Use cases where the enterprise can benefit from the technology in this preview
One common use case for is to offload analytic data marts from poorly performing warehouses. This allows line of business users to leverage their existing OLAP application tools and get faster reports in real time from the in-memory column organized data mart stores. The choice to offload the enterprise data warehouse (EDW) to a data mart becomes dead simple. Where typical data mart environments having to support a specific LOB require EDW-like activities to get up and running, the technology preview show how this complexity could be removed. Architecture, capacity planning, storage choices, tooling, optimization and tuning, index creation all go away. Create tables, load and go becomes a reality.
A second popular use case is to create data marts directly off of the transactional database for use by lines of business. Many lines of business find that the number of transactional data systems is sufficient to support the business, however the data from these OLTP or ERP systems has not been transformed into actionable information. By directly accessing a transactional database to turn the data into actionable information you create a dynamic where mixed workloads compete for processing power. This can become a problem as now resources both human and machine are spent on supporting queries which are running the business (transactional) and queries which are analyzing the transactions (analytical). Additionally, the analysis of the of the data is a mystery since most times, the transactional data is not organized in a fashion which would suit an analytical workload, i.e.. Data is not optimally organized to be analyzed. A solution to this dilemma is through the creation of a data mart to offload transactional data which would then become actionable information for analytics to run against. Given the columnar method of data store, and the simplicity of the loads, data does not have to be indexed and organized to support the business queries. As well, the data mart can now contain and handle the historical data which would be dumped out of the transactional database.