MCP & A2A

Link

Khái nhiệm và kiến trúc

Agentic MCP (Model Context Protocol) is a protocol that helps AI models, like chatbots, connect to external systems such as databases or business tools. It’s like giving the AI a set of tools to fetch information or perform tasks during a conversation, making it smarter and more helpful.

A2A (Agent-to-Agent) Architecture is about letting different AI agents talk to each other. Imagine a team where each agent has a job, like one handles customer queries and another manages tickets — they can work together smoothly using A2A.

MCP

Core Components

MCP follows a client-server architecture with the following key components:

  1. MCP Hosts — Programs using LLMs (like Claude Desktop or IDEs) that initiate connections to access external data and tools
  2. MCP Clients — Protocol clients embedded within the host application that maintain 1:1 connections with servers
  3. MCP Servers — Lightweight programs that expose specific capabilities through the standardized protocol
  4. Data Sources — Both local (files, databases) and remote services (APIs) that MCP servers can access

Works

MCP standardizes three main types of capabilities:

  1. Tools — Functions that can be called by the LLM (with user approval)
  2. Resources — File-like data that can be read by clients (API responses, file contents)
  3. Prompts — Pre-written templates to help users accomplish specific tasks

SSE & STDIO

A2A Architecture

Core components

A2A architecture centers around facilitating communication between agents with these key components:

  1. Client Agent — Formulates tasks and communicates them to remote agents
  2. Remote Agent — Acts on tasks to provide information or perform actions
  3. Agent Card — JSON metadata file describing an agent’s capabilities and endpoints
  4. Task Management — Defines task objects with lifecycle stages and outputs
  5. Messaging System — Allows agents to exchange context, replies, and artifacts