IBM's roadmap for its DB2 Universal Database platform includes new features and refinements of existing functions.
While the company will continue to improve the database to take advantage of further advancements in hardware, new features such as better XML support and web services enhancements are being added, Janet Perna, general of data management at IBM, said.
"Databases are not at a point where there is nothing to do [to improve] them [anymore]," she said.
"Today, DB2 does support XML and it does it through extensions of SQL," Perna said in an interview at the IBM developerWorks Live conference. "We can ingest, bring in XML documents and store them as a BLOB in a column and as multiple columns."
What IBM really wants to achieve is to use the language XQuery, which is the evolving query standard for XML, and to store the XML data natively, Perna said. This would enable improved performance and is part of IBM's Xperanto database project.
IBM has introduced XML for Tables as a first stage of its XQuery implementation. This enables XML documents to be broken up into components and placed in multiple tables. So, for example, an employee name on a document could be stored in one table and the department listed on the document could be stored in another table, making for easier access, Perna said.
"It lets you take the information and parse it in a way that you can direct where you want pieces to go," she said.
Currently available in a pre-release form on the IBM alphaWorks site, XML for Tables will be included within DB2 eventually.
In early May IBM plans to release Fixpack for DB2 8.1.2, a DB2 service pack. The package will feature a function called WS Consumer, for building SQL statements that can access web services data from web applications. WS Consumer is already available for download on IBM's website.
IBM plans to keep maximising the database to exploit new hardware, such as 64-bit Intel Itanium processors, Perna said.
Decreasing database downtime remains a priority, with IBM seeking the goal of zero downtime, said Perna.
"We're not there yet but we're close. On 390 [model mainframes] we're close to zero downtime," she said.
She noted the growth in data sets.
"The demands on the database have become tremendous because we're now talking about petabytes of data," she said.
Also on the agenda is improving security via role-based security and enabling self-management of the database, with customers seeking zero management, she said.