BT

Burton Says SOA Is Rising From Ashes

by Boris Lublinsky on Nov 09, 2010 |

Nearly two years after proclaiming that SOA was dead, the Burton group has changed their mind and now writes that SOA is set for a comeback.

Quoting Chris Howard from Burton, Maxwell Cooter writes :

... there's a different climate now. There are several reasons for [SOA] comeback... there is a still a need for SOA but it had been sidelined as a technical issue and that IT had failed to sell SOA as a transforming methodology... SOA projects failed because there was too much concentration on the technology and because, in the financial climate of the past few years, major transformational projects were canned

According to Cooter:

... [the] companies would need to stop talking about SOA as a technology but in terms of a business case. Technologists must recast SOA as an architectural approach rather than a technology-first solution... Service oriented environments require a higher level of abstraction in design, which must be decoupled from the underlying technology

Picking on Burton’s report, David Linthicum notes that the initial SOA failure was, for the most part, due to the overselling the SOA technology (ESBs as "SOA-in-a-Box,") not to the approach itself. In his opinion, SOA is a foundation of many things including good enterprise architecture, business/IT alignment, effective usage of cloud computing and more.

According to Linthicum:

The core purpose of SOA is to define a way of doing something that provides an end-state architecture that's much more changeable and thus much more agile, and ultimately provides more value to the business.

For a few years now InfoQ has been promoting SOA as an architectural style based on a functional decomposition of enterprise business architecture and introduces two high-level abstractions: enterprise business services and business processes. Enterprise business services represent existing IT capabilities (aligned with the business functions of the enterprise). Business processes, which orchestrate business services, define the overall functioning of the business. This architectural style is a foundation of many current and future implementations regardless of the technologies that are used for the actual implementation. It is nice that analysts are sharing our opinion again.

Hello stranger!

You need to Register an InfoQ account or to post comments. But there's so much more behind being registered.

Get the most out of the InfoQ experience.

Tell us what you think

Allowed html: a,b,br,blockquote,i,li,pre,u,ul,p

Email me replies to any of my messages in this thread

Selling approaches doesn't make money by Hermann Schmidt

This sounds like the realization that current "SOA technologies" don't deliver and are too expensive. Surprise! Vendors sell SOA with their products, not with a mindset.

To me, this even seems like a retreat to more abstract grounds because the technology approach didn't quite cut it. In abstract land, life is so much easier. Yes, there is work to be done here. However, at the end of the day, some technology must carry it all. Plenty of opportunities to fail. Someone has to take care of it.

...that architecture strategies that involve cloud computing must have a service orientation foundation.


No. They need not. They should.
Cloud computing in general doesn't force anyone to think more service-oriented. You can build monolithic apps just fine. Cloud is not a SOA migration path. Conservative minds will still find ways to avoid change, and guess what, stuff may work for them.

Re: Selling approaches doesn't make money by Felipe Oliveira

No make sense you have a monolithic application in current time of diverse platforms and languages. Everything has to be connected and SOA provide that your application to use others APIs like Twitter, Facebook, SalesForce and distribute your business model too.

From a technical point, it makes more sense you distribute only parts that have very intensive use. This allows for horizontal scalability and reduced maintenance costs.

Best Regards,

Felipe Oliveira

Big bang dead, bottoms-up alive and well by greg schott

Prescience wasn’t required to predict that top-down SOA projects, driven by large vendors hawking SOAware, were doomed to fail. Now that the hype has subsided, projects are being scoped correctly. They are often not tagged as “SOA” initiatives, but service orientation is presumed. They begin as "bottoms-up" or incremental approaches and, after demonstrating success, are being rolled out broadly. In fact, the number of web visitors evaluating open source Mule has tripled in the past year as organizations continue to drive toward service orientation . Greg Schott, MuleSoft

Soa tools must be technologically open. by Cenk Ozan Kahraman

I am using the "Java Killer"'s soa suite. There is only one technology you can use, and it booms. If I could choose my own technology for creating Soa applications, that would be good. Message line is: Don't use soa suite.

Re: Big bang dead, bottoms-up alive and well by Steve Snodgrass

Very well said!

Re: Selling approaches doesn't make money by Hermann Schmidt

That may be right for enterprises, who are selling stuff over the web. For others, where software is mainly for internal use, there is no immediate requirement to open up by exposing a business model. There, cost of development and operation is what counts and makes the cloud (in whatever flavor) interesting.

The point I was trying to bring across: if the costs of monolithic apps in the cloud is lower than in classic infrastructure, they will be built. Enterprises may consider a new architecture, but they don't have to.

I don't see any intrinsic force in cloud computing, which inevitably drives enterprises to go service-oriented.

Re: Big bang dead, bottoms-up alive and well by Boris Lublinsky

Bottom up is even more dangerous.
Creation of services that make sense only in the realm of existing project leads nowhere. This will become SOA for applications, with all drawbacks of application centric implementations

Re: Selling approaches doesn't make money by Boris Lublinsky

I see yor point, but...
No (very few) application is an island. This means that just lift and redeploy would not work, still have to connect to others. The other interesting observation is coming from Netfix talk at QCon SF. They kinda looked at this and decided that in order to move to Amazon they do have to rearchitect

Re: Big bang dead, bottoms-up alive and well by Steve Snodgrass

You are making the incorrect assumption that bottom up = services that only make sense in the realm of the existing project. Even with a bottom up approach, services should strive to be as reusable and system agnostic as possible. We can't predict the future however so incremental changes may need to be made as the Service Oriented Architecture matures and services are used in ways not originally envisioned. IMO, a bottom up approach is the only approach which has any likelihood of success when dealing with a large corporate legacy IT infrastructure. My company is currently helping a Fortune 500 corporation to adopt better SOA practices. I've recently begun blogging on our company website about SOA adoption in real world corporate situations and I will discussing the challenges we have faced and will face as we move forward with this particular project. It is early yet but we are already beginning to see efficiencies emerging as the result of our adherence to solid SOA methodologies. The latest blog post can be found here if you'd like to read it: knbintegration.com/blog/?p=38

Re: Selling approaches doesn't make money by Hermann Schmidt

Don't get me wrong, enterprise apps (EAs) traditionally have many interfaces. That doesn't make them service-oriented. Old school EAs have tightly tailored interfaces for specific purposes. The technology would have allowed service orientation already. Take queuing systems for instance. A perfect transport for services. Not for the web, but for the corporate intranet.

Cloud platforms may restrict data exchange to other transport protocols/techniques. However, that doesn't profoundly affect the design and content of an interface (its semantics).
I can happily design an HTTP binary interface without taking care of backward or forward compatibility, and with no reusable content.

That is the core of my critique. Service orientation doesn't come through new platforms. It's not about how things are hosted, exposed, and transported. It is about the design of software.

As for HTTP for instance, this would be the leap from using it as just another transport protocol to a true resource-oriented architecture.

Personally, I do hope that cloud computing is taken as a starting point to rethink stuff and drop ballast.

Allowed html: a,b,br,blockquote,i,li,pre,u,ul,p

Email me replies to any of my messages in this thread

Allowed html: a,b,br,blockquote,i,li,pre,u,ul,p

Email me replies to any of my messages in this thread

10 Discuss

Educational Content

General Feedback
Bugs
Advertising
Editorial
InfoQ.com and all content copyright © 2006-2013 C4Media Inc. InfoQ.com hosted at Contegix, the best ISP we've ever worked with.
Privacy policy
BT