Oracle may have soft feelings about its flagship Database 10g Server software, but the company is now showing other database products some love.
Oracle today launched Oracle TimesTen In-Memory Database Release 6.0, the first product upgrade since Oracle purchased the company in June.
The TimesTen software is a far different database than Oracle’s disk-based database software because it runs as part of the application, accessing a relational database that’s always in memory, using a fraction of the CPU, as compared to a fully cached database.
“Unlike Oracle 10g Database and 10g RAC, this is database technology optimized for the applications tier,” said Jim Groff, former CEO of TimesTen, who joined Oracle as a senior vice president.
“It runs in the same tier as the rest of Oracle Fusion Middleware and enterprise applications,” Groff continued. “But because we moved the database engine as well as critical data forward in that tier what can we do is get blazing-fast response time. The path to the data is a lot tighter than it is in a typical multi-tier network.
That response time allows data queries to be answered within microseconds, compared to milliseconds or whole seconds typical of traditional databases like Database 10g.
This makes it a good solution to bandwidth challenges posed by computer paradigms, such as service-oriented architectures (SOA)
Groff said demand for a middle applications tier is rising, with such segments as Web services, business activity monitoring, complex event processing and RFID
“The role of data in that middle tier is rising rapidly and you need to take data out of the back-end realm and create a complementary management capability,” Groff said.
On that score, TimesTen 6.0 complements Database 10g through a cache connect option, combining the speedy response times of the TimesTen software with the Database 10g’s ability to scale. Cache Connect loads a subset of Oracle data into TimesTen, propagates updates in both directions, automates SQL requests for non-cached data and resynchronizes data after failures.
The new version also features 400 percent greater throughput; an advanced data replication option to thwart failures; and broader support of major J2EE application servers and standards such as SQL and Java.
The Oracle Times Ten In-Memory Database product is generally available with prices ranging from $12,000 to $24,000 per CPU based on the size of in-memory data store. The Replication and Cache Connect to Oracle options are each available for $6,000 to $12,000 per CPU.