
Understanding
SOA with Web Services, $39.99
by Eric Newcomer and Greg Lomow
Addison-Wesley, 2005
444 pages
Examples in XML
http://www.amazon.com/exec/obidos/ASIN/0321180860/larkware-20
Eric Newcomer's UNDERSTANDING WEB SERVICES was the best of the first crop of Web services books - a no-nonsense guide to the confusing thicket of XML standards that helped many people (including myself) get a handle on what was going on as we all tried to figure out what this XML-over-HTTP stuff was all about. That's one reason I was happy to see this new title cross my desk; the recent explosion of new standards and new ways to hook things together has left things more confused than ever. The book doesn't disappoint. Newcomer and co-author Greg Lomow do an excellent job of explaining why service-oriented architectures are important, show how Web services enable building SOAs (while pointing out that plenty of organizations managed to put together SOAs on older technologies long before Web services entered the pictures), and help navigate through all those confusing standards.
The authors start by digging into some of the key concepts that any SOA needs to fulfill: service contracts, the differences between line of business services and technical services, the benefits of reuse and incremental adoption, the characteristics of a service that distinguish it from a simple component, and so on. This acts as positioning material that then allows them to discuss why they thing Web services fit so well into SOAs, and why they're being increasingly used in this fashion. Here the authors start drilling down into some particulars, showing how Web services can fit on top of a variety of transports (WebSphere MQ, CORPA IIOP, Tibco Rendezvous, and so on) and how they can be used to enhance integration and interoperability. They give examples from both the legacy world (CICS and IMS) and the modern world (bridging J2EE and .NET). It's clear that they've spent plenty of time in the trenches with businesses using this stuff on top of many platforms, and aren't wedded to any particular vendor.
Most of the rest of the book is an extended tour of the alphabet soup that can be so difficult for the beginner to penetrate. They dig into the competing standards in business process management and choreography, the whole WS-* stack of protocols, the various metadata specifications, the knotty issues of security and reliability, and transactions, to name a few subjects. Where a protocol is reasonably finished, they'll tell you; where the standardization effort is incomplete, or patents threaten to encumber things, you get suitable warnings. The end result is that you can use this book as a roadmap to tell you which parts of a SOA are likely to be easy to implement, and which parts are likely to prove to be troublesome, as of early 2005. Of course, the field is moving fast; parts of this book will undoubtedly be outdated within a year. But that's the problem with dead trees in general. If you need to get up and running with this subject, this looks like a good place to start as of right now.