What's New

Partitioned Databases

DB2 has extended the strength of its database manager to the parallel, multinode environment. DB2 Universal Database Extended Enterprise Edition allows a database to be partitioned across multiple, independent computers connected by a LAN. Each database partition or node (node was the term used in DB2 Parallel Edition for AIX) is a part of a database that consists of its own data, indexes, configuration files, and transaction logs. To the end-user and application developer, the database still appears as a single entity on a single processor. There are two main benefits that this provides. First, this enables an application to use a database that is simply too large for a single processor to handle efficiently. Second, SQL operations can operate in parallel on the individual database partitions, thereby speeding up the execution time of a single query or utility.

There are two types of query parallelism:

• Intra-partition parallelism refers to the ability to break up a query into many parts. This usually means subdividing what is usually considered a single database operation, for example an SQL query, into multiple parts, many or all of which can be executed in parallel within a single database partition.

• Inter-partition parallelism refers to the ability to break up a query into multiple parts across multiple partitions of a partitioned database, on one or more machines.

As your needs change you may find that your configuration is no longer appropriate. DB2 Universal Database lets you scale your configuration. Whether you are adding memory, storage, processors, database partitions, or changing environments (moving to SMP or MPP) DB2 can meet your current and future needs.

DB2 Universal Database offers many options on how best to match your hardware and application requirements with a specific DB2 product configuration.

Version 5

Single Partition on a Single Processor

This environment consists of memory and disk, but contains only a single CPU. It has been given many names: standalone database, client/server database, serial database, and single node/non-parallel environment.

Single Partition with Multiple Processors

This environment is typically made up of several equally powerful processors within the same machine (symmetric multiprocessor or SMP). Resources such as memory and disk space are shared. Machines with multiple processors have more memory and disks than those with a single processor.

Multiple Partitions Each with Its Own Processor

In this environment there are many database partitions, but each is on a separate machine and that has its own processor, memory, and disks. Machines are connected by a communication facility. This environment is known as a massively parallel processing (MPP) environment or a shared-nothing configuration.

Version 5.2

Multiple Partitions with Multiple Processors

In a multipartition environment each partition can have multiple processors. This configuration combines the advantages of SMP and MPP parallelism. Queries can be performed in a single partition across multiple processors, or they can be performed in parallel across multiple partitions. This environment is called an SMP cluster.

Windows NT and Solaris

In addition to AIX, DB2 Universal Database Extended Enterprise Edition is now available for the Windows NT and Solaris operating systems.

[ DB2 List of Books | Search the DB2 Books ]