Online Analytical Processing
Online Analytical Processing (OLAP) is counted in addition to the data mining to the methods of analytical information systems. OLAP is still assigned to the hypothesis- based analysis. The analyst needs to know before the actual examination, which requests he would like to put to the OLAP system. His hypothesis is then confirmed or refuted by the analysis result.
OLAP systems obtain their data either from the operational databases of a company or of a data warehouse (data warehouse). This prevents that the analysis data with the transaction-oriented data bases in contact, and the efficiency is deteriorated. Similarly, the power of an OLAP system on the used data storage form and their connection to the analysis client depends.
In contrast to online transaction processing (OLTP ) is here to perform complex analysis projects in the foreground, causing a very high data volumes. The goal is to win by multidimensional viewing these data, a decision support analysis result. As a special target group here is named in his decision-making role management.
The structure underlying OLAP is an OLAP cube ( cube english ), which was created from the operational database. This follows a multi-dimensional data point oriented logic as opposed to line-oriented logic for online transaction processing (OLTP ).
Species
A distinction is made between ROLAP ( " relational OLAP " ) that accesses a relational database, and MOLAP ( " multidimensional OLAP " ) that accesses a multidimensional database. HoLAP ("H" for " hybrid " ) is an intermediate form between ROLAP and MOLAP. Each type has advantages and disadvantages.
MOLAP store numbers in the form of data points. This MOLAP store has a performance advantage over OLAP systems, the data on relational basis as records.
Predicted OLAP systems have better performance as OLAP systems, the expected runtime.
In memory systems have better performance than disk-based systems, but they must economize carefully with the memory.
ROLAP scaled better, but depending on the performance of relational sources used slower than MOLAP. This is ROLAP because the data in addition to the part may already precalculated aggregations in a versatile, but possibly slower relational database be stored while the data in MOLAP are in proper, quickly accessible form as a data point. One advantage of ROLAP is again that less space is needed because data is retrieved from existing databases. This is useful especially in the evaluation on the basis of mass data in complex data warehouse environments.
HoLAP often provides a good compromise between ROLAP and MOLAP.
A fourth type of architecture is referred to as dolap (D for desktop). Here, the base data is first imported locally in the analysis client to carry out a local analysis. A disadvantage of a possibly weak hardware design, however, can be seen here. Time intensive for OLAP is not the analysis of the data, but the creation and refresh the applied cubes.
Another, increasingly popular type is memory- based OLAP dar. up all data held in RAM and calculates all values in real time. This technique was limited in the past with regard to the amount of data. Due to the increasing popularity of 64- bit computer architectures (see 4 GB limit ) can be analyzed but nowadays even large amounts of data with memory- based OLAP.
OLAP tools are often characterized by multidimensionality. Through this multi-dimensionality, relevant economic indicators ( eg sales or costs sizes) based on different dimensions (for example, customers, regions, time ) can be viewed and rated multidimensional. For pictorial representation OLAP cubes are used. These cubes are divided into different dimensions, which are in turn subdivided into elements. These elements form a compression tree, or more generally a directed acyclic graph representing the aggregations.
Requirements for an OLAP system
12 rules by Codd
The OLAP term was coined by Edgar F. Codd in 1993. He first formulated 12 rules that he has extended to last 18 rules. This evaluation rules presented the first request list to a OLAP system The meaning of these rules for the valuation of an OLAP system can not today be classified as very high. This is due in particular to their strong application-oriented focus and her partly controversial rules. The rules were developed from the cooperation with the company Arbor, which had just introduced the OLAP database Essbase - Essbase is now developed and distributed by Oracle under the product name Hyperion Solutions.
Because of its pioneer status, the rules are often quoted:
FASMI rules by Pendse and Creeth
Pendse and Creeth presented in 1995 (Ref.: Pendse ) under the acronym FASMI five vendor- independent evaluation of rules in order to describe the OLAP concept. FASMI stands for "Fast Analysis of Shared Multidimensional Information" and states as follows:
In summary it can be stated that the FASMI rules more responsive to user needs than on technical requirements. Overall, however, they are less specific than the rules by Codd, so much more systems can be assigned to the OLAP according to this definition.
Market Overview
In The OLAP Report, 2006, the international OLAP market shares as follows:
In addition there are with the company Palo Jedox from Freiburg a German competitor in the field of open source software.