Zero Trust AI Agents: OpenZiti A2A MCP

Описание: What happens when Zero Trust networking meets AI-powered agents?

In this session, we introduce the fusion of OpenZiti, a powerful overlay network enabling zero-trust access, with the A2A (Agent-to-Agent) protocol — enabling AI agents to communicate, delegate, and collaborate securely by design.

🚀 Highlights:

How A2A lets agents act autonomously across networks

Securing AI workflows using OpenZiti's identity-based access

Enforcing trust boundaries in distributed, intelligent systems

Live demo: A LangGraph-based agent protected by OpenZiti, handling infrastructure queries through pyATS

🔐 Implications:

Zero trust for AI agents = no open ports, no VPNs, no lateral movement

Agents only communicate through encrypted, authenticated overlay

Redefining secure AI orchestration for ops, security, and automation

🧠 Why this matters:
As AI agents evolve from local tools to autonomous networked collaborators, security can’t be an afterthought. OpenZiti and A2A bring identity, policy, and control to this next frontier — ensuring your AI agent only talks to who it should.