Estimating querying and maintenance costs for restructuring data cubes

Abstract

In their daily routine, enterprise decision makers use to analyze huge amounts of information in order to sustain their decisions and, consequently, ensuring success of enterprises business activities. Through time, on-line analytical processing systems have contributed decisively to the decision making process improvement, not only by granting extremely flexible data manipulation mechanisms, but also allowing the materialization of the analysis indexes required. However, that analytical "power" uses to exhaust the computational resources, especially disk space and processing time, especially materializing specialized views. Besides, as time goes by, multidimensional databases become very large, being its management very difficult. Aiming to optimize maintenance and operationality of such databases, we design a system that is able to restructure them in useful time and reduce multidimensional query processing time, according to the exploitation trends of knowledge workers. In this paper, we present the system's structure, its correspondent cost model, query and maintenance algorithms, restructuring strategies, and, finally, its distribution through several processing OLAP nodes.