Fragmentation of Multidimensional Databases.
D. Munneke, Kirsten Wahlstrom, Mukesh Mohania · 1999
. As diverse demands are placed on On-Line Analytical Processing (OLAP) systems, distribution of the supporting multidimensional database (MDDB), and thus fragmentation, become an issue. We consider a recent multidimensional data model, and show that two basic strategies of fragmentation are possible: slice 'n dice and sever. The operators that realise these methods are composed from the basic set given in the data model, and the strategies are shown to satisfy the fragmentation correctness rules. 1 Introduction Multidimensional databases (MDDB) exploit the multidimensional nature of On-Line Analytical Processing (OLAP) [Codd et al., 1993]. As the popularity of OLAP grows, an increasingly diverse group of users are making demands of the systems, possibly with conflicting goals. In order to meet the performance requirements of the systems, implementation as distributed databases will become an important consideration. Fragmentation is an important part of distributed design [ Ozsu and ...