Publishersubscriber

Publishersubscriber, commonly written as publish-subscribe or pub/sub, denotes a messaging pattern in which information producers (publishers) emit messages that are categorized by topics or channels, and consumers (subscribers) receive messages based on their interest. Publishers and subscribers are decoupled: publishers do not need to know who subscribes, and subscribers do not need to know who publishes. This decoupling enables scalable, asynchronous distribution of events across distributed systems.

Most implementations rely on a mediator, often a message broker, that routes messages from publishers to subscribers. Publishers publish to a topic or channel; subscribers subscribe to topics of interest. The broker may offer filtering, durable storage, and delivery guarantees. Some brokerless systems route messages directly between peers. Topics, queues, and channels are common abstractions, with topics supporting many subscribers and queues typically supporting load-balanced delivery to a set of consumers.

Key characteristics include loose coupling, scalability through horizontal expansion, and asynchronous delivery. Delivery semantics vary: at-most-once, at-least-once, and exactly-once are supported in different systems; ordering is often limited to topics or partitions. Durability options allow messages to persist for late subscribers. Pub/sub is widely used in microservices, event-driven architectures, real-time dashboards, and notification systems.

Common examples include MQTT for IoT, Apache Kafka for event streaming, Google Cloud Pub/Sub, and Redis Streams. The pattern contrasts with point-to-point messaging, where a message is delivered to a single consumer.