In-Depth

Oracle Caches In on Its Acquired TimesTen Technology

The Oracle In-Memory Database Cache keeps frequently accessed data in memory for faster performance. We look at the strengths, weaknesses, opportunities, and threats of this technology.

• By Mike Schiff
• 05/14/2008

With the release of the Oracle In-Memory Database Cache option (which allows frequently accessed data to reside in memory), Oracle demonstrates that it is committed to supporting acquired technology and leveraging it to augment its own functionality and improve its performance.

Our SWOT -- strengths, weaknesses, opportunities, and threats -- analysis looks at the option from many angles:

Strengths:

• The In-Memory Database Cache should significantly enhance the performance of applications that require frequent or near-instant access to a database subset. It supports both read and write SQL operations, thus allowing it to be used in both operational and analytic environments.
• The cache option includes the TimesTen In-Memory database, TimesTen-to-TimesTen replication, and the Cache Connect to Oracle. It provides automatic data synchronization between the in-memory cache and the Oracle database as well as with an active standby configuration.
• It allows Oracle to credibly claim that although IBM may have acquired in-memory database technology in its November 2007 acquisition of Cognos (who had previous acquired Applix in October 2007 and its in-memory TM1 OLAP server), Oracle has integrated the in-memory database technology that it acquired with its acquisition of TimesTen in 2005 with its own.
• Oracle is continuing to offer its TimesTen In-Memory database as a standalone product, albeit at the same price as the Oracle In-Memory Database Cache.
• This is not an intention announcement; the Oracle In-Memory Database Cache is currently available and can be used with Oracle Database 10g R2 and Oracle Database 11g.

Weaknesses:

• Like many of its other options, the Oracle In-Memory Database Cache is only available with the Enterprise Edition of the Oracle Database.
• The Oracle In-Memory Database Cache is priced on a per-processor basis and costs an additional 90 percent of the per-processor price of the Oracle Database Enterprise Edition.
• The TimesTen In-Memory database can only directly integrate with the Oracle database.

Opportunities:

• The Oracle In-Memory Database Cache certainly has appeal in very-high-performance environments such as real-time analysis and execution (e.g., stock and bond trading, call centers, communications, defense and homeland security applications, etc.); it can be used in both operational and analytic environments since it supports both read and write access.
• Oracle can make a strong and credible case for using a very large multi-terabyte Oracle Database 11g together with a smaller In-Memory Database Cache as enabling technology for these types of applications.
• Since it can be deployed in an active standby configuration, with cache tables replicated in real-time, Oracle can position the In-Memory Database Cache as part of a high-availability solution.
• Oracle can cite its continuing investment in TimesTen technology as evidence of its commitment to other acquired technology.
• Oracle can credibly position itself as both a tools and a solutions vendor by stressing that it is committed to enhancing its database and middleware toolset while continuing to expand its enterprise application and business intelligence product portfolio.

Threats:

• Oracle is certainly not the only database vendor with in-memory database capabilities. IBM can be expected to compete aggressively, at least on the analytical side, with the Applix TM1 technology it acquired with its acquisition of Cognos.
• Oracle’s competitors may falsely suggest that Oracle requires the use of its In-Memory Database Cache option in performance-critical applications while reminding prospects that this will increase the per-processor price by 90 percent.