A New Allocation Technique for Methods and Attributes in Distributed Object-Oriented Databases Using Genetic Algorithms

Abstract: With the wide increase in using distributed object-oriented databases, it became important to find an  efficient technique to store large-scale databases on the different sites. The allocation of object-oriented database involves allocation of both methods and attributes within the classes. The main objective of this paper is to introduce a new technique for allocating the distributed object-oriented database methods and attributes among N sites. The proposed technique uses genetic algorithm to find the best allocation (optimal or near optimal) of the object-oriented database methods and attributes to the available sites. A cost function that computes the total data transfer during the execution of queries is developed. The genetic algorithm steps use this cost function to evaluate the possible allocations of methods. Validation of the proposed technique is done via simulation. The experimental results of the proposed technique depict that it has a great impact in reducing the total time required to find the best allocation and in most of the cases it reaches the optimal allocation of the methods.

Keywords: Object-oriented database, allocation, methods, attributes, genetic algorithm, distributed database.

Received November 19, 2006; accepted June 17, 2007