In the past, analytics got pushed out to a small number of Line of Business (LOB) and private / public sector Executive (CxO) decision makers on a regular basis – perhaps daily, after a nightly batch data update. But in today’s fast-paced business world, there are many more people who need access to data and analytics, and they need current, up-to-the-minute data, not yesterday’s news. Also, they want the ability to do ad-hoc drill down into specific areas, and see immediate results. Call centers, mobile sales applications, supply chain logistics, medical centers, real-time fraud detection or other applications supporting operational decision making, are all becoming more “real time.” This is Operational analytics – allows you to having have insights at your fingertips, when you need it, where you need it and how you need it.
A high percentage of queries for operational analytics systems—often up to 80 percent— are interactive lookups focused on data about a specific customer, account or patient. To deliver the correct information as rapidly as possible, the data warehouse supporting these systems must be optimized for the right balance of analytics performance and operational query throughput.
The IBM® PureData System for Operational Analytics helps organizations meet these complex requirements with an expert integrated data system that is designed and optimized specifically for the demands of an operational analytics workload. Built on IBM Power Systems servers with IBM System Storage and powered by IBM DB2–based InfoSphere Warehouse software, the system is a complete, out-of-the-box solution for operational analytics. Designed to handle more than 1,000 concurrent operational queries, it delivers mission-critical reliability and scalability with outstanding performance.
The PureData System for Operational Analytics is capable of handling mixed workloads i.e. handling the simultaneous presence of operational and analytic workloads. A larger CPU share is provided to the operational queries by the InfoSphere Warehouse/DB2 Workload Manager. This ensures the large number of concurrent operational queries (point queries, random reads and writes of a few records) is not affected by the fewer complex analytical queries (comprising of serial reads involving a large number of records).
The underlying InfoSphere Warehouse 10.1 Work Load Manager (WLM) simplifies the management of workloads - effectively grouping and isolating maintenance operations from business critical applications. It ensures completion of important requests by giving priority to user defined service classes. It helps maintain service level agreements defined by policies such as CPU limits i.e. the percentage of resources DB2 can consume, CPU shares - the percentage of the limit a service class can consume, concurrency capacity & Limits while maintaining operational functions.
Essentially queries are assigned to appropriate service class based on priority & policy which enables efficient distribution of workloads. The WLM assigns workloads based on priority and concurrency and monitors and readjusts to maintain a consistent operating environment.