Optimize Distributed Workload - Expert Oracle RAC Performance Diagnostics and Tuning - page 508

Workshop
To understand how such a distribution could be implemented, we discuss a typical workload configuration and
management scenario using an example. A global operational organization has five applications that they would like
to configure on a four-node RAC. Defined by the business needs, the applications are required to meet the following
requirements:

TAPS : This application is used by the client services to record customer interactions that
happen throughout the day. It's an OLTP application and requires a 24/7 uptime.

•

FIPS : This is a homegrown application used by various members of the organization.
Considering the various time zones, except for a few hours between 3:00 AM and 7 AM EST,
this application is also required to be functional most of the time.

•

SSKY : This is another reporting batch application that runs two or three times a week and
during weekends.

•

GRUD : This online reporting system is a subcomponent of both TAPS and FIPS . These reports
are triggered subprocesses and should support both applications. The load or usage of this
application is not very high, and an infrastructure to queue all reporting requests is in place;
hence, small outage of this system is acceptable.

•

TICKS : This is a critical seasonal application that runs twice a month. The application's
criticality is high such that during these two periods of execution, it should complete on time
and should have a very minimal to no failure rate.

•

All the previous applications are to be configured over a four-node RAC cluster. Each node in the cluster has
a public IP address, private IP address, and a VIP address. The cluster-level scan address has been configured.
The database is configured using the policy managed option and has two server pools defined. Table 15-6 details
the various characteristics of the application configuration across the available instances in the cluster.

Table 15-6. Application to Instance Mapping

Applications Database

Services

Type of
Service

Server Pool Preferred

Instances

Pool
Credentials

Priority/Criticality

SCDB_POOL SCDB_1,

SCDB_2,
SCDB_3

TAPS TAPS Client

Application

Max 3
Min 3

HIGH

SCDB_POOL2 SCDB_4 FIPS FIPS Client

Application

Max 1
Min 0

STANDARD

SCDB_POOL2 SCDB_4 SSKY SSKY Scheduler Job Max 1

Min 0

STANDARD

SCDB_POOL SCDB_1,

SCDB_2,
SCDB_3

GRUD GRUD Client

Application

Max 3
Min 3

LOW

SCDB_POOL SCDB_1,

SCDB_2,
SCDB_3

TICKS TICKS Seasonal

Application

Max 3
Min 3

HIGH