In-Depth

Dashboard-Driven Analysis, ROLAP-Style

It’s an in-database, in-memory ROLAP engine that—officials say—could be just the Rx for dashboard-driven analysis.

• By Stephen Swoyer
• 03/01/2006

Last month, business intelligence (BI) start-up AppFusion announced Advanced Modeling, the newest deliverable for its eponymous BI & Collaboration platform. It’s an in-database, in-memory ROLAP engine that—AppFusion officials say—could be just the Rx for many dashboard-driven analysis efforts.

ROLAP has a number of advantages, including faster load times, standard SQL accessibility, better support for textual descriptions, and more flexible data modeling. Such advantages are typically offset by a host of (mostly symmetrical) disadvantages, including slower query performance and more demanding processing and system memory requirements. In-memory ROLAP is one way to offset these shortcomings, however.

That’s just what AppFusion’s new Advanced Modeling deliverable purports to do. It allows business users to perform hierarchical or customized aggregations, time series analyses, iterative calculations (e.g., simultaneous equations) and financial modeling (e.g., budget and sales forecast variances) right from the AppFusion BI & Collaboration Dashboard environment.

AppFusion officials have a somewhat hyperbolic take on the potential advantages of the new tool. “Advanced Modeling eliminates the need for writing pages and pages of SQL statements, or maintaining a complicated SQL abstraction layer or OLAP cubes that have to be designed, populated and continually processed,” said AppFusion CEO Arshak Navruzyan in a statement. “By… lowering the threshold for creating Actionable Dashboards, we’re enabling business users to open up different parts of their business to the use of our Actionable Dashboards for a true Enterprise-Wide adoption of BI.”

Hyperbole notwithstanding, Advanced Modeling does appear to deliver the dashboard goods. It lets users query results in the form of multidimensional arrays and then apply formulas—including interdependent calculations with inter-row and inter-array references—to calculate new array values. The idea, AppFusion officials say, is that organizations can now deploy BI & Collaboration-based dashboards (powered by the new Advanced Modeling component, of course) to replace Excel spreadsheets, among other solutions.

Of course, they can already accomplish the same thing manually—using SQL join and union operations, for example. This is pain-staking and laborious, AppFusion officials argue, and doesn’t take into account the complexity (in terms of data modeling and algorithms) involved in doing so. The new Advanced Modeling plug-in also eliminates other traditional steps—e.g., building cubes outside the database, copying data into the cube-building application, processing the data, and loading the results into the database—associated with manual approaches. The new plug-in can exploit database-specific parallel query features to further accelerate performance.