IBM is launching a beta program for a T Rex version of its DB2 8 database that will run mainframe applications faster, explosively grow potential address storage and revolutionize the alteration of databases.
The Somers, N.Y., company on Tuesday announced T Rex, its next-generation mainframe. Starting with 32 chips, the eServer z990 will grow to 64 chips by 2004.
The T Rex version of DB2 includes 100 new or enhanced features that focus on improved performance, scalability, continuous availability, application porting and security, all geared to exploit the new z/OS and zSeries architecture, officials said. Officials also said that this is the first IBM product to engineer an exclusive 64-bit solution for the mainframe.
Potential address storage space will jump from 2 gigabytes to 16 exabytes. Virtual storage under 64-bit will enable one, large, virtual space, rather than a number of smaller hyperspaces and data spaces in addition to the address space. Limits for storage use have been lifted so that customers can scale farther with more memory and more effectively. Outages are also expected to be reduced, and monitoring is expected to be simplified.
All this works out to less housekeeping for the DB2 application, freeing it up to do a "lot more real data management, instead of maintenance of spaces and back-and-forthing to the disk," said Jeff Jones, IBMs director of strategy, Data Management Solutions.
"Longer table names, longer SQL statement lengths, longer key sizes, longer character strings—a whole bunch of things have grown immensely," said Jones. Inflating these features means DB2 will be able to bite into gargantuan databases, he said. "Much more can be done in memory in one, big step: difficult queries, large databases can be processed much, much faster."
In addition, in this upgrade DB2 acquires Online Schema Evolution, which enables DBAs (database administrators) to alter databases on the fly without bringing them down. Thats a boon for big databases that would otherwise take days to change, as DBAs stop the system, copy data out, change the structure of the database to reflect whatever changes are being done to partitions and, finally, reload the database, Jones said. This new feature can handle schema changes such as adding columns, adding partitions and managing multiple versions of an object in the database at the same time.