Event Hubs

Event Hubs are primarily designed to ingest high volumes of event and telemetry data in a high throughput manner from a variety of devices and services. They are similar in principle to Azure Queues and Topics, but with different characteristics. While Azure Queues and Topics are used for enterprise messaging scenarios in which transaction support, dead-lettering, ordered delivery, and guaranteed delivery are of prime concern, Event Hubs are biased towards very high throughput and event processing scenarios.

Partitions

The key technology that provides this high throughput is streaming, using a partitioned consumer pattern. This is where each consumer only reads a specific subset or partition of the message stream. With Azure Queues and Topics that use a competing consumer model, if too many consumers target the same queue, this will result in resource contention and scalability issues.

Messages sent to Event Hubs can be forced to target specific partitions by using partition keys...