Wherever possible, isolate the database server from the application??™s CPU
requirements. The data distribution techniques described in the networking chapters
of this book will result in data being stored in its most appropriate place, and the CPU
requirements of the application may be separated from the I/O requirements against the
database.
Consider using Oracle??™s Real Application Clusters (RAC) technology to spread the database
access requirements for a single database across multiple instances. See Chapter 10 for
an in-depth review of RAC features along with step-by-step instructions on how to create
a RAC database.
Use the database resource management features. You can use the Database Resource
Manager to establish resource allocation plans and resource consumer groups. You can
use Oracle??™s capabilities to change the resource allocations available to the consumer
groups. See Chapter 5 for details on creating and implementing resource consumer
groups and resource plans via the Database Resource Manager.
Use Parallel Query to distribute the processing requirements of SQL statements among
multiple CPUs. Parallelism can be used by almost every SQL command, including select,
create table as select, create index, recover, and the SQL*Loader Direct Path loading
options.
The degree to which a transaction is parallelized depends on the defined degree of parallelism
for the transaction. Each table has a defined degree of parallelism, and a query can override the
?–
?–
?–
?–
?–
Chapter 8: Database Tuning 245
default degree of parallelism by using the PARALLEL hint.
Pages:
385
386
387
388
389
390
391
392
393
394
395
396
397
398
399
400
401
402
403
404
405
406
407
408
409