BI in Action

So here's a typical BI-related dilemma. By the time you get data into a warehouse or even a database, then analyze it, the information's old, and the decision you're trying to make is in danger of being out of date or irrelevant. You don't have time to turn real-time data into traditionally manageable forms. You've got to analyze and triage the data as it arrives.

That's what SQLstream, a company that just announced release 2.0 of its innovative solution, intends to help you do. The solution basically analyzes data as it streams into a business environment, using queries based on business rules and goals.

Previous approaches to this problem, including proprietary hardware and throwing lots of boxes at the problem, create their own challenges. SQLStream decided instead to apply SQL, the popular and well-understood Structured Query Language, to the problem. User-defined SQL queries continuously applied to data "on the wire," with results pushed to the information consumers as changes occur. It's basically querying data that's coming - "the future" - rather than the past.

This approach can triage the data and focus on information generating useful initial query results, instead of trying to capture everything and wade through it with analysis, as is done with most data warehouses. Also, the SQLstream approach does not remove the need for or value of a data warehouse. Where a warehouse exists, SQLstream helps to reduce volumes, and focus the warehouse on what it does well - historical analysis - while adding proactive/real-time data stream analysis features previously unavailable. And where no warehouse exists, SQLstream helps to translate high volumes of streaming data into information that can actually be used and acted upon by mere business mortals.

I've chatted with corporate management and seen a credible demo of the software. It has numerous applications, such as detection and prevention of fraudulent online transactions, for data-intensive businesses, and what business isn't data-intensive these days? It's native, standards-compliant SQL, which minimizes developer training (and, hopefully, whining). Also, the results of initial queries can help to refine and improve those very queries. Finally, queries and other application elements are reusable, to make new applications easier, faster, and more consistent to build. (There's even a Java interface that generates live code compliant with the Java Database Connectivity (JDBC) application programming interface (API).)

If your enterprise has challenges related to managing and leveraging high volumes of rapidly changing information, or to getting more business value out of database and data warehouse investments, you ought to check out SQLstream 2.0. Ditto if you are, say, attempting to map enterprise information resources and their interdependencies. Or perhaps to gain insight as to how users are using those resources and the effects of that use on business and IT infrastructures. Just thinking out loud, here...

(Oh, and by the way, if you do check out or have already checked out SQLstream, let me know your experiences and thoughts, please. After all, we now already know what I think...)