Event Driven Architecture

1. Overview

EDA is a software architecture paradigm that promotes the production, detection, consumption of, and reaction to events.
Events can be defined as significant occurrences or changes in the system's state that can be originated internally or externally.
EDA is commonly used in systems where decoupled components need to communicate asynchronously.

Events: Key occurrences or changes to which the architecture is responsive. Events can trigger actions or workflows.
Event Producers: Entities that generate and publish events. They are responsible for notifying the system when something significant happens.
Event Consumers: Components that listen for and handle events. They react based on the event's information.
Event Channel/Bus: A medium through which events are transmitted from producers to consumers, often enabling a publish-subscribe model.
Event Processing: Logic that transforms or directs events as necessary to facilitate actions within the system.

Callbacks are functions that are passed as arguments to other functions and are invoked once a specific event occurs or the task is complete.
They are primarily used for asynchronous operations, allowing a program to continue without blocking waiting for an event or result.

Hooks are predefined points within a system where custom code or functions can be inserted to influence the behavior of the system at runtime.
They allow for modification and extension of existing functionality without altering the core codebase, often used in plugin architectures.

EDA uses both callbacks and hooks to efficiently manage and respond to events.
Callbacks provide a way to react to asynchronous events immediately, often simplifying the chaining of operations.
Hooks provide extensibility and customization, ensuring that various components or extensions can react to events by inserting additional logic.

Operate on the basis of requests and responses, typically client-server models.
The flow of operations is initiated by a client request, leading to synchronous processing.
Suitable for operations requiring immediate feedback or results.
Often involves direct interaction patterns such as REST or RPC.

Operate on events that are emitted as significant state changes.
Asynchronous and decoupled, enabling components to react independent of each other.
Suitable for scenarios requiring real-time data processing and distributed system responsiveness.
Use indirect interaction patterns like message queues and publish-subscribe systems.

Synchronous vs. Asynchronous: Request-driven is typically synchronous, while event-driven is inherently asynchronous.
Initiation: Request-driven is initiated by a specific client request, whereas event-driven is initiated by system or external events.
Decoupling: EDA emphasizes a higher level of decoupling between components than request-driven systems.
Scalability and Flexibility: EDA often provides greater scalability and flexibility, as components can evolve independently.