TestProNews Q1 - 2006
An Overview of Olap Technology
The acronym OLAP stands for On-Line Analytical Processing. In a formal sense, OLAP
can be described as an approach to provide fast answers to complex analytical queries.
A key characteristic of OLAP is that it enables users to analyze and view multi-dimensional
data from many different perspectives. In other words, OLAP gives engineers and
managers the ability to perform trend, comparative, and time-based analysis, among
other things, in different dimensions such as time, region, operator, product, etc.
The term OLAP was introduced in 1993 by E.F. Codd and Associates, but the ideas
and technology have their origins long before then. Multi-dimensional analysis was
first addressed in 1962 by Kevin Iverson in “A Programming Language” (APL). Oracle
entered the OLAP market when it acquired the Express business from IRI Software
in 1995. With a similar approach, Microsoft purchased Panorama in 1997 and released
a significantly enhanced product in 1998 named OLAP Services.
Figure 1: Market share for OLAP. Source:
The OLAP Report
In September 2000, Microsoft announced the second version of their OLAP product
under the name Analysis Services and it has been the leading OLAP server for several
years (Figure 1). One month later, Oracle announced the integration of OLAP capabilities
with its Oracle product which would replace Express, indicating that Express was
at the end of its development life.
With the release of SQL Server 2005 in November this year, Microsoft introduced
its new version of Analysis Services which provides a unified foundation for all
traditional reporting, OLAP analysis, Key Performance Indicator (KPI) scorecards,
and data mining. Both Oracle and Microsoft have bundled their OLAP implementation
with their respective database product under a broader category named Business Intelligence
(BI).
Fast Access to Multi-dimensional Data
The goal of OLAP is to provide fast access to multi-dimensional data. Fast means
that the system should deliver responses to user queries within five seconds. This
speed is not easy to achieve, especially with large data sets, and a variety of
techniques are used to achieve this performance goal. A common technique to ensure
consistently fast query times is to pre-aggregate data. This is usually done through
a highly com-pressed multi-dimensional representation of data known as a cube.
A cube is a representation of multi-dimensional data in a dimensional model. The
cells of a cube contain data values. The intersection of one member on each dimension
represents a particular data value. In Figure 2, the selected data in the cube represents
the number of units that passed in Q1, tested in the United States by A. Hansen.
After the information has been selected, the user can immediately look at the cube
from a different perspective or even change its dimensions. OLAP allows the user
to view data in the cube from many different perspectives very easily and very fast.
In our example, the user may decide to refine their query by looking only at the
units that passed on the first attempt (i.e., first pass yield.) With a drag-and-drop
operation, they have immediate access to this information.
Figure 2: Multi-dimensional data represented in a cube
Conclusion
OLAP technology has opened a new world of possibilities and greater performance
in many different industries and applications. OLAP and Business Intelligence technologies
have been proven for many years in applications such as finance, marketing, and
sales. Arendar is bringing the power and benefits of OLAP technology to the test
and measurement world. While before a complex query could take hours or days, OLAP
technology enables test engineers and managers to get results in seconds, allowing
engineers and managers to have the right information at the right time to help them
make the right decision.
Ask the Expert
Email questions concerning this article to:
expert.testpronews@vi-tech.com
|
|
|
|
Read about the latest information for integrating automated test systems throughout the enterprise. |
|
|
|
 |
Download PDF |
|
|
