Building for the future
- By Brian Robinson
- Mar 11, 2001
In some ways, the future of business intelligence (BI) applications runs parallel to the expectations of other parts of the information technology field. For example, visualization, currently only in embryonic form, couldprovide a powerful method of expanding BI analysis into the realm of full-scale modeling of events and "what-if" scenarios. The need to provide BI data to wireless devices will also require an expansion in the way BI applications present that data to users, in near-real time.
But a company called QueryObject Systems Corp., Roslyn Heights, N.Y.,is betting on a fundamental shift in the BI model, from one in which organizations know who the users of data are and so can tightly define user requirements,to one in which data will need to be delivered to a broad range of unknown types of users across the Internet.
That means the way that data is viewed in a BI system will have to change. In the opinion of QueryObject's leaders, that calls for a way to store and index massive amounts of data that will be available, and that can be matched with other organizational data, on a random-access basis.
"Over the last 15 years, the focus in BI has been on how people canform queries and generate reports," said Matthew Doering, senior vice president and chief technology officer at QueryObject. "We've more or less accomplished that. Now the problem is one of many people asking [different] queries at the same time. The need in the future will be for much more robust data structures."
Although those types of systems are not widely available today, agencies can begin to move in that direction by designing systems that are not strictly focused on answering only one set of questions using one set of data.
Brian Robinson is a freelance writer based in Portland, Ore.