SOA Advisor
SOA Past and SOA Future
Service Oriented Architecture's definition broadened and matured in 2006 to incorporate governance as a central tenet.
- By Dana Gardner
- 01/01/2007
While SOA made quite a splash throughout 2006, the year can best be characterized as a period when the SOA hook was baited. Mergers, product rollouts, standards maturity and a mish-mash of marketing approaches in 2006 have set up 2007 as the year for reeling in some big SOA catches.
Software infrastructure vendors and business applications providers -- as well as the global systems integrators -- are poised to extend SOA from the conceptual and project stages to the type of systemic and broad enterprise-wide initiatives that will best demonstrate SOA's long-term value and recurring productivity benefits.
Such moves into common, horizontal use of SOA required major vendors -- and open source projects -- to assemble and position themselves as increasingly complete solutions. At the same time, SOA components providers raced to demonstrate their wares and services as best-of-breed.
Becoming Mainstream
SOA became a mainstream concept in 2006, evidenced in part by a slew of mergers, acquisitions and partnerships. There was the HP acquisition of Mercury/Systinet, Red Hat's purchase of JBoss and IBM's purchase of Filenet, to name a few. Large vendors, including Microsoft, more deeply embraced SOA. Startups realigned themselves around the SOA message and value. Vendors became more service-oriented in terms of how they craft their products and deliver solutions to their customers.
SOA's definition also broadened and matured in 2006 to incorporate governance as a central tenet. As enterprise IT assets are exposed as services beyond the pilot stage, there comes a need to control, manage, and document which assets are being exposed as services, under what circumstances, to whom and for how long. Governance has emerged to bring trust and reduced risk to the use of SOA on an enterprise scale. Governance also drove several mergers and acquisitions.
The enterprise service bus (ESB) concept matured in 2006 as well. Several major vendors have improved or delivered ESBs, including IBM, BEA, JBoss, Cape Clear and open source projects such as Mule, ServiceMix, and CXF. ESBs went from an architectural pattern into real products for most SOA vendors. Many architects also speculated that Microsoft's Windows Communications Foundation forms an ESB equivalent for Vista and Windows legacy installations -- but it's unclear whether the technology will evolve into a larger, cross-framework ESB.
Big Fish in a Small Pond
Analysts expect 2007 will see a continuation of SOA-related acquisitions. Names that come up as acquirers include SAP, Oracle and Microsoft.
I expect the business applications vendors, including Microsoft, will begin to flex their muscles on SOA. If you provide both the infrastructure and the applications, then you can more readily produce SOA benefits on your terms.
Also expect to hear more about Event-Driven Architecture (EDA). EDA brings a real-time and responsive nature to SOA that ties into business intelligence (BI), and allows applications services to exploit BI solutions. It also opens the door to policy-based governance.
The new year will play host to market advances in the use of modeling. Just as EDA can exploit policy and BI for dynamic provisioning, modeling advances will give those closest to the line of business a greater role in defining and managing the business processes that SOA enables. Expect more products, specifications and methodology around the modeling of services that can leverage activity from 2006 around governance, ESB and BI. Included in this will be much-improved tools, graphical interfaces and innovative approaches to services choreography.
In 2007 we should also expect to see uptake in so-called Enterprise 2.0 capabilities, which borrow from Web 2.0 social networking and open collaboration approaches such as blogs, wikis, podcasts and virtual environments. We'll see these things mashed up with traditional enterprise messaging, Office applications and groupware products. Microsoft Live Office and Live Clipboard offer a glimpse at Redmond's intentions here.
Make no mistake, Web and Enterprise 2.0 uptake may encourage even broader use of SOA. We're already seeing rich Internet applications (RIAs) grow popular as a way to assemble SOA services as composite applications, while delivering a user experience more in line with rich client-side architectures.
Therefore, 2007 will mark the beginnings of large-scale SOA build-outs. Organizations that have not yet done SOA groundwork are going to remain in small-pilot implementations. But expect 30 percent of those organizations that have passed the pilot stage to move on SOA in a big way.
The proof of SOA's long-term viability will come when organizations that heavily embrace SOA show significant competitive advantage in their markets. Also look for more productive IT management, swifter return on new business services, and lower ongoing IT operating costs. All this will help SOA adopters free up more discretionary money for new growth-oriented, IT-enabled business ventures.
About the Author
Dana Gardner is president and principal analyst at Interarbor Solutions, an enterprise IT analysis, market research and consulting firm. Gardner tracks and analyzes Web services, application-development tools and application optimization techniques. He is also the producer of the podcast series, BriefingsDirect.