A Hadoop load balancing method on heterogeneous clusters includes the steps of: evaluating the performance of servers based on the number of central processing unit (CPU) cores and the size of an available memory in each server; determining the number of Map slots representing the maximum number of Map Tasks, which are able in each server, in proportion to the performance of each server; performing the Map Task allocated to each server according to the determined number of the Map slots; and determining a server to execute a Reduce Task according to the priority resulting from the performance of the servers. Accordingly, the load balancing is performed based on the performance of the servers, thereby improving the performance to execute Hadoop Mapreduce.;COPYRIGHT KIPO 2016
展开▼