Load balancing solution based on AHP for Hadoop
Huixiang Zhou, Qiaoyan Wen · 2014
Apache Hadoop is an open-source software framework for cloud computing, the server load balancing algorithm named as Balancer which Hadoop provided can load balance for each DataNode, but Balancer algorithm only consider the storage space factor, in this case it will easily lead to load imbalance while operation of the system for a long time. To solve this problem, we consider the file number of visited, frequency of concurrent access, bandwidth, CPU power, memory utilization and other factors, we propose a load balancing solution combined APH, our solution can more accurately calculate the load capacity for each DataNode for Hadoop, and it can automatically adjust the load capacity of Hadoop distributed system while user requests a file again.