IBM is due to officially unveil its DB2 9 database, previously codenamed “Viper,” this month. As well as fully detailing functionality and pricing for the hybrid relational and XML database, the company has committed to a July 28 worldwide shipping date for the product.
Big Blue is positioning DB2 9 as one of its most important database releases in recent years, given its capacity to store and manage not only structured relational data but also unstructured XML data such as audio, video and Web pages natively.
Cindy Taylor, data management executive for IBM Canada Ltd. in Markham, Ont. said that DB2 9 is a “very significant release,” particularly when considering the new storage compression and XML capabilities.
“It’s a unique way of dealing with (XML data) without having to shred it,” Taylor said. That additional ability contrasts with rival products from Microsoft and Oracle, which rely on the older, slower approach of reformatting XML data or placing it in a large object before storing it as relational data.
The storage features enables users to compress database tables’ rows by scanning for duplicate data and building dictionaries to assign short, numeric keys to those entries, which according to IBM, can result in providing disk, I/O and memory savings.
“The database doesn’t have to have multiple repositories anymore…you can integrate relational and XML data and use them in the same apps,” Taylor said. This allows database administrators to maintain things like digital signatures, which is a requirement for privacy and compliance legislation — to be able to show direct access to the actual document, Taylor noted.
But one industry analyst notes that while the new features may in the short term give IBM a leg up in a tight database market, it is too early to tell if the offering can make significant gains against Oracle. “DB2 9 does look promising,” said Wayne Kernochan, president of Infostructure Associates. However, it’s premature to say what kind of performance advantage all the new features in the IBM database will end up providing for users, he added.
While some organizations may find the native XML handling in DB2 9 useful, that kind of capability isn’t likely to persuade existing customers of other databases to jump ship to IBM.
“Enterprises find it very difficult with existing applications to switch databases,” Kernochan said. “People just tend not to do it.” However, for customers with new applications, DB2 9 is likely to prove attractive, he added.
When DB2 9 ships, it will be the first time that IBM has simultaneously released three editions of the database — enterprise, workgroup and developer — based on a single code base.
Other key features include a new storage compression technology known as Venom and improvements in data management and security capabilities. DB2 9 can now support three types of database partitioning at the same time — range partitioning, multidimensional clustering and hashing. The new database includes a number of technologies like range partitioning and some autonomic capabilities that IBM acquired when it purchased rival database player Informix in 2001.
DB2 9 will initially ship on Microsoft’s Windows operating system, IBM’s AIX flavor of Unix and a number of Linux distributions including Red Hat, Suse, Ubuntu and Mandriva. IBM does plan to support other operating systems such as HP-UX since DB2 has typically run on 34 different operating systems.
IBM also said it intends to bring out a version of DB2 9 for its System z mainframe later this year.
QuickLink 067369