A Comparative Analysis on Performance of Minimum Spanning Tree and K-Means Clustering Based Vertical Fragmentation Algorithm
Abhishesh Dahal, Shashidhar Ram Joshi · 2019
Distributed database is an emerging technology in context of database which is supposed to possess a remarkable benefits for data storage in upcoming future. One of the major distributed database design issue rely upon fragmentation of the relations. Among the fragmentation task, vertical fragmentation exhibits an entangled behavior in comparison to other due to its multi solution nature. The main objective of research highlighted in this paper is to find the best vertical fragmentation algorithm to be utilized for splitting the attributes of a database relation. Two partitioning algorithms namely Minimum Spanning Tree based vertical fragmentation algorithm and K-Means clustering based vertical fragmentation algorithm are employed to generate variety of fragments of a database relation. Further, those generated fragments are compared through partition comparator on the basis of data access cost utilized. The final outcome of the partition comparator concludes the best vertical fragmentation algorithm for utilization. As per the experiment, K-Means Clustering based fragmentation generates less data access cost in comparison to Minimum Spanning Tree based fragmentation.