Introduction to Pub-Sub

A pub-sub design must say which subscriptions receive a publication and how long it remains available to each. Decoupling producers from subscriber addresses is useful, but retention, lag, and acknowledgement still connect their behavior.

Use cases of pub-sub

Common use cases include:

• Update delivery: A push delivery path can reduce polling, while a pull subscriber can control its own fetch rate. Pub-sub can use either mode.

• Log ingestion: Multiple consumers can process a log stream for analytics or archival. Facebook’s Scribe is a historical log-aggregation example; its archived project is not a current managed pub-sub service.

• Monitoring: One event stream can feed several analyses, each with its own progress and processing delay.

• State propagation: Publications can carry changes to replicas, caches, or client views. Ordering, duplicate handling, and missed-update recovery are still required for those views to converge.