The RAC database provides the technology of cache fusion, which simplifies application deployment issues.
Before cache fusion had been fully implemented, many of the
applications had to selectively access the multiple instances in a
way that would not result in update contentions, thereby minimizing
pings and false pings.
With the full implementation of cache fusion,
many types of applications are able to connect and use the multiple
RAC instances while avoiding the performance issues that result from
cross instance updates. There are two broad types of data access
methodologies:
-
OLTP:
Traditionally, online transaction
processing (OLTP) users access the database resources or data
blocks randomly. The transaction life, or access duration, is
usually very short. With this method, there is limited conflict or
contention with the data sets.
-
DSS:
Decision support systems (DSS) and data warehousing (DW)
applications focus more on analysis of the data and the creation
of various reports. A data warehouse is a relational database that
is designed for analysis rather than transaction processing. A
data warehouse usually contains historical data that is derived
from transaction data as well as from other sources. The warehouse
separates the analysis workload from transaction workloads and
consolidates data from several sources.
Using tools such as online analytical processing (OLAP) extraction
engines or other statistical tools, large data sets are processed.
With this method, for any given query a large number of data blocks
are read and analyzed. Performance is a crucial factor with this type
of access.
The OLTP and DW databases have traditionally been separated into
different servers and instances. The data warehouse is updated or
refreshed by loading data from the OLTP on an appropriate schedule.
Data warehouses typically use an extract, transport, transform, and
load (ETL) process, which involves complex and time-consuming steps
such as data exports and network copies. The mixed OLTP and DW
databases used to compete often for resources, resulting in update
contentions. This created performance issues.
With multiple instances, the RAC database is in a perfect position to
segregate the activity on different nodes while still maintaining
single database storage. This not only results in better performance
levels for both types of data activities, OLTP and DSS, but also gives
administrative flexibility and cost savings.
Database Consolidation
The main purpose of database consolidation is to reduce cost.
Today, when so many software choices are supposedly free, DBAs
see the number of MySQL and PostgreSQL installations growing.
However, Oracle continues to dominant in the area of Enterprise
Applications. LAMP (Linux,
Apache, MySQL, and PHP) development is growing, but many
Enterprise
level applications, in the authors' experience, do not support MySQL
or PostgreSQL.
For 2015, Fortune magazine ranks Oracle Corporation as the sixth (6th)
most profitable technology company, making $4.3 billion.
Google came in seventh (7th) place.
MySQL and PostgreSQL have a long climb ahead of them before
they are major players in the Enterprise Application market.
As CTOs are forced to cut costs, one trend is to consolidate databases
and use Oracle Standard Edition where the features of Oracle
Enterprise Edition are not needed.
This could be called license scale-down.
The cost savings of using Standard Edition in place of
Enterprise Edition can be significant.
It is important to understand what features and options are
only available to Enterprise Edition.
