TimesTen Release 6.0, which was under development prior to the acquisition, is available immediately at a price of $12,000 to $24,000 per CPU, depending on the size of the in-memory data store.
Oracle has broken out two pieces of the original product and now offers them as add-on options: a replication service called Replication " TimesTen to TimesTen, for hot backup in case of server or network failure; and Cache Connect to Oracle, a service that uploads a subset of Oracle data into TimesTen, then updates both the cache and the parent database on an ongoing basis.
The two add-ons are each priced at $6,000-$12,000, depending on the size of the data store in memory. Both are available immediately.
"By pulling forward data into an in-memory database, you get great performance," says Jim Groff, formerly CEO of TimesTen and now senior VP of Oracle's TimesTen product line. A data retrieval or record update that might take a few tenths of a second in a typical relational system will take nine millionths of a second with TimesTen, he says.
One improvement still to be made is getting TimesTen to recognize and use Oracle's modified SQL query development language, called PL/SQL, Groff says. Many applications written for Oracle's database use PL/SQL.
In-memory database systems are a special type of database where both the database software and the data it's managing reside in a server's memory rather than on a disk. That allows the system to act on calls for data much more quickly than traditional databases. TimesTen is often used to host stock price information for equities trading at financial services companies. It's also used in telecommunications and network services for high-speed tasks.
Other changes in Release 6.0 include expanded support for SQL constructs and Java Database Connectivity (JDBC) interfaces. Throughput for updates to in-memory data caches has been increased by 500%, Groff says. With larger addressable memories made possible by 64-bit operating systems, TimesTen sometimes work with caches that contain 20-to-30 gigabytes of data, compared to the two-to-five gigabytes that were typical one or two years ago.