Intelligent Enterprise

I'm just back from Oracle's 11g launch event in midtown Manhattan, and I have to say I came away impressed. The database is at the center of Oracle's world, and company president Charles Phillips strutted the vendor's stuff — with his usual low-key swagger — on topics ranging from the firm's 30th anniversary to its 47-percent marketshare ("more than IBM and Microsoft combined" he asserted) to the long list of 11g upgrades and new features.

I lost track of how many of those features Phillips claimed to be unique to Oracle (and I'm following up with analysts to verify), but a few of the more impressive ones included OLAP cube-based management of materialized views and the use of standby servers to drive testing and performance improvements.

Materialized views are a technique for speeding multidimensional queries, such as those exploring sales across regions, products and customers. However, as the number of materialized views mounts along with query volumes and complexity, managing those views becomes difficult. In 11g, Oracle is using an OLAP cube to store up to millions of materialized views so they can be managed more efficiently.

"As data changes in the base data tables, those changes are automatically refreshed to the materialized views stored in the cubes," explained Chuck Rozwat, Oracle's executive vice president of server technologies. "The bottom line is more automation and a better ability to handle large amounts of data and complex queries with less of an impact on performance."

The OLAP engine used is neither Essbase, Hyperion's OLAP product, nor Express, a legacy product acquired by Oracle, but it was designed by Express engineers, said Andy Mendelsohn, senior vice president of Oracle's Database Group. "The big breakthrough here is that we're able to use the OLAP cubes as a transparent performance accelerator inside the [relational] database," said Mendelsohn. "The users are still happily using their SQL applications, and they won't even know they're using OLAP. We're moving OLAP from a specialized market to a much broader market among all our SQL users." He added that more than 60 percent of Oracles data warehousing customers use materialized views.

Data Guard was previously introduced by Oracle as a redundancy feature, but Phillips said customers complained about licensing a standby server that sat idle, for the most part, waiting to take over when a production server fails. To make use of the processing capacity, Data Guard can now port reporting and I/O-intensive backup activities over to the standby server, improving performance on the production server by offloading resource-intensive demands.

Data Guard also supports two other new features. First, database upgrades and patches can be implemented on 11g without taking the database down, which is achieved by switching between the production and standby servers. Data Guard also works hand-in-hand with a new Real Application Testing feature that brings Tivo-like replay capabilities to the database tier. Instead of having to recreate an extreme production environment in a test suite, you can simply capture a peak load period on the production server and then replay it on the standby server. Mendelsohn estimated the feature could cut test periods by as much 80 percent.

About the only aspect of today's event I found disappointing was Oracle's secrecy on when it might release a Windows version of 11g — a vagueness Phillips chalked up to regulatory scrutiny. "All we're announcing today is Linux in this quarter," said Phillips. "We'll have to come back to you on the other platforms because of the rules around disclosures these days."

Needless to say, the 11g database is very material to Oracle's overall financial performance, so a slip in a launch date of just a few days could have a major impact on quarterly performance. According Oracle User Group survey results shared today, 35 percent of members expect to upgrade to 11g within one year.