My Azure Study Note - 15 - Messaging- Service Bus, Event Hubs

Service Bus

• Fully managed, full blown message queueing service

• Durable

• Supports point-to-point (Queue) and pub/sub (Topic) scenarios

• Compatible with AMQP protocol

• Compatible with JMS 2.0 API (Premium only)

 Advanced features:

• Message sessions (guarantees FIFO)

• Dead-letter queue

• Scheduled delivery

• Transactions

• Duplicate detection

• And more…

Availability:

• SLA: 99.9%

• Can be configured for geo-disaster recovery 

Event Hubs

• Big Data streaming platform and event ingestion service

• Note: No “messaging” in the description

• Basically a managed Kafka implementation

• Can receive millions of events per second

Event Producers

- Components generating the events

- Can be done by anyone with the client /

HTTP client

- Simple connection and API

Partition

- Single event stream

- Can think of it as a single queue

- Guarantees order

- Limited availability

- Better to spread messages across

partitions to improve availability

- …but then order is not guaranteed

- Max 32 partitions on a single Event Hubs

Consumer Group

- Logical group of receivers, belong to the

same application

- Example:

- Receivers for processing telemetry

= Consumer Group

- Receivers for storing the telemetry

= Consumer Group

- Event receiving is done via AMQP

protocol 

SLA:

• Basic and Standard tier: 99.95%

• Dedicated: 99.99% 

Event Hubs Throughput Units

• Throughput is measured in Throughput Units (TU)

• 1 TU =

• Ingress (Input) – 1MB / sec or 1000 events / sec

• Egress (Output) – 2MB / sec or 4096 events / sec

• Prepurchased, billed by the hour