SOA is Dead; Long Live Services

[Jane Somerfield]

Have anyone read this article. Any comments?

http://apsblog.burtongroup.com/2009/01/soa-is-dead-long-live-services.html

I vote for it!

[Ulf Dittmer]

It's been clear for a while that SOA has been way overhyped -particularly by the tools and server vendors- so the premise is not exactly new. But while the SOAP stack built on top of WS-* may have piqued, the interest in simpler WS techniques like REST is still building up. (On the Java side, JAX-RS and the Jersey library that implements it is a sweet little thing.) Any rearchitecting towards SOA that looks like it might turn into a big-money item is of course dead in the water now.

[Peer Reynders]

Lets not "throw out the baby with the bath water". While I'd like to see SOAP/WS-* de-emphasized, I believe that there is value in Service-Orientation and its Service-Orientation Design Princples. Even in this context it is clear that the majority of the services needed are entity services which can be realized as RESTful web services. It remains to be seen how much plumbing is actually required to realize "business agile" task services (or equivalent) for the majority of organizations (not just the ones with deep pockets). Service-Orientation will probably help to formalize and unify many EAI best practices and guidelines.

In a way this is reminiscent of "the mainstream" moving to Java Enterprise Application servers, getting disillusioned and then getting all excited about Spring/Hibernate and the like - in the end all of "them" are still around.

[Soalib Incorporated]

Jane Somerfield wrote:Have anyone read this article. Any comments?

http://apsblog.burtongroup.com/2009/01/soa-is-dead-long-live-services.html

I vote for it!

Those who voted for it is an indication that they have not been able to grasp the true value of SOA. You may be surprised to see what we were able to do using SOA. We at Soalib (http://soalib.com) have built services that can be consumed from embedded systems, mobile, legacy systems, PC, mainframes and supports all major programming languages (Java/J2ME, C/C++, C#/VB .NET), all major databases, and the list is growing. We were able to control a robotic arm using a cellphone, were able to do video conferencing using web services, it is all possible. It just takes time to build a SOA architected system. It took us 5 years. Are there any shortcuts to it, well people who have taken shortcut are now saying "SOA is dead". SOA is not a new concept, its been there all the time for the past 30 years, its just a new acronym of an old concept.

[William Brogden]

To: soalib incorporated
Subject: Naming policy

If you expect to get respect around the ranch you are going to have to respect our naming policy - it was clearly displayed when you signed up.

Bill

[Jane Somerfield]

Soalib Incorporated wrote:

Jane Somerfield wrote:Have anyone read this article. Any comments?

http://apsblog.burtongroup.com/2009/01/soa-is-dead-long-live-services.html

I vote for it!

Those who voted for it is an indication that they have not been able to grasp the true value of SOA. You may be surprised to see what we were able to do using SOA. We at Soalib (http://soalib.com) have built services that can be consumed from embedded systems, mobile, legacy systems, PC, mainframes and supports all major programming languages (Java/J2ME, C/C++, C#/VB .NET), all major databases, and the list is growing. We were able to control a robotic arm using a cellphone, were able to do video conferencing using web services, it is all possible. It just takes time to build a SOA architected system. It took us 5 years. Are there any shortcuts to it, well people who have taken shortcut are now saying "SOA is dead". SOA is not a new concept, its been there all the time for the past 30 years, its just a new acronym of an old concept.

Five years is too long. It is because you are using the "beautiful" SOA. Are they completed yet? Can I find your service through UDDI?

[Ulf Dittmer]

That's where the waters start to get muddy. UDDI is one of the parts of the WS-* protocol stack that truly is dead. While SOAP and WSDL are widely implemented and used (as are some of the other standards, like WS-Security), UDDI failed to take off, even though it's one of the oldest WS standards. But the market moved into a different direction, away from a public market that peddles individual services (or actually, it didn't move towards that), so the need for public registries went away. But I realize that's not the point you're trying to make...

[Jane Somerfield]

Soalib Incorporated wrote:
...SOA is not a new concept, its been there all the time for the past 30 years, its just a new acronym of an old concept.

Why do I keep seeing so many thick new books coming out on SOA? I do not get it. Here is one just released a few days ago:

http://www.amazon.com/Design-Patterns-Prentice-Service-Oriented-Computing/dp/0136135161/ref=pd_bbs_sr_1?ie=UTF8&s=books&qid=1232135479&sr=8-1

[Peer Reynders]

Jane Somerfield wrote:Why do I keep seeing so many thick new books coming out on SOA?

You did read the article that you linked to, right?

Anne Thomas Manes wrote:Although the word “SOA” is dead, the requirement for service-oriented architecture is stronger than ever.

Thomas Erl is is one of the few authors who makes some sense in all the (now fading) buzz.
The practice of Service-Orientation (with its design principles) is possibly as important to enterprise architecture as Object-Orientation is to application development.