Why? Isn’t there enough hype around SOA already?
I disagree it is all hype. Instead, there seems to be a low barrier to entry for discussing Service Oriented Architecture, and this is a good thing. Discussions on SOA concepts have people thinking about abstractions that will help to provide their applications with longevity beyond a 3-year lifecycle. Programmers and architects are considering the scalability implications of integration strategies. We finally have something related to architecture besides yet another DataGrid article. Now we need to clarify the message and provide examples. We need to listen to the industry responses and clarify when the message isn’t succinct.
For example, more and more people (subconsciously or otherwise) replace the term “Service Oriented“ with “Web Service Implemented.“ While web services and GXA (whose name is now unceremoniously replaced simply with “Web services specifications“) certainly form the current and most obvious implementations, the architectural concepts are not inextricably bound to SOAP and XML.
… you can use CORBA or COM IDL to do “pure” protocol-based integration a la service-orientation provided one has sufficient discipline. Darwin has a way of sorting out technologies that rely on programmer discipline that none of us can really change.
While everyone busily quotes Don Box, Harry Pierson, Rich Turner, and Pat Helland, concepts are less clear because the sound bytes are terse and abstract. Concepts are contorted to look like products, and various vendor products compete to show how simple XML and Web Services really are. We still have more questions to answer. Why not rely on a binary format over the wire now and create wrappers that provide class interfaces to mask the underlying transmission type? Why think about XML and the implied bloat when I can drastically reduce the transmission size? What does ASMX really buy me that COM and CORBA didn’t?