ناقل الأحداث ووسطاء الرسائل

In a choreography saga, services communicate by exchanging events. For this to work reliably in a distributed system, we need robust communication mechanisms.

Imagine an order being created, then needing payment, and finally shipping. Each step is an event triggering the next service.

• Order Service publishes OrderCreated
• Payment Service subscribes to OrderCreated, publishes PaymentProcessed
• Shipping Service subscribes to PaymentProcessed, publishes ShipmentScheduled

An Event Bus is a pattern that enables different parts of an application (or different services) to communicate by sending and receiving events without knowing about each other directly.

Think of it as a central hub where events are broadcast. Any interested party can 'listen' for specific events.

It decouples services, meaning they don't need to know who will process their events, only that an event occurred.

Consider an internal event bus within a single application:

• A user clicks a 'Save' button.
• The UI component publishes a UserSaved event to the event bus.
• A logging module, a notification module, and a data synchronization module might all subscribe to UserSaved.

They all react to the event without the UI knowing about them. This is simple and effective for in-process communication.

While an event bus is great for in-process communication, distributed microservices need more. This is where Message Brokers come in.

A message broker is a dedicated service (like RabbitMQ, Kafka, or AWS SQS) that acts as an intermediary for messages between different applications or services.

It provides advanced features essential for reliable communication across networks.

Message brokers offer crucial benefits for choreography sagas:

• Durability: Messages persist even if services are down, preventing data loss.
• Guaranteed Delivery: Ensures messages reach their intended consumers.
• Decoupling: Senders and receivers don't know each other, promoting service independence.
• Load Balancing: Distributes messages efficiently among multiple instances of a consuming service.
• Asynchronous: Services don't wait for responses, improving performance and responsiveness.

While both facilitate event-driven communication, an Event Bus is typically a lightweight, in-memory mechanism within a single application or process.

A Message Broker is a robust, external system designed for inter-process and inter-service communication across a network. It provides features like message queues, topics, persistence, and complex routing.

You can think of a message broker as a powerful, distributed implementation of an event bus concept.

Message brokers are the backbone of choreography sagas. They allow services to:

• Publish Events: A service completes a step and publishes an event to a specific topic (a named channel on the broker).
• Subscribe to Events: Other services interested in that event subscribe to the topic and receive the event to trigger their next step.

This decentralized flow, driven by events through a broker, defines the choreography pattern.

Here's a simplified example of how an OrderService might prepare and conceptually "publish" an OrderCreated event using a message broker.

The broker handles the actual sending to subscribers.

Now, let's see how a PaymentService would conceptually "subscribe" to and process the OrderCreated event from the broker.

It listens on the order_events topic.

Message brokers are crucial for choreography sagas. Which of the following are key benefits provided by message brokers for reliable asynchronous communication?

In this lesson, we explored how Event Buses and Message Brokers facilitate reliable asynchronous communication, especially for choreography sagas.

• An Event Bus is a pattern for in-process event communication.
• A Message Broker is an external, robust system that implements this pattern for distributed services, offering features like durability and guaranteed delivery.
• They enable services to publish events and subscribe to topics, forming the core of how choreography sagas progress.

Next, we'll look at how to handle compensation when things go wrong in these event-driven flows!