A Performance Comparative of Vertical Fragmentation Table Using Bond Energy and Graph Based Vertical Partitioning Algorithm

The cost and the speed of data access on a query depends on the numbers of attributes. It is because the query function will search for data from various attributes of the database table, whereas not all the attributes needed on the accessed tables. To separate the attribute that is not necessary, it can be done using the Vertical Fragmentation methods. The purpose of this research aims to compare the vertical fragmentation of database tables using Bond Energy algorithm and Graph-Based Vertical Partitioning algorithm to create a distributed database. The results of both algorithms will determine which attributes will be separated. Then, the results of the two algorithms are compared and evaluated using Partition Evaluator to show the amount of the access data costs that are used, so we can get a conclusion which is better algorithm to be applied. Based on the experimental results, GBVP algorithm produce better partition with fewer computational complexity and generate higher Partition value. 5 Partition Evaluator table has a higher value respectively at 4.98; 10.41; 17.95; 4.64; and 14.07. For the query execution time, 6 tables produces a lower execution time by using GBVP algorithm with the value of each table at 0.003s; 0,002s; 0,002s; 0,002s; 0,003s and 0,003s.