IBM has added new technology to its DB2 database software, designed to make data analysis faster and more economical.
IBM DB2 10.5 with BLU Acceleration features what the company calls “dynamic in-memory technology” that loads terabytes of data in Random Access Memory, speeding up query workloads. In addition, the platform can process row-based and column-based tablets simultaneously within the same system. Users can perform analytics on compressed data without needing to decompress it first, which could save on storage space.
IBM claims that BLU Acceleration can operate multiple times faster than other in-memory database systems. “Some queries that took 7 minutes were shown to have dropped to 8 milliseconds, thanks to the innovations in BLU Acceleration,” Bob Picciano, general manager of IBM Information Management, wrote in a June 26 statement.
But IBM faces significant competition in the in-memory arena. SAP has pushed its HANA in-memory platform as a competitive differentiator, baking the technology into everything from its Business Suite ERP offering to its CRM software. Over the past few quarters, the company has even indicated that it would expand HANA beyond its traditional enterprise audience, focusing on small- to midsize businesses with a need to store and analyze data.
When IBM first announced BLU Acceleration in April, it wasn’t the only company rolling out something to challenge HANA. Oracle used this year’s Collaborate ’13 to unveil its new Oracle In-Memory Applications for Oracle Engineered Systems, which it claimed would leverage DRAM, flash memory and the InfinitiBand network fabric to run processes much faster than other commodity hardware.
If that wasn’t enough competition, Microsoft has been working on a memory-optimization platform known as “Hekaton,” which will layer in-memory technology into the company’s next major SQL Server release. “In-memory computing is a core element of Microsoft’s strategy to deliver a data platform that enables customers to analyze all types of data while also accelerating time to insight,” Ted Kummert, corporate vice president of Microsoft’s Business Platform Division, wrote in a November 2012 blog posting. “Our approach to in-memory computing is to provide a complete portfolio for all application patterns, built into our existing products that enable rapid insights on any data, structured or unstructured.”
Can IBM’s own in-memory solution stand out in that sort of marketplace?