Performance and reliability optimization for distributed computing systems

This paper discusses a performance and reliability optimization approach for distributed systems under a given budget constraint using the genetic algorithm (GA). The overall effectiveness of a distributed system is measured in terms of the average network throughput. This measure computes performance and reliability from network connectivity point of view. In order to carry out this optimization a distributed genetic algorithm (DGA) based scheme is developed. To demonstrate the effectiveness of the proposed approach, the results obtained from the distributed genetic algorithm approach is compared with single machine genetic algorithm and optimal solutions computed using exhaustive searches. Moreover, a brief discussion on the speed up over single machine implementation is also included.