Research carried out by the Aberdeen Group shows that 61% of IT strategists rank “integrating systems and processes” as among their highest priorities — this task received a higher percentage than any other strategic descision. In these rapidly times, with cost-effectiveness a key consideration in any IT decision, enterprise integration cuts the costs of a business process while improving customer satisfaction by shortening time to delivery. Like business process reengineering 10 years ago, enterprise integration does this task primarily by automating the transitions between steps in a business process.

However, actually trying to integrate a business in a systematic and strategic way often results in an almost comically complex point solution that makes further improvement even more difficult. Cost-effective enterprise integration necessarily attempts to automate flows between existing applications that have grown over the years into redundant, ill-documented monstrosities. The CIO who attempts to create a “composite” application that invokes existing applications to carry out subprocesses finds that the output of each subprocess requires extensive effort to convert into the input to the next one, and that despite the fact that data is duplicated between applications, the whole process is repeated when adding a new subprocess or creating a new composite application. Meanwhile, the administrator must deal not only with the new data created by the composite application but also with changes in the data underlying each application that affects the composite application — and database administration is already one of the largest cost centers in typical large enterprise IT management.

Until now, efforts to tackle composite application creation have focused on the functional, or application, side of the job, not the data side. Standards-based approaches, such as Web services and “business process standards,” seek to define the interface between the composite “executive” that directs sub-applications to carry out business processes and the sub-applications themselves — decreasing the effort needed to add new process-connecting information flows. These approaches are worthy as far as they go, but they do not go far enough.

Applications will run into serious difficulties when forgetting the data. Conversely, those who identify and surface metadata across their major applications will reap major benefits in speed to implement and upgrade composite applications, knowledge of proprietary competitive advantage information, and flexibility to perform “what-if analyses.” Moreover, Aberdeen asserts that a new technology that allows this type of “data minding” across application data sources is ready to use: EII, or enterprise information integration.