An Efficient Server Minimization Algorithm for Internet Distributed Systems

The increasing use of online services leads to an unequal distribution of the loads among the servers. As a result, the problem is to balance the loads among the servers such that the total number of active servers is minimized. One of the possible solutions is to transfer the loads from the underutilized server to a suitable server and make the underutilized server to sleep mode. In this paper, a server minimization algorithm (SMA) is proposed for the solution of server minimization and the load balancing problem. The proposed algorithm reduces the number of servers by merging the loads of the two least loaded servers. Then it determines the standard deviation of the server loads for load balancing. The proposed SMA is compared with an existing load balancing algorithm using the number of minimized servers, load standard deviation and load factor. The simulation results show the efficacy of the SMA.