Azure Service Bus
Download as PPTX, PDF•3 likes•3,396 views
This document provides an overview of Azure Service Bus and its capabilities for cloud-scale messaging and events. It describes the core components of Service Bus including queues for brokered messaging, topics for publish/subscribe messaging, Relay for enabling communication between private and public clouds, and Event Hubs for high-volume event ingestion. It also outlines common usage scenarios and provides a brief history of distributed systems and messaging technologies to provide context. The document concludes with a demonstration of Service Bus topics and Event Hubs.
Azure Service Bus
- 1. Azure Service Bus Cloud-scale Messaging and Events Azure-in-the-ATL December 2014
- 2. • Third Tuesday of the month, generally 6-8 PM • Microsoft offices in Alpharetta • Updates on www.meetup.com/Azure-in-the-ATL • Formats • 1+ hour, single speaker • 2 or 3 “lightning” talks (20-30 mins each) • Panel discussions • ? About our group
- 4. Agenda • Messaging and Events – What and Why? • Azure Service Bus overview • Queues • Topics • Relay • Event Hubs
- 5. A Brief History of Distributed Software Systems • In the beginning, it was all punchcards • And then we got… • Operating systems • Multiple CPUs • Multitasking • Threads • Server farms and data centers and clouds and stuff • Stupid cat pictures • browsers and web servers and stuff
- 6. Stuff Over Here Wants to Talk to Stuff Over There • Which means we got… • Ethernet • RPC (DCOM, CORBA, etc.) • Enterprise Messaging (TIBCO, etc.) • XML messaging • SOAP • REST • Most of these fall down at scale due to conceptual and/or practical limitations • Messaging and Events provide a useful set of core concepts to reason about moving data around at scale (“in the cloud”)
- 7. Messaging and Events • Messaging – transmission of data from sender to receiver via some intermediary • Implications • Sender and receiver don’t necessarily know about one another • Intermediary can be arbitrarily complex and interesting (multiplexing, durability, security, etc. etc.) • Data goes only one way • Events – data transmitted from sender to receiver as a side effect of some interesting occurrence
- 8. Azure Service Bus – Big Picture • Umbrella technology that implements (most) messaging for Azure • Scales to potentially millions of concurrent message senders (Event Hubs) • Supports various standard messaging patterns • Producer/consumer (competing or cooperative) • One-way async • RPC-ish • Reactive-style streamed event consumption • Scenarios • IoT • Consumer messaging • MSMQ or TIBCO type messaging • Hybrid cloud
- 9. Azure Service Bus – Queues • FIFO brokered messaging technology • Similar to MSMQ in usage (transactions, security, sequencing, etc.) • Scenarios • Rate leveling • Independent scaling of application tiers • Fault tolerance and enhanced reliability • Asynchronous processing • Differences with Azure Queues • http://msdn.microsoft.com/en-us/library/azure/hh767287.aspx
- 10. Azure Service Bus – Topics • Brokered pub/sub messaging • One-to-many fanout scenarios • Topics are where messages are sent • One or more subscriptions are created against a specific topic • Can be all messages, or filtered • Scenarios • Domain Event pattern • Scenarios where multiple consumers should process the same message
- 11. Azure Service Bus – Relay • Niche technology that enables secure communication between private and public clouds • WCF-based • Uses custom WCF bindings and behaviors • Integrates with Azure AD Access Control • Can be very useful for securely exposing legacy WCF endpoints to cloud-hosted code
- 12. Azure Service Bus – Event Hubs • Implements very high-scale message ingest and consumption • Supports AMQP and HTTP APIs • Also various language bindings (.NET, Java, etc.) • Event delivery to readers is async via Consumer Groups (logical stream copy, similar to topic subscriptions) • Uses partitions (within a CG) to logically segment data for fast consumption • Only 1 reader per partition • # partitions == # of readers in a group • Readers can use per-partition checkpoints to mark progress and provide resiliency • EventProcessorHost class simplifies administration of reader tasks
- 13. Demo – Service Bus Topics and Event Hubs
- 14. Not TicketMaster Service Bus Event Hub InventoryManagement Auditing Analytics Service Bus Topic