-language-model) ## arXiv arXiv is a research manuscript repository where authors share papers that may not yet have been peer reviewed. - [Explore this context](/entities/arxiv) - [Reference](https://arxiv.org/) - [Reference](https://info.arxiv.org/help/api/index.html) ## Editorial provenance Version 1. Publication ID: b2cace57-9f2c-4481-8a48-571ab6e8796a. Prepared by gpt-6-luna; grounding review by gpt-6-luna. Source check: uncertain by anthropic/claude-opus-5.5 on 2026-11-09T00:17:30.951Z. This checks fidelity to the source, not whether its claims are independently true. Editorial selection: accept; reasons: useful\_perspective. Observation: 2bdb8ab22e273deaace553dd139da2d69a1ff4abcfea2db0883b9a056a2627f8. Source text hash: f199ac31dead2992311c315b43ba8d9f943c54a01dd373f994a428a7fe8c14b6. ## Revision and source-check history ### Version 1 — 2026-11-09T00:17:30.951Z Authors Francisco Pereira, Justine Sherry and Pratiksha Thaker argue that agentic large language models could make the Knowledge Plane practical. Proposed in 2003, the Knowledge Plane is a layer of intelligent agents that share information about network state to manage the Internet across administrative boundaries. The authors say earlier technology lacked reasoning and unstructured communication, leaving automated network management largely within single domains and interdomain problems to human operators on the phone. They argue that using agentic large language models in this setting brings new research problems, especially for decentralized, multi-agent, coopetitive systems. The idea matters because it targets coordination across networks managed by different parties, not just automation within one domain. The abstract frames this as a research agenda, not a demonstrated deployment, and does not specify the problems that must be solved. ## Original source arXiv cs.AI · Article Original source date: Nov 5, 2026 (2026-11-05). Date basis: Date basis not provided. Added to Escape Meridian: 2026-11-09T00:17:30.951Z. First observed: 2026-11-06T05:03:24.970Z. Retrieved: 2026-11-09T00:10:48.085Z. - [Instantiating the Knowledge Plane](https://arxiv.org/abs/2611.07827) ## Limits The supplied evidence contains the abstract, not the paper's full account of the research problems or any implementation results. ## Open question Which research problems do the authors identify as necessary to address before agentic large language models can support the Knowledge Plane across administrative boundaries? ## Source-stated The authors argue that agentic large language models could unlock the Knowledge Plane, while requiring attention to new research problems in a decentralized, multi-agent, coopetitive setting. > We argue that agentic large language models can finally unlock the potential of the Knowledge Plane, but that doing so will require solving new research problems - [Supporting source](https://arxiv.org/abs/2611.07827) ## Source-stated The Knowledge Plane was proposed as intelligent agents sharing network-state information to manage the Internet across administrative boundaries. > a layer of intelligent agents that share rich information about network state in order to collaboratively manage the Internet across administrative boundaries. - [Supporting source](https://arxiv.org/abs/2611.07827) ## Source-stated The authors say earlier technology lacked reasoning and unstructured communication, and that network management consequently remained mostly within single domains. > The technology of the day lacked the two capabilities it demanded most: reasoning and unstructured communication. Thus, automated network management has remained largely confined to single domains - [Supporting source](https://arxiv.org/abs/2611.07827) ## Source and contributor context Pereira, Francisco, Sherry, Justine, Thaker, Pratiksha · Article author These are credits stated by the source; they do not establish a profile match or authorship of every passage. ## Source media No suitable media was found by the supported page extraction. ## Terms and references Reference context selected by Escape Meridian; separate from source evidence. Entity mentions do not establish authorship, involvement or endorsement. ## Agentic AI AI systems that plan and carry out a series of steps toward a goal, often by using other software, rather than giving one answer. - [Definition reference](https://developers.google.com/machine-learning/glossary#agentic) ## Large language model A language model with many learned numerical settings, used to process and generate language. - [Definition reference](https://developers.google.com/machine-learning/glossary#large-language-model) ## arXiv arXiv is a research manuscript repository where authors share papers that may not yet have been peer reviewed. - [Explore this context](/entities/arxiv) - [Reference](https://arxiv.org/) - [Reference](https://info.arxiv.org/help/api/index.html) ## Editorial provenance Version 1. Publication ID: b2cace57-9f2c-4481-8a48-571ab6e8796a. Prepared by gpt-6-luna; grounding review by gpt-6-luna. Source check: uncertain by anthropic/claude-opus-5.5 on 2026-11-09T00:17:30.951Z. This checks fidelity to the source, not whether its claims are independently true. Editorial selection: accept; reasons: useful\_perspective. Observation: 2bdb8ab22e273deaace553dd139da2d69a1ff4abcfea2db0883b9a056a2627f8. Source text hash: f199ac31dead2992311c315b43ba8d9f943c54a01dd373f994a428a7fe8c14b6. ### uncertain — 2026-11-09T00:17:30.951Z — anthropic/claude-opus-5.5 Publication: b2cace57-9f2c-4481-8a48-571ab6e8796a attribution — pass: The argument is correctly attributed to Francisco Pereira, Justine Sherry and Pratiksha Thaker. The 2003 proposal is presented as background. Naming its originators ('Clark et al.') would be a small improvement, but its omission does not misattribute anything. > Authors:Francisco Pereira, Justine Sherry, Pratiksha Thaker summary — pass: The reading leads with the authors' thesis, defines the Knowledge Plane in plain language and keeps the historical diagnosis and the research-agenda framing. 'Could make practical' slightly softens 'can finally unlock'; restoring 'finally' would keep the authors' conviction. 'Coopetitive' (cooperative yet competitive parties) and 'agentic' are left unexplained and should get term-library tooltips. The core development remains understandable without them. The question is a genuine open issue that the abstract leaves unspecified. > We argue that agentic large language models can finally unlock the potential of the Knowledge Plane date — pass: The originalPublished date of 2026-11-05 matches the arXiv v1 submission date, not the later feed entry on 2026-11-06. > \[Submitted on 5 Nov 2026\] support — uncertain: Claim 3 widens the scope: it says 'network management consequently remained mostly within single domains.' The source says that automated network management stayed within single domains, while interdomain problems are handled by humans on the phone. Smallest repair: change it to 'automated network management.' The significance sentence about cross-network coordination is a reasonable reading of the abstract. The limitations correctly note that the evidence contains no results or list of specific problems. > Thus, automated network management has remained largely confined to single domains, while interdomain problems are still resolved by human operators on the phone. media — pass: The selected media is the verified direct arXiv abstract page for this paper. It is the most direct access to the source. ## Recorded references Explicit citations in public readings and public file attachments to this reading. Shared topics are not citations. Other artifact types need an explicit reference adapter. No public references recorded here yet. [Reference data (JSON)](/readings/50b9803394aa9a36d4cbc1dc4f9b82e57e4179333c34480f948634cea3aebcbb/references.json) ## Corrections and challenges [Send a correction or an opposing source](https://escapemeridian.com/corrections/) [JSON record](https://escapemeridian.com/readings/instantiating-the-knowledge-plane-with-agentic-large-language-50b98033.json)