期刊名称：International Journal of Computer Trends and Technology
电子版ISSN：2231-2803
出版年度：2014
卷号：16
期号：3
页码：90-92
出版社：Seventh Sense Research Group
摘要：Distributed file systems in clouds such as GFS(Google File System) and HDFS(Hadoop Distributed file systems) rely on central servers to manage the metadata and the load balancing. DFS are keys building block for cloud computing application based on the reduce programming. In this file system node at the same time provide computing and storage functions, a file dived into a number of parts allocated in specifically nodes therefore Details Reduce functions can be implemented in parallel to all the nodes. In this cloud computing environment, failure standard and the nodes may be updated, replaced and added in the system. Files is also dynamically append, create, and delete. These results in load imbalance in a distributed file system, those files parts are not distributed in the same way as much as possible between the nodes developing distributed file systems in production systems tight are based on the central node for the parts transfer. This depended is apparently inaccurate in a largescale failurechances environment because there will be central load is put under a significant level workload that is simple scaled with system size, and may thus become the management deterrent and the single point of failure. In this paper, a fully supplied load rebalancing algorithm is displayed to cope with the load imbalance problem. In proposed system sending the files divided different parts of files using load rebalancing and DES algorithm after stored into cloud.
关键词：load balancing ; Distributed file systems ; clouds
