Trade-offs in Decentralized Agentic AI Discovery Across the Compute Continuum

TL;DR

This paper compares different decentralized overlay networks like Chord, Pastry, and Kademlia for agent discovery across cloud and edge environments, analyzing their trade-offs in reliability, startup, and overhead.

Contribution

It provides a systematic comparison of major structured-overlay families for decentralized agent discovery in diverse compute environments.

Findings

Kademlia offers lower control-plane overhead.

Chord shows higher discovery reliability in churn scenarios.

Pastry balances startup time and reliability.

Abstract

Agentic systems deployed across the compute continuum need discovery mechanisms that remain effective across cloud, edge, and intermittently connected domains. In some emerging agentic architectures, decentralized discovery is already an active design direction, placing DHT-based lookup on the path toward agent directories. This paper studies the trade-offs among major structured-overlay families for agent discovery, comparing Chord, Pastry, and Kademlia as candidate indexing substrates within a shared control-plane framework. Using a benchmark subset centered on a 4096-node stationary comparison and a representative 4096-node churn benchmark, the paper characterizes how discovery reliability, startup behavior, and control-plane overhead vary across these overlays. The goal is to clarify the operating points they expose for agent discovery across edge-to-cloud environments.