Improving Performance of Object-Oriented Software for Distributed Systems using Restructuring Technique

Abstract

Recent years have seen increasing demands for faster computers that have led to the availability of distributed systems after the great advances in technology which decreased the prices of computers and processors units. The growing importance of distributed computing systems is reflected by the large applications that embrace them. At present, distributed computing permeates almost all aspects of computer science and engineering. Meanwhile, Object-Oriented programming is widely accepted for designing and implementing software systems application domains such as; air traffic control, high­ speed communications, e-commerce and military systems. More and more software systems are built as distributed-object applications. Within this distributed heterogeneous environment, objects are the main building blocks that form the communication of distributed system. For such systems, performance is a key issue.

The main objective of this research work is to improve the performance of object-oriented based distributed systems. Two techniques were introduced to achieve such improvement. The first one is the restructuring technique which handles the software/hardware (SW/HW) mismatch problem. This problem often occurs in distributed computing environments. A novel restructuring technique was proposed to solve the mismatch problem taking into consideration both user and system requirements and/or constraints. The technique efficiently maps existing object-oriented software components into an available hardware satisfying the user and system performance requirements while achieving minimum total resource cost. A queuing model that considers the cost of both the execution and communication activities at each processing node is developed and used.