Enterprise data integration has clearly "arrived." The road had many twists and turns, yet data integration has not just survived, it has grown in strength and stature. How do we apply our collective learning from market developments to position ourselves better for 2007 and beyond?Enterprise data integration, just a few years ago, meant no more than executing bucketfuls of spaghetti-like extract-transform-load (ETL) processes to load bulky and often unwieldy data marts and data warehouses. That was then. Spurred by product development and refocusing, artful solution convergence, and a flurry of mergers and acquisitions, the data integration landscape is now dramatically different. The primary goal remains to bring data from its source(s) to its destination(s) in a timely manner and useful form, but that is now a very loaded statement. You still have ETL, but in addition, you get access to a wide variety of data sources, services and applications in real-time, near-real time and batch modes. There's also data profiling, cleansing and standardization, query federation and virtual data models as well as master data management or "data verticalization" through hubs. These product hubs and customer hubs are glued together with integrated metadata management and service-oriented architectures, ready for consumption in your applications. Driven more by vendor innovation and "big picture" thinking than by customer demand, data integration moves ever closer to being a much-respected fixture in IT shops.
If you haven't looked at data integration solutions lately, do so today. In particular, customers who need data provisioning through enterprise application integration (EAI) and service-oriented and enterprise service bus architectures (SOA/ESB) would do well to take a close look at data integration technologies as well.
So where is enterprise data integration headed? For many vendors and customers, the primary purpose of integrating data across the enterprise is business intelligence (BI) or its latest avatar, corporate performance management (CPM). There's a lesson in this: if you are looking to maximize your return on BI/BPM investments, consider strengthening the "back end" data integration.
BI or CPM need not be the raison d'etre for data integration efforts. As SOA and collaborative solutions flourish in your organization, data integration becomes an integral component of the enterprise architecture and, thus, a key enabler of the business architecture. Data visualization and reporting solutions will remain important beneficiaries of data integration, but let your vision go beyond BI and BPM.Enterprise data integration has clearly "arrived." The road had many twists and turns, yet data integration has not just survived, it has grown in strength and stature. How do we apply our collective learning from market developments to position ourselves better for 2007 and beyond?... BI and performance management need not be the raison d'etre for data integration efforts.