Build composable, interoperable multi-agent systems with A2A on AWS
Build a composable, interoperable multi-agent system on AWS with the Agent-to-Agent (A2A) protocol. A LangChain orchestrator on Amazon ECS communicates with a Strands agent on Amazon Bedrock AgentCore across organizational boundaries, using bidirectional, OAuth-authenticated calls, with neither side depending on the other's framework or SDK.
First seen by Oriel on .
AI summary
This AWS reference architecture demonstrates interoperable, bidirectional multi-agent collaboration across organizational and framework boundaries using A2A, Agent Cards, and OAuth without shared secrets or exposed agent internals.
- The enterprise runs a LangChain orchestrator and context agent on private AWS Fargate subnets, while the Strands Agents SDK compliance agent runs in Amazon Bedrock AgentCore.
- At startup, the orchestrator reads the SaaS Agent Card and creates one LangChain tool per published skill, allowing model-driven routing without hardcoded integration logic.
- Each organization uses its own Amazon Cognito pool; onboarding exchanges only public URLs, client IDs, scopes, and endpoints, while OAuth bearer tokens authenticate both directions.
- The compliance agent verifies the registered context Agent Card, calls back for policies and account metadata, then performs synthetic scans without exposing internal tools or prompts.
- Three Jupyter notebooks deploy the system in roughly 28 minutes across two same-Region AWS accounts; always-on NAT, AWS Fargate, load balancer, and VPC endpoints incur charges.
#a2a #multi-agent systems #aws #oauth
Generated by gpt-5.6-sol. Check the original for details.
Aliases
| Source | External ID | URL |
|---|---|---|
| AWS Machine Learning Blog | 9681452550bf3bb15b0e292427676f986d6d624a | https://aws.amazon.com/blogs/machine-learning/build-composable-interoperable-multi-agent-systems-with-a2a-on-aws/ Read discussion |