In-Memory Analytics
Brief overview
In-memory analytics allow organizations to retrieve transactional data, close to real-time, by storing the data in-memory. In-memory systems extract information from disk-based databases and stores it in an In-memory database. The Disk-based database also stores a back-up in a memory-cached database (Acker). The process of extracting information from disk-based databases takes milliseconds, and retrieving data from the in-memory databases only takes microseconds (Acker). The information is analyzed directly from a dashboard using a temporary data cube when it is complete. In-memory analytics not only speeds the process of accessing data, but creates new capabilities that are not possible using standard disk-based data warehouses (e.g. real time analysis and increased complex querying capabilities.)
Class material topic relationship
In class, we learned that the analysis of big data required organizations to retrieve data from a data warehouse and store it in a data mart, which feeds an analytical data cube to be used by analyst. However, the process for analyzing data from a disk-based data warehouse is time-consuming and does not produce dynamic information. In addition, organizations using disk-based data warehouses can only answer business questions if asked at least a day in advance. In-memory analytics eliminates the need for a data marts by storing transactional in-memory.
Case study
A consumer goods manufacturing named Unilever just switched from a hard-disk system to an in-memory database with an analytical platform called Hana. Unilever uses their new capabilities to make almost real time analysis for forecasting cost fluctuations of materials, while including other factors. The old disk based system only permitted them to retrieve quarterly information. The standard software BusinessObjects Planning and Consolidation software would take 446 seconds to retrieve the same records that now only take 30 seconds to get. Unilever’s goal is to centralize all its ERP Systems allowing it to do complex analysis on every aspect of their operations.
Acker, O., Gröne, F., Blockus, A., & Bange, C. (2011). In-memory analytics – strategies for real-time CRM. Journal Of Database Marketing & Customer Strategy Management, 18(2), 129-136. doi:10.1057/dbm.2011.11
Nick, Heath, (2012, Dec 5). How in-memory computing is helping Unilever to maximize profits. Retrieved from http://www.techrepublic.com/blog/cio-insights/how-in-memory-computing-is-helping-unilever-to-maxmise-profits/39749693/