期刊名称：International Journal of Computer Trends and Technology
电子版ISSN：2231-2803
出版年度：2018
卷号：61
期号：1
DOI：10.14445/22312803/IJCTT-V61P104
出版社：Seventh Sense Research Group
摘要：Cloud Computing provides different services to the users with regard to processing data. The main concepts in cloud computing are big data and big data analysis. Hadoop framework is used to process big data in parallel processing mode. Job scheduling and optimized resource allocation can help improve performance of Hadoop. In the existing system Hadoop architecture has been enhanced in order to reduce computational complexity while processing big data. It also takes care of efficient resource allocation and processing textual data such as DNA sequence. Their architecture was named as H2Hadoop that improves the ability of NameNode to assign jobs to the TaskTrackers (DataNodes) in a given cluster. By adding control features to NameNode, their architecture can intelligently assign tasks to the DataNodes where required data is present thus reducing resource utilization pertaining to CPU time, number of read operations etc. However, the existing system can be improved to have more focused approach by considering data locality awareness to the job scheduling process. In the proposed system, an algorithm is proposed to have data locality aware job scheduling. This algorithm is named as Data Locality Aware Job Scheduling (DLAJS) algorithm. The algorithm explores the data locality aware to know how far efficient job scheduling. Thus, consuming less cloud resources such as CPU, memory and execution time.
关键词：Cloud computing; Big data; Hadoop; MapReduce framework; Data-locality; Job scheduling
Loading...
