Présentation du protocole STOMP

Welcome to STOMP! When building real-time apps with WebSockets, you often need more than just raw data transfer. You need structure, routing, and a way to subscribe to specific messages.

STOMP (Simple Text Oriented Messaging Protocol) provides this structure. Think of it as a higher-level protocol that sits on top of WebSockets.

With raw WebSockets, clients and servers send arbitrary bytes or strings. If you want to send a "chat message" or a "user joined" event, you have to define your own rules:

• How do I know what kind of message it is?
• Who should receive this message?
• What if I want to send data in a specific format like JSON?

Without a protocol, you'd implement all this manually.

STOMP solves these problems by defining a standard way to send and receive messages. It's like agreeing on a common language for your real-time communication.

• It's text-based, making it easy to debug.
• It uses a simple frame format for all communication.
• It supports common messaging patterns like publish-subscribe.

All communication in STOMP happens via "frames." A STOMP frame consists of three main parts:

• COMMAND: What action is being performed (e.g., SEND, SUBSCRIBE, CONNECT).
• HEADERS: Key-value pairs providing additional info (e.g., destination, content-type).
• BODY: The actual message payload.

These parts are separated by newlines.

When a client wants to establish a STOMP connection over a WebSocket, it sends a CONNECT frame. This tells the server it wants to speak STOMP. Note the empty line before the null byte (^@) which terminates the frame.

Here's what a simple CONNECT frame looks like:

If the server accepts the connection, it responds with a CONNECTED frame. This confirms the STOMP session is ready. It also includes headers like the server's version and session ID. Remember, frames end with a null byte (^@).

To send a message, clients use the SEND command. This frame requires a destination header, which tells the STOMP broker where to route the message. The message body contains the actual data, often JSON. Don't forget the null byte (^@) terminator!

Clients receive messages by using the SUBSCRIBE command. They specify a destination to listen to and an id for their subscription. The server then sends MESSAGE frames to this client when new data arrives for that destination. Frames are always null-byte (^@) terminated.

STOMP uses destinations to route messages. These are like addresses. The most common types are:

• Topics (/topic/...): For broadcast messages, where all subscribers receive a copy. Great for chat rooms.
• Queues (/queue/...): For point-to-point messages, typically handled by one consumer. Useful for private messages or task queues.

This built-in routing is a major advantage over raw WebSockets.

STOMP provides significant benefits for real-time applications:

• Structured Messaging: Standardized frame format.
• Message Routing: Built-in support for topics and queues.
• Subscription Model: Easy for clients to listen to specific events.
• Simpler Client-Side: Many client libraries exist that handle STOMP frames, simplifying development.

It adds a robust layer for complex messaging patterns.

Which of the following is a primary benefit of using STOMP over raw WebSockets?

In this lesson, you learned about the STOMP protocol and why it's essential for structured real-time communication over WebSockets.

• Raw WebSockets lack built-in message structure and routing.
• STOMP provides a simple, text-based frame format (COMMAND, HEADERS, BODY).
• Key STOMP commands include CONNECT, CONNECTED, SEND, and SUBSCRIBE.
• It enables powerful features like topic-based broadcasting and queue-based point-to-point messaging.

Next, we'll see how to integrate STOMP with Spring Boot!