Guide to The Online Analytical Processing (Olap)

Google+ Pinterest LinkedIn Tumblr +

In computing, especially in the field of databases, online analytical processing (OLAP) is a type of computer application-oriented analysis on information in several axes in order to obtain summary reports such as those used in financial analysis. This application is different to OLTP (online transaction processing).

This process was been defined by Edgar Frank Codd in 1993 through 12 rules that a database must comply with if it wants to embrace the concept of OLAP.

   a) Multidimensional conceptual view
   b) Transparency
   c) Accessibility
   d) Consistency of response times
   e) Client-server architecture
   f) Independence of the dimensions
   g) Management of sparse matrices
   h) Multi-user access
   i) No restrictions on transactions between and within dimensions
   j) Easy manipulation of data
   k) Simple reports
   l) Unlimited size and unlimited number of elements on the dimensions

This concept has been applied to a virtual model of representation of data called OLAP cube or hypercube that can be implemented in different ways.

To optimize response times, the summary of the information is usually calculated in advance. These aggregations are pre-calculated values or the basis of the performance amplifications of this system. Some systems use data compression techniques to reduce disk storage space due to pre-calculated values


There are several similar versions of drivers that can be adapted to store data on different types of database to implement the concept of OLAP:

    – R-OLAP (relational OLAP)
    – D-OLAP (Dynamic OLAP or Desktop)
    – M-OLAP (Multidimensional OLAP)
    – H-OLAP (Hybrid OLAP)
    – S-OLAP (Spatial OLAP)

Each OLAP system has some benefits (although there is disagreement about the specifics of the benefits between providers).

Some MOLAP implementations are prone to the explosion of the database, this phenomenon leads to the need for large amounts of storage space for the use of a MOLAP database when certain conditions are met: high number of dimensions, pre-calculated results and sparse multidimensional data. The usual techniques for mitigation of the explosion of the database are not as efficient as desirable.


The M-OLAP is optimized for OLAP multidimensional analysis. It is a form of multidimensional hypercube to represent data in the form of a cross of n dimensions, these dimensions may be more or less dense, thus characterizing the density or sparsity of the cube.


In the world of business intelligence, R-OLAP is a technique for modeling and data storage based on a relational structure. It leverages existing resources (licenses, material resources …) and, as such, does not require additional investment in a multidimensional database.

Examples of engine R-OLAP: Microsoft Analysis Services, Oracle 10g, Informix and MetaCube MicroStrategy DSS Agent.


The H-OLAP is a hybrid between M-and R-OLAP OLAP. The multidimensional structure of a hypercube is used to aggregate data. When access to a basic level of detail finer is needed, conventional relational tables are used: the mechanism of the drill through. Example of motor H-OLAP: Oracle OLAP, Microsoft Analysis Services


Platform supporting visual exploration and analysis of spatio-temporal easy and timely data in a multidimensional approach at several levels of aggregation via a map display in tabular or statistical chart.

The underlying idea is that the representation of the data should be tabular as is the case for relational databases. We must be able to present the data as we wish.



About Author

Leave A Reply