Book Image

SOA Patterns with BizTalk Server 2009

By : Richard Seroter
Book Image

SOA Patterns with BizTalk Server 2009

By: Richard Seroter

Overview of this book

SOA is about architecture, not products and SOA enables you to create better business processes faster than ever. While BizTalk Server 2009 is a powerful tool, by itself it cannot deliver long-lasting, agile solutions unless we actively apply tried and tested service-oriented principles. The current BizTalk Server books are all for the 2006 version and none of them specifically looks at how to map service-oriented principles and patterns to the BizTalk product. That's where this book fits in. In this book, we specifically investigate how to design and build service-oriented solutions using BizTalk Server 2009 as the host platform. This book extends your existing BizTalk knowledge to apply service-oriented thinking to classic BizTalk scenarios. We look at how to build the most reusable, flexible, and loosely-coupled solutions possible in the BizTalk environment. Along the way, we dive deeply into BizTalk Server's integration with Windows Communication Foundation, and see how to take advantage of the latest updates to the Microsoft platform. Chock full of dozens of demonstrations, this book walks through design considerations, development options, and strategies for maintaining production solutions.
Table of Contents (18 chapters)
SOA Patterns with BizTalk Server 2009
Credits
About the author
About the reviewers
Preface
Index

Using queues within asynchronous scenarios


Queue-based technology is an underrepresented but powerful way to exchange data and events between disconnected clients. WCF has full support for Microsoft's queuing implementation (MSMQ) and BizTalk has an adapter specifically targeted at the netMsmqBinding WCF binding.

Why introduce yet another layer in your service communication? BizTalk has queuing logic, so what benefit do we get by having our service client send a message to an external queue that BizTalk acts upon? First of all, you get delivery assurance in the case of the service being offline. As you are not travelling over an inherently unreliable transport like HTTP, you can be confident that your message will arrive only once at its destination because of the intermediary queue. Also, a queue enables you to implement a level of soft throttling by allowing the queue to get pummeled by inbound requests but allow the service to process them at its leisure.

What we will demonstrate here...