Characterizing Network Centralization and Observability in the Remote MCP Ecosystem
A measurement study of 179 remote MCP servers finds heavy infrastructure concentration (HHI 0.736) and a security-observability tradeoff in platform OAuth.
The paper introduces a three-tier observability framework (catalog metadata, passive compliance signals, live vulnerability analysis) applied to a stratified sample of 179 remote Model Context Protocol (MCP) endpoints from two public registries. The Herfindahl-Hirschman Index over ASN distribution is 0.736, well above the 0.25 high-concentration threshold, and 95% of commercial PaaS-hosted servers enforce gateway-level OAuth 2.1 with PKCE. Authentication correlates strongly with hosting platform choice rather than operator configuration, creating a security-observability tradeoff that constrains automated scanning for tool-poisoning vectors without prior credentials.
- HHI of 0.736 over ASN distribution shows heavy concentration in remote MCP hosting.
- 95% of commercial PaaS-hosted MCP servers enforce gateway-level OAuth 2.1 with PKCE.
- Authentication is strongly determined by hosting platform choice, not operator configuration.
- Platform-level security limits automated tool-poisoning scanning without prior credentials.
From the original article
1 min read at arxiv.orgThe Model Context Protocol (MCP) has emerged as the dominant interface for connecting autonomous agents to external data sources and execution environments. The ecosystem's transition from local process execution to remote Streamable HTTP deployments introduces unmeasured architectural and security constraints at scale.
By Muhammad Abdullah Sohail · arxiv.org
A short excerpt; the article and its images belong to the publisher. Summary, key facts and assessment above are AI-written from the article and can be wrong.