Partition and Allocation of Objects in Distributed Application Development

In recent years there has been significant interest in the use of object-oriented techniques for the production of distributed systems. However, only a limited amount of work has been reported on the incorporation of distributed system design issues into object-oriented development methodologies. The identification of processes within a software system (partitioning) and the allocation of these processes to processors in the system is a fundamental problem in the design of distributed systems. Techniques for evaluating the three objectives of task partitioning: minimizing intermodule communications, exploiting concurrency, and limiting the size of processes are presented. The potential concurrency in the object model is graded into three categories. This grading of the potential concurrency between objects/classes is new and is anticipated may assist in the partitioning decision making process. Communication and execution costs required for software allocation are determined from the event diagrams and event traces. The aim of the paper is to use the object-oriented model to produce information that fits into existing non object-oriented decision techniques such as traditional graph theoretic or graph heuristic allocation techniques. CR Categories: Object-oriented design methods and software engineering practices; distributed systems.