SOA is obviously the new buzzword of the day. Among the many acronyms, one that is seen very often is "Same Old Architecture." In many ways, this is true. The key differentiator between the paradigms that have been prevalent in the past and this new incarnation of "service-orientation" is that the new definition of services is targeting the business as well as the technical side of the house. Same old architecture - different politics.

Mind you, I am not saying that this not needed. The processes and governance that has been formalized around SOA make for a very effective IT renovation roadmap. If you are interested in how SOA assists in IT renovation, check out the book Enterprise SOA: Service-Oriented Architecture Best Practices by Dirk Krafzig, Karl Banke, and Dirk Slama (The Coad Series). I recently published a review on this at my blog: http://ajitsagar.javadevelopersjournal.com/read/1062164.htm.

There is a tendency (propagated by vendors) for the industry to conclude that SOA means Web services. While Web services provide the ideal platform for implementing SOA, they are not the only option. Everyone automatically assumes Web services whenever the term SOA is mentioned. Well, adding Web services to SOA definitely gives SOA a bit of an oomph, and differentiates it from being the "Same Old Architecture." After all, if you weren't using Web services to implement SOA, what's new about your solution?

While it is true that Web services offer a very attractive platform for realizing SOA, they are not the only technology available to do so. In fact, the main message behind SOA is not the "Web" but rather the "service." The main objective of SOA is to help organizations move toward a Service Oriented Enterprise (SOE). The main problem in organizations that SOA addresses is the ability to use architecture as a common tool for IT and business to achieve a common objective - IT agility.

I've found it a little surprising that while the term WSOA (Web Service-Oriented Architecture) is around the corner for everyone on the SOA bandwagon, no one really talks about MSOA (Message Service-Oriented Architecture). Although I'm sure I haven't invented this term, I haven't found it in my Google searches. You will find "Message Oriented Architecture" or "Messaging Service," but not MSOA. Messaging is another way to leverage existing investments to realize SOA. It provides the protocol and the semantics. What it doesn't provide right off the bat is a standard service registry akin to UDDI. So shouldn't WSOA be a subset of MSOA? After all Web services provide a mechanism to exchange messages in a loosely coupled architecture and eliminate the tight coupling mandated by APIs.

As I was saying earlier, SOA is not the new concept - what's new is its application. If you take away the whole message of "platform and language independence," which never really happens when you actually implement something, is the concept any different from what Java proposed with Jini? In fact, I would say that Jini is one of the first software architectures that promoted the concept of SOA - although they made the mistake of adding the qualifier "network" to it. So it got associated with devices, not software applications. As the "computer" moved away from being the "network," Sun's message faded into burst bubbles of .com, and Jini slipped through the cracks. When you look at Sun's site, you wonder "Whither Jini?"

In the article "Jini Network Technology Fulfilling Its Promise" (http://java.sun.com/developer/technicalArticles/Interviews/waldo_qa.html) Jim Waldo has provided some very interesting insights into where Jini is today ("today" is relative since the article is about a year old). I found two statements in the article that talk about Jini from the SOA perspective:

• Probably the biggest misconception is that it is concerned primarily with devices.
• The message about devices hid the fact that Jini software is really a general service-oriented architecture.

It is all about the message. Sun has never really made any money on software. If Jini was marketed properly, it could have been the leading concept behind the new wave that is SOA. I remember back in 2000, in one of my previous lives, I had worked on a whitepaper where we borrowed the concepts from Jini and applied them to transactional marketplaces - the setup would be similar to what is touted in SOA today. Perhaps if Sun had used the word "messaging" instead of "network" and "service" instead of "device," Jini would have evolved into the key architecture implementation for SOA today.

On a side note, it's interesting to see how SOA has been globally accepted. I recently received an invitation to a Web services and SOA conference in China (details are available at www.ajitsagar.javadevelopersjournal.com/read/1088114.htm), where SOA seems to be gaining a lot of traction. This should be an interesting experience, and I hope to chronicle it in one of my future editorials.