C H A P T E R 15
Executing SQL in Parallel
Parallelism can help improve performance on particular operations simply by assigning multiple resources to a task. Parallelism is best used on systems with multiple CPUs, as the multiple processes used (that is, the parallel processes) will use those extra CPU resources to more quickly complete a given task.
As a general rule, parallelism is also best used on large tables or indexes, and on databases with large volumes of data. It is ideal for use in data warehouse environments, which are large by their nature. Parallelism is not well suited for OLTP environments, just because of the transactional nature of those systems.
In order to use parallelism properly, there are several important factors to ...