R.S. Tsankova and M.R. Neykova (Bulgaria)
modelling, multidimensional databases, OLAP, memory consumption
The OLTP (On Line Transaction Processing) technology, based on relational databases is suitable for storing operational databases but OLAP (On Line Analytical Processing), based on multidimensional databases is more efficient for data analysis in Decision Support Systems (DSS). High levels of data aggregation and low response time are the main advantages of OLAP technology. High memory consumption is a serious disadvantage. Seeking compressing methods is one of the challenges for database developers. In this paper we propose algorithms for generation of effective multidimensional data models. The starting point is a multidimensional interpretation of a relational database where each aggregating attribute is mapped into a separate dimension in the multidimensional model. A way to optimize the model regarding its storage size is to decrease dimensions number by attributes uniting in common dimensions. We present different variants of incorporating attributes in common dimensions and discuss some requirements and limitations. Our goal is to decrease the multidimensional database size without worsening model’s functionality.
Important Links:
Go Back
