Restructuring distributed object-oriented software using hierarchical clustering

The software restructuring techniques present solutions for the software-hardware mismatch problem in which the software structure does not match the available hardware platform. In Distributed Object Oriented (DOO) systems, software engineers face many challenges to solve the software-hardware mismatch problem. One important aspect of DOO software systems is the efficient distribution of software classes among the different nodes while maintaining low-coupling and high software quality. In this paper, we present a new methodology for efficiently restructuring the DOO software systems to improve the performance and to solve the software-hardware mismatch problem. In our method, we use the hierarchical clustering technique to opt the classes to be grouped together and according to the customer hardware organization, we pick the level of the hierarchy that have the appropriate number of clusters to be allocated to the set of available nodes in the customer distributed system.

tex.type: Proceedings paper