Web cluster load balancing via genetic-fuzzy based algorithm

摘要

In this genetic-fuzzy based Generalized Dimension Exchange (GDE) method is proposed to uniformly distribute the unprecedented Web cluster workload. Fuzzy set theory is used to capture the vagueness of the workload during redistribution period. Fuzzy set theory is used to capture the vagueness of the workload during redistribution period. According to the experts’ subjective evaluations, a fuzzy inference system is established to aggregate the fuzzy web performance metrics into a so-called load-weight index which indicates the servers workload intensity. Based on the load-weight index, the genetic-fuzzy algorithm is applied to equally redistribute the workload among in the servers. Finally, a simulation of 20 load-weight indices in a topology of 3-cube form Web cluster is implemented to illustrate the functionality of the proposed method.