The Evaluation of the Improved Redundant Power Consumption Laxity-Based (IRPCLB) Algorithm in Homogeneous and Heterogeneous Clusters

In a server cluster, one server is usually selected to perform a request process from a client. Once the server stops by fault, the client is suspended to wait for a reply. Even if the request is performed on another server on detection of fault of the server, some QoS requirement like response time may not be satisfied. Hence, each request is redundantly performed on multiple servers to be tolerant of server faults. Here, more number of servers a request process is redundantly performed, the more reliable but the more amount of electric power is consumed. Thus, it is critical to discuss how to realize a reliable and energy-aware server cluster in presence of server faults. The redundant power consumption laxity-based (RPCLB) algorithm is proposed to redundantly and energy-efficiently perform a request process in our previous studies. In this paper, we newly discuss an improved RPCLB (IRPCLB) algorithm where once a process successfully terminates on one server, meaningless redundant processes are forced to terminate on the other servers. We show the total power consumption of servers and response time of each process can be reduced in the IRPCLB algorithm than the RPCLB and round-robin (RR) algorithms in both heterogeneous and homogeneous clusters.