Partitioning Multimedia Objects for Optimal Allocation in Distributed Computing Systems

Object partitioning is an essential mechanism to improving the performance of complex object-based systems. An essential step in developing an ee-cient information system in such a distributed environment is to optimally distribute data among the applicable sites. As a result, network communication overhead becomes a concern during object partitioning. For multimedia (complex) objects, the communication overhead is greatly aaected by the way related or unrelated objects are aggregated for distribution. In this paper, we present two models of multimedia data models and discuss the techniques for eecient distribution of the objects. One model exploits the Mincut Algorithm and introduces the concept of the binding strength to interpret the relationships between objects in order to develop a linear time partitioning algorithm. The other model utilizes user speciied or system determined degree of partitioning and bipartite matching to determine the aggregation and optimal allocation of objects.