Event Streaming differs from traditional publisher-subscriber mechanism. Both these technologies fall under message oriented middleware.
| Criteria | Pub-Sub | Event Streaming |
|---|---|---|
| Implementation | Using message brokers like ApacheMQ, | Using event streaming platform like kafka |
| Message Storage | Deleted after messages consumed | Stored regardless of consumption |
| New Subscribers | New subscribers cannot access old messages | Newly added subscribers on topic can access old data |
Event driven architecture is useful in a herogenous environment, when the event can be used to perform different activites across services.
For example, if a customer places an order on an online store, besides having a subscriber that fulfills the order, we can have a search service that subscribes to index the event into the search database. Both these services, order-fulfillment and search-service can be built upon RDBMS and elastic respectively.
