Message Brokers

1. Why

robustness and scalability require decoupled components
message brokers act as asynchronous intermediaries for interservice coms
instead of direct, synchronous calls, applications exchange messages through the broker
see Event Driven Programming

Decoupling: Producers and consumers operate independently, enabling asynchronous communication and enhancing fault tolerance.
Increased Performance & Responsiveness: Asynchronous messaging prevents producers from blocking, optimizing performance even under heavy load.
Simplified Development & Deployment: Location transparency abstracts service discovery and routing complexities.
Load Levelling: Message queues act as buffers, absorbing traffic spikes and preventing cascading failures.
Reliable Delivery: Brokers offer various message persistence and acknowledgement options, ensuring message delivery even in failure scenarios.

Message Ordering: Guaranteed within a partition/queue, not across multiple. Impacts scalability and consistency.
Delivery: "At-least-once" ensures delivery (potential duplicates). "Exactly-once" guarantees single processing (more complex). Choose based on use case needs.
Durability: In-memory queues offer speed but risk data loss. Disk-based persistence prioritizes reliability over speed.
Message Flow: Push-based brokers deliver messages proactively. Pull-based brokers require consumer action.
Routing: Content, topic, or header-based routing mechanisms allow for granular control over message delivery and consumption.

High-throughput data streaming for event-driven architectures and real-time analytics.
Guaranteed order within partitions and robust fault tolerance.
Ideal for mission-critical applications.