Usable Agent Discovery for Decentralized AI Systems

TL;DR

This paper investigates decentralized agent discovery in large-scale AI systems, analyzing how different overlay structures perform under various churn scenarios to optimize robustness and efficiency.

Contribution

It provides a comparative analysis of structured and gossip-based overlays for agent discovery under two-level churn, highlighting their respective strengths and trade-offs.

Findings

Structured overlays excel in stability and efficiency during node churn.

Gossip-based overlays are competitive and faster when service readiness is prioritized.

Different overlay strategies are optimal depending on churn and performance requirements.

Abstract

Large-scale agentic systems run on distributed infrastructures where many software agents share physical hosts and are discovered via peer-to-peer mechanisms. Discovery must handle node-level churn from failures and host departures and agent-level churn from demand-driven activation, deactivation, and state changes. Their interaction reshapes classic trade-offs between structured and unstructured overlays. We study decentralized agent discovery under this two-level churn, assuming nodes host multiple agents, overlays are structured or gossip-based, and agents switch between warm and cold states. Using Kademlia as a structured and Cyclon+Vicinity as a gossip baseline, we compare stable, node-churn-only, agent-cooling-only, and combined regimes to see when routing efficiency, resilience, and service readiness align or favor different designs. Structured overlays are more robust and efficient in stable and node-churn regimes, while gossip-based overlays remain competitive and can be faster when readiness dominates.