In-Depth
The Emergence of Integration-as-a-Service
Thanks to aggressive service enablement and the efforts of DI vendors, another integration paradigm -- call it Integration-as-a-Service -- has emerged.
Data integration (DI) doesn't seem like the most obvious candidate for Software-as-a-Service (SaaS), being mostly about connectivity between back-end systems such as databases, operational applications, flat files, or other sources, after all.
Thanks both to aggressive service enablement and to the efforts of DI vendors, which have aggressively SOA-ified their integration platforms, another paradigm -- call it Integration-as-a-Service (IaaS) -- has emerged.
IaaS prescribes the use of Web services, well-defined interfaces, and ubiquitous calls between and among service-enabled applications and data sources to deliver a loosely-coupled integration experience.
In other words, write analysts Noel Yuhanna and Mike Gilpin in a recent Forrester Wave research bulletin, "IaaS offers a virtualized data services layer."
One of its most conspicuous drivers is a marked preference for both agility and automation -- to say nothing of rapid ROI -- among adopters, Yuhanna and Gilpin argue, citing feedback from IaaS users. That IaaS offers a logical -- if still imperfect -- onramp to cloud data is an added bonus, they say.
"Every organization undertaking an IaaS initiative wants quick return on investment … and shorter time-to-deployment. Most want agility and automation in metadata and data discovery, integration, data quality and transformation, and availability," they write. "Although cloud data integration with IaaS is still in its early stages and is evolving rapidly, organizations are looking for solutions
that can integrate on-premise data with cloud data. Most IaaS vendors already offer support for REST that enables such integration, with more sophisticated and integrated features in development for the near future."
Like most organizations that are striving to be agile, shops that adopt IaaS typically want more flexibility to support a wider range of use cases. More to the point, Yuhanna and Gilpin argue, they're shifting from a reactive stance (in which integration needs are determined by specific application requirements) to a proactive posture -- in which "Organizations want to cut implementation times with easier approaches that support the full range of use cases the business is looking to implement."
IaaS may better (or more flexibly) handle the requirements of semi-structured or unstructured data integration, Yuhanna and Gilpin point out.
"Although in the past the majority of IaaS has focused primarily on structured data, Forrester estimates that today a third of IaaS deployments are looking to broaden their scope to include all types of data such as geospatial, images, video, audio, binary, XML, and other content," they write. "These requirements reflect the emergence of an increasing number of newer applications that require this integration, which is easier to do once in the access layer rather than multiple times across many applications."
The biggest driver for IaaS could best be described as the Field of Dreams scenario. For years, shops were told that if they aggressively service-enabled their existing assets, good things -- e.g., flexibility, resilience, synergy, and ROI -- would happen. Having embarked on strategic SOA efforts, they understandably want to start realizing the ROI. More precisely, the allure of service enablement was the promise of leveraging existing resources -- by both exposing and in many cases repurposing applications and data sources -- to avoid costly rip-and-replace scenarios. IaaS is consonant with that promise.
"Most organizations expressed their unwillingness to build a data services framework from scratch, preferring to use existing components and processes to get a jump-start on such projects," Yuhanna and Gilpin write.
Forrester names a "gang of four" that it says leads the IaaS pack. Not surprisingly, all four of Forrester's picks -- Composite Software Inc., Denodo Technologies, IBM Corp., and Informatica Corp. -- offer established and respected DI offerings.
Forrester also listed both Microsoft Corp. and Red Hat as IaaS comers. Although Microsoft can make a creditable claim to a DI future -- SQL Server has included a built-in DI facility for a decade now, and BizTalk Server is a mature offering -- it's also limited (in IaaS as in other segments) by its Microsoft-only focus. Red Hat, for its part, isn't typically thought of as a DI player, but Forrester's analysts stress that the company has the goods thanks to its 2007 acquisition of the former MetaMatrix.