SAP’s Sybase division announced the latest version of its IQ columnar database on Thursday, a release focused on providing enough scalability to allow many thousands and types of users within a company to access the system.
To achieve this, Sybase IQ 15.3 employs a MPP (massively parallel processing) grid architecture, wherein all compute notes share processing power, memory and access to data, which is accessed through a SAN (storage area network).
There’s been much talk of late in the tech industry about the value of analyzing “big data,” the term often used for the massive volumes of Web, sensor and other information being generated by companies today, said David Jonker, director of product marketing.
That vision can’t be fully realized, however, unless a wide variety of business users get access to such information, according to Jonker. “We believe this transition requires a new kind of platform. The traditional enterprise data warehouse isn’t holding up.”
The MPP grid architecture will allow IT staffers to create “virtual data marts” for various users within a company by committing groups of physical servers to them, Jonker said.
Sybase’s move got a nod of approval from one observer.
“Virtual data marts are a good idea, and too few vendors are making it easy to spin them out. They let departments start doing analytics very quickly, yet allow IT to keep partial control,” said analyst Curt Monash of Monash Research.
The spirit of what Sybase is proposing, if not the same technical approach, is embodied in efforts such as rival Greenplum’s Chorus platform for data-mart building.
Other features in Sybase IQ 15.3 include a Web services API (application programming interface); a Ruby on Rails language driver; an API for analyzing multimedia content; and a Web-based administration console.
Sybase IQ 15.3 arrives amid a protracted and intense marketing blitz for SAP’s in-memory computing engine, HANA (High Performance Analytic Appliance). In-memory computing places data to be processed in memory instead of reading it off of disks, giving a performance boost.
There’s certainly a way for IQ and HANA to work together, according to Jonker. “IQ is a natural fit for storing a lot of data on traditional disk-based storage,” he said. In the longer term, IQ could serve as a “nearline” storage platform for HANA, aimed at data that needs to be accessed at some point, but which is not critical to have in HANA minute-to-minute, he said.
