All Is Not Lost
Performing iterative calculations against transaction data can become time intensive, especially when there are millions of rows in the table.Performing iterative calculations against transaction data can become time intensive, especially when there are millions of rows in the table. Thats why online analytical processing (OLAP) has become so essential. But OLAP necessarily summarizes the data along the parameters of interest, called dimensions. In doing so, the ability to dig down to a transaction level of detail is lost. A different way of storing the data may just be the answer to provide the speed associated with OLAP and the detail provided by transaction-oriented schemas. Alterian has recently released its suite of products that complement traditional data warehouses and tools. Alterians core engine is rooted in column-based analytic technology (CBAT), which stores data in columns as opposed to rows. The technology, coupled with some clever (and proprietary) indexing schemes, can deliver some impressive performance gains. Alterians Kirk Fallbacher claims that in one instance a report that took 16 hours to complete was produced in just 45 minutes after the product was installed.
The company doesnt see CBAT replacing OLAP or transaction oriented systems, but as a complement to them. That concept was well received by analytical tool powerhouse SPSS, which recently inked a deal to include the Alterian technology into its own products.