Performance Optimization of Tree Structured Grid Services Considering Influence of Error Rate

Grid computing is an emerging computing paradigm that will have significant impact on the next generation information infrastructure. Due to the largeness and complexity of grid system, its quality of service, performance and reliability are difficult to model, analyze and evaluate. In real time evaluation, various noises will influence the model and which in turn accounts for increase in packet loss and bit error rate (BER). Therefore, a novel optimization model for maximizing the expected grid service profit is mandatory. This paper presents the novel tree structured model, is better than other existing models for grid computing performance and reliability analysis by not only considering data dependence and failure correlations, but also takes link failure, packet loss & BER real time parameters in account. Based on this model, an algorithm for evaluating the grid service time distribution and the service reliability indices are suggested. The algorithm is based on the graph theory and probability theory.