CAJ | 학술논문

随着计算机技术的发展与信息化的普及,大规模数据处理已经成为计算机产业界的焦点问题和学术界的研究热点.Hadoop是云计算的分布式开源的实现平台,在大规模数据处理方面有着高容量、低成本等优势,也是一个运行在大型集群之上的并行处理系统.Hadoop云平台上的调度算法是对用户作业进行资源分配和调度执行的程序,它的选择合适与否直接关系到用户作业的执行效率.因此,对Hadoop上作业调度算法的研究和改进优化就显得极为重要.本文通过实验验证作业重调度具有更好的响应时间和吞吐量,具有良好的负载均衡,并且大大提高了Hadoop系统在异构平台上的性能.
수착계산궤기술적발전여신식화적보급,대규모수거처리이경성위계산궤산업계적초점문제화학술계적연구열점.Hadoop시운계산적분포식개원적실현평태,재대규모수거처리방면유착고용량、저성본등우세,야시일개운행재대형집군지상적병행처리계통.Hadoop운평태상적조도산법시대용호작업진행자원분배화조도집행적정서,타적선택합괄여부직접관계도용호작업적집행효솔.인차,대Hadoop상작업조도산법적연구화개진우화취현득겁위중요.본문통과실험험증작업중조도구유경호적향응시간화탄토량,구유량호적부재균형,병차대대제고료Hadoop계통재이구평태상적성능.
With the development of computer technology and the popularization of information technology,large-scale data processing has become the focus of the computer industry and academia.Hadoop is a distributed open source platform for cloud computing.It has advantages of high capacity and low cost in large scale data pro-cessing.It is also a parallel processing system running on a large cluster.Scheduling algorithm on the Hadoop cloud platform is a program that performs resource allocation and scheduling for user tasks,and it is directly relat-ed to the execution efficiency of users.So it is very important to study and improve the scheduling algorithm of Hadoop.In this paper,we show that the job scheduling has better response time and throughput,and has a good load balance,which greatly improves the performance of Hadoop system on heterogeneous platforms.