Taxonomy of LLM agent communication protocols reveals federated future
A new ArXiv paper by Linus Sander, Habtom Kahsay Gidey, and Alexander Lenz proposes a five-dimension taxonomy for classifying LLM agent communication protocols, finding that no single protocol can maximize versatility, efficiency, and portability simultaneously — pointing toward a federated, layered protocol stack.
Score breakdown
The taxonomy gives protocol designers and adopters a structured framework for navigating an otherwise fragmented interoperability landscape, while the finding that no single protocol can satisfy all constraints simultaneously reframes the field's goal from convergence to federation.
- 01The taxonomy covers five dimensions: counterparty, payload, interaction state, discovery mechanism, and schema flexibility.
- 02Nine actively maintained open-source protocols with demonstrable adoption were analyzed.
- 03The iterative method comprised five rounds: three empirical-to-conceptual and two conceptual-to-empirical.
Linus Sander, Habtom Kahsay Gidey, and Alexander Lenz present a technical taxonomy on ArXiv designed to bring order to the fragmented landscape of LLM agent communication protocols. The authors followed an established iterative methodology, defining the taxonomy's purpose, meta-characteristic, and ending conditions before conducting five iterations — three empirical-to-conceptual and two conceptual-to-empirical — across nine actively maintained open-source protocols with demonstrable adoption. The resulting taxonomy organizes protocols along five dimensions: counterparty, payload, interaction state, discovery mechanism, and schema flexibility.
Classification of the nine protocols surfaces several recurring architectural patterns.
Classification of the nine protocols surfaces several recurring architectural patterns. All sampled agent-to-agent protocols combine hybrid payloads with session-state persistence. Most protocols support multiple predefined schemas, while two negotiate schemas at runtime — a pattern the authors interpret as a trend toward greater schema flexibility. Decentralized discovery, by contrast, remains rare across the sampled protocols.
The paper's analysis draws two temporal conclusions. In the short term, there is convergence pressure toward protocols that unify agent-to-agent and agent-to-context (tool and data) communication. Over the long term, however, the authors argue that no single protocol is likely to simultaneously maximize versatility, efficiency, and portability, and that the field will more probably evolve toward a federated, layered protocol stack. The taxonomy is offered as a practical framework for protocol selection and as a map of open research gaps, including privacy and policy enforcement.
Key facts
- 01The taxonomy covers five dimensions: counterparty, payload, interaction state, discovery mechanism, and schema flexibility.
- 02Nine actively maintained open-source protocols with demonstrable adoption were analyzed.
- 03The iterative method comprised five rounds: three empirical-to-conceptual and two conceptual-to-empirical.
- 04All sampled agent-to-agent protocols combine hybrid payloads with session-state persistence.
- 05Two protocols negotiate schemas at runtime, indicating a trend toward schema flexibility.
- 06Decentralized discovery remains rare among the sampled protocols.
- 07The authors conclude no single protocol can maximize versatility, efficiency, and portability simultaneously, predicting a federated, layered protocol stack.
Topics
Summary and scoring are generated automatically from the original article. We always link back to the publisher and never republish images or paywalled content. Last processed Jun 18, 2026 · 10:40 UTC. How this works →