Valley based vertical partitioning in distributed database

Abstract

A Vertical Partitioning approach is defined as the method of dividing the attributes of a relation. An efficient Vertical Partitioning method always puts frequently accessed attributes of a relation together in a fragment. Various Partitioning algorithms have been proposed by several researchers. Still there is a scope of further improvement in previously developed algorithms. In this paper a new algorithm is proposed for Vertical Partitioning in Distributed Database. The proposed algorithm is named as Valley Based Vertical Partitioning Algorithm (VBVPA).This algorithm makes use of Clustered Affinity Matrix (CAM), which is derived using Attribute Usage Matrix (AUM) and Frequency Matrix (FM)