Is Open Source the way ahead for SOA?

Posted by Mark Little on Jun 01, 2007

Sections

Operations & Infrastructure,

Enterprise Architecture,

Development,

Architecture & Design

Topics

Open Source ,

SOA ,

SOA Platforms ,

SOA Appliance

Tags

SOA Adoption ,

Adoption

Dana Gardner suggests that there is growing evidence of a synergy between open source and SOA. He references a recent report by Forrester which Dana says adds to:

"... the compelling economics that bind open source products and SOA activities."

The report mentions that of 400 executives surveyed, 71% of them view open source as either "important" or "very important" for consolidating their IT infrastructure. Plus, 57% said that they saw open source as "important" or "very important" for facilitating a migration to SOA. According to Forrester, 78% believed that the open standards approach typically taken by open source software is a major factor driving their view of its value for SOA next-generation enterprise architectures.

Dana believes that a series of recent announcements in the community show that 2007 will see open source drive SOA adoption even further.

Most importantly is the expected delivery in Q4 of this year of a comprehensive open source SOA platform stack from Red Hat and its JBoss unit. The goal is to take the best of open source platform, middleware, tools, and SOA infrastructure (such as ESBs) and package them for volume viral delivery to the market, with a simplified pricing model (based on a “server instance”) for support.

Furthermore, recent interviews with IONA Technologies around CXF (the Apache merger of XFire and the Celtix ESB), the CEO of SourceLabs (who are attempting to “professionalize” the adoption of open source components, making it “safe and easy,” as well as “dependable”), and a new SOA project in Eclipse from Deutsche Post World Net, are further evidence that open source is not lagging behind the traditional proprietary vendor approaches, as has often been the case in the past. As the article mentions:

SourceLabs expects open source for SOA to be huge, and it is gearing up to manage the movement to and support of open source code products for SOA ...

However, it's not all a bed of roses. Celtix wasn't really going anywhere when it was within ObjectWeb. The Eclipse STP project (an attempt to provide a toolset for SOA) is still struggling to gain acceptance over the WTP project. As with all successful open source, the community adoption/following is often as important as (and sometimes more important than) the technology. Whether 2007 is really going to see open source leading the way with SOA adoption remains to be seen. However, it does seem clear from the amount of open source efforts in and around SOA, that developers and adopters are taking open source solutions seriously.

1. Back to top

   Is SOA the way ahead for Open Source?

   by Chris Norton

   If by SOA you mean WS-*, I'm not sold. Open source projects tend to emerge from some fairly concrete need, and grow to the extent that they address that need in a useful way. The WS-* leviathan is so overwrought with interlocking dependencies that it is hard to imagine how an open source team would ever muster the manpower to get started, let alone produce anything useful.
   
   I just don't see it. The buzz that I hear in the open source world is about Ruby, JavaScript, AJAX, RIA, REST, LAMP, and other very concrete things that get out of your way and let you do interesting things with networked computers. The buzz is there because those things work well independently, and they work well together.
   
   That's not to say that IBM, BEA, and other big vendors won't release big chunks of useful code as open source, it's just that I never, ever hear any of my peers saying "Dude, did you hear about that new WS-ResourceLifetime spec? So killer. I can't wait to use it."

   Reply

2. Back to top

   Re: Is SOA the way ahead for Open Source?

   by Jim Murphy

   I agree with your underlying premise but think of it a little differently.
   
   Most of WS-* is useless and will never be implemented. Some of it is essential to replace proprietary, vendor specific solutions. The incumbent mega-vendors have incentives to keep you locked into their type system, platform, world view, and way of approaching things. With the world interested in standard commoditized plumbing they have retrenched by confusing the situation with WS-*. Being able to do a low risk swap Windows machines for Linux boxes at any point (because of standard and loosely coupled services) does not get MS very excited and same goes for the other vendors - they're all in the same camp no one is worse then the other really.
   
   Open source doesn't have the same legacy to protect or lock-in to create. So while the big vendors continue battle each other in spec committees solution driven open source projects can be quietly delivering on the promise of SOA by providing the tools and infrastructure needed to realize it.
   
   All that said - my company ain't goin open source anytime soon but I appreciate how a neutral point of view fits best with SOA.
   
   Jim

   Reply

3. Back to top

   Re: Is SOA the way ahead for Open Source?

   by Andrew Clifford

   I can't see RedHat driving the acceptance of SOA with a "comprehensive open source SOA platform stack." The last thing I need is a tightly-coupled package. I like all the non-functional features provided by an ESB (BPEL included). Can I get this best of breed in a decoupled open source offering? Probably not there yet. Opensource is a more powerful driver than SOA. SAO has vendor-lock all over it right now with proprietary offering. As more opensource projects implement WS-* and ESB defacto specs the more adoption will continue.

   Reply

4. Back to top

   Re: Is SOA the way ahead for Open Source?

   by Burr Sutter

   "Ruby, JavaScript, AJAX, RIA, REST, LAMP " - these solutions are primarily targeted at the web application space which is the largest sector for application developers to focus on today. Much like client/server not that long ago when everyone needed to understand event-driven GUI and windows development.
   SOA is here to address application to application interoperability and integration. This is a fundamentally different problem space than webapp dev. The integration space is a relatively smaller market in terms of technologists than general web app designers/programmers. If your organization is streamlined enough to only need a few web applications then SOA may not be for you.
   
   SOA does not equate to WS-*, however, SOAP & WSDL are two very useful solutions when it comes to MSFT/.NET to Unix/Linux/Java interop often needed in application to application requirements.

   Reply

5. Back to top

   Too many SOA products to begin hypothesizing any winners

   by Frank Cohen

   The report doesn't include WSO2's new ESB, it doesn't talk about TIBCO ActiveMatrix, it doesn't consider what may happen if JBI 2.0 succeeds, and there are a hundred or more open source projects for SOA development not even considered. It seems to me that it is way to early to be talking about emerging winners. It's still going to be another year or two before management understands SOA, let alone which tools will get them there.
   
   -Frank Cohen
   www.pushtotest.com

   Reply

• Older >