File merging strategy for optimization of distributed file system
Fagen Gong · Journal of Computer Applications · 2011
The performance of distributed file system has an important effect on the performance of the whole distributed system.This paper took Hadoop Distributed File System(HDFS) as the research object.In order to solve the poor performance problem of HDFS in term of processing small files,this paper analyzed the existing problem and presented a novel file merging strategy to optimize system I/O performance.The experiment results show this merging strategy can effectively improve the performance of distributed file system.