Operational workloads are essentially a combination of analytics and transactional processing, said Bernie Spang, director of strategy and marketing for database software and systems at IBM. The platform needs to be able to perform complex analytics while loading operational data and letting applications access the data warehouse, in other words.
IBM has added to its PureSystems line of integrated server bundles with what it calls PureData models for analyzing petabytes of data.
As this story from ComputerWorld U.S. notes, there are three intial bundles optimized for transactional, operational and big data analytics. IBM says organizations can have the system up and running in 24 hours, and are big enough to handle more than 100 databases on a single system.
“The family as a whole was designed to dramatically simplify enterprise computing,” said Pete McCaffrey, director of PureSystems category marketing. “We’re extending the family to take these same attributes of simplicity and extend it to the world of big data.
For example, he said, the system could be used to detect fraud while a transaction is in process, or in a call centre, looking for insight on a customer who’s on the phone. This requires high performance hardware under the hood, Spang added.
“For instance, our operational analytics model is built on Power Systems technology and leverages the high threading capacity in the Power architecture [used in IBM’s line of high-end servers], and how the AIX [operating] system takes advantage of that.”
All three models of PureData Systems come in different size configurations with an “easy path to add capacity,” he added.
For operational workloads, the largest configuration comes with 96 CPU cores, 768 GB of memory, 14.4 TB of flash storage and 324 TB of spinning disk, said Spang.
Bundles of servers and storage aren’t new –here’s a link to a recent story we carried on companies offering bundles of server, storage and networking, the so-called data-centre-in-a-box. The difference here is that IBM says it has created solutions tailored for particular industries. It’s the claimed speed of deployment that’s also interesting.
(With files from Brian Bloom)