Performance optimisation of the decision-support queries by the horizontal fragmentation of the data warehouse

The horizontal fragmentation of the data warehouse is considered as one of the important performance optimisation techniques of the decision-support queries. This optimisation is reached only if the large data volume of the fact table is horizontally fragmented. For that, the fragments of the fact table are always derived from the fragments of the dimension tables. Unfortunately, in this type of fragmentation, the fragments number can dramatically increase, and their maintenance becomes quite hard and costly. Thus, to reduce the number of the fragments and to further optimise the decision-support queries performances, we propose to fragment horizontally only the fact table by exploiting jointly: the selectivities of the selection predicates, their occurrence numbers, and their access frequencies. To prove the effectiveness of our fragmentation technique, we present at the end, a set of experimental results conducted under Oracle 10g on the APB-1 benchmark.