Oracle and IBM are building native XML indexing capabilities into their respective database platforms to help customers rein in new applications not easily integrated with traditional data structure models.
The next release of the Oracle 10g database will feature XML Path Indexing, said Sandeepan Banerjee, Oracle Corp.s director of product management for XML, in Redwood Shores, Calif. Currently available upon request in the products beta version, XML indexing will allow users to “walk” XML structures, or take a specific direction to reach a point in an XML document.
The absence of native XML indexing requires custom coding to parse the XML to target content, said Banerjee.
For its part, IBM will feature three forms of XML indexing with the release of its new DB2-hybrid database, code-named Viper, due in the first half of next year, with a full, open beta due later this year, said Bob Picciano, vice president of database servers at IBM, in Armonk, N.Y.
Vipers XML index options include a Path-specific index, with an enhanced B+ tree on path expression geared toward fixed schema and known access patterns; a Values index, including indexing on all elements and attributes but no full-text search designed for arbitrary XML without a full text; and a complete index, providing an index on all elements and attributes with full-text search.
“[Viper] will feature all three [forms of XML indexing], and the usage will vary based on the types of documents and collections and the type of workload an application has,” said Picciano. “Customers will use it to understand how big their documents are, how often they get updated [and] what is the size of a transaction.”
Aris Prassinos, distinguished member of technical staff for Motorola Inc., in Schaumburg, Ill., said he is eager to try the native XML indexing capabilities in Oracle 10g.
Prassinos said that he was disappointed the feature was not part of the previous Oracle 10g beta release. “Some of our customers are using different database schemas,” he said, “so were waiting for this XML indexing technology so we dont have to go to each customer and create specific XML indexing for them or know explicitly what an XML field is called. Transparent XML indexing can be very useful for us.”
Prassinos division at Motorola is using an Oracle 10g database to power its AFIS (Automated Fingerprint Identification Systems) integrated suite of applications and Printrak product. Used in conjunction with the FBI and other government agencies, the technology allows for the capture, processing, search for and storage of fingerprints, palm prints, facial images and other biometric data.
Any improvements in XML indexing would benefit Motorola customers that deploy AFIS around the world and use systems that must conform to national or state standards for shared data, said Motorola officials.
Among the Oracle 10g second-phase upgrades introduced in June that he found favorable, Prassinos said the databases enhanced security, backup and replication capabilities are vastly improved. In particular, he lauded Oracles new database change notification function, which helps indicate where improvement occurs and is needed, and OracleBackup to avoid costly third-party backup assistance.
“Were very interested in [OracleBackup]. It would cost about $10,000 to use Legato [Systems Inc.] for the deployment we have, but this new product has the same features [as Legato] but the cost is much lower,” said Prassinos.
In addition, Prassinos said Oracle made great strides in its Oracle 10g second-phase release in the direct mode functions within Enterprise Manager and new distributed LOB support capabilities.
As part of its Oracle 10g second phase, Oracle published APIs for clustering to push its grid capabilities and is offering transparent encryption to bolster security, according to Oracle officials.