The Cost of Saying Hello: An Empirical Study of Handshake Overhead in AI Agent Communication Protocols

arXiv:2610.30151 [cs.AI] | Submitted 12 August 2026 | 15 pages, 4 figures, 4 tables

Author: Pranav Varshney

Subjects: Artificial Intelligence (cs.AI); Machine Learning (cs.LG)

DOI: https://doi.org/10.48550/arXiv.2610.30151

Code and data: https://github.com/pvarshh/handshake-benchmark

Abstract

Every interaction between AI agents begins with a handshake: the initiating agent must locate its counterpart, learn what it can do, and establish a session before the first useful message can be sent. Surveys of agent communication protocols name handshake overhead as an efficiency criterion, yet none report measurements. This paper presents the first empirical, phase-decomposed benchmark of the handshake in three major agent protocols: MCP (in both its 2025 initialize and 2026 discover eras), A2A, and ACP. Using each protocol's official SDK, we instrument every request for round trips, per-phase latency, wire bytes, and token cost across cold-start, warm-repeat, capability-mismatch, and degraded-metadata scenarios (N=50 per cell, on loopback and under an emulated 50 ms RTT).

Round-trip count dominates handshake cost. A legacy MCP handshake takes three round trips and 230 ms at 50 ms RTT; A2A needs one round trip and 61.0 ms. Round trips are not the whole bill: when an LLM reads the counterpart's metadata, handshake cost scales with the counterpart's capability inventory at 24-107 tokens per capability, so a 200-capability counterpart spends 11% of a 200k-token context window before the agent reads its own task. MCP's own 2026 revision cuts its handshake from three round trips to two and, with era pinning, to zero on warm reconnects—protocol evolution that corroborates the cost we measure.

We release the full harness, raw data, and analysis at https://github.com/pvarshh/handshake-benchmark.

Key Findings

Round-Trip Count Is the Dominant Cost Driver

Legacy MCP (2025 initialize): 3 round trips, 230 ms at 50 ms RTT
A2A: 1 round trip, 61.0 ms at 50 ms RTT
MCP 2026 revision: reduced from 3 to 2 round trips; zero round trips on warm reconnects with era pinning

Metadata Consumption Adds Hidden Token Cost

When an LLM must read the counterpart's metadata, handshake cost scales with the size of the counterpart's capability inventory:

24-107 tokens per capability
A 200-capability counterpart consumes 11% of a 200k-token context window before the agent even reads its own task

Protocol Evolution Corroborates the Measurement

MCP's 2026 revision—cutting handshake round trips from three to two, and to zero on warm reconnects with era pinning—directly corroborates the overhead measured in this study.

Methodology

Protocols benchmarked: MCP (2025 initialize and 2026 discover eras), A2A, ACP
Implementation: Each protocol's official SDK, with instrumentation on every request
Metrics captured: Round trips, per-phase latency, wire bytes, token cost
Scenarios: Cold-start, warm-repeat, capability-mismatch, degraded-metadata
Sample size: N=50 per cell
Network conditions: Loopback and emulated 50 ms RTT

Resources

The full harness, raw per-run data, and analysis are publicly available at https://github.com/pvarshh/handshake-benchmark.

Citation

Varshney, P. (2026). The Cost of Saying Hello: An Empirical Study of Handshake 
Overhead in AI Agent Communication Protocols. arXiv:2610.30151 [cs.AI].
https://doi.org/10.48550/arXiv.2610.30151

via ArXiv AI

The Cost of Saying Hello: An Empirical Study of Handshake