一种基于访问趋势的数据网格动态副本创建策略
发布时间：2019-06-27 17:53
【摘要】：当前的副本创建策略在选取热点数据时往往没有对访问突变做出反应,造成副本频繁创建和删除的网格动荡.针对上述情况,本文引入文件访问趋势的概念,提出一种基于访问趋势的数据网格动态副本创建策略(ADRS),ADRS根据历史数据访问情况计算出文件的访问趋势,然后根据文件的访问趋势决定数据副本的创建和删除.仿真结果表明,在随机访问模式下,对比最近访问最大权重算法(LALW),ADRS能有效地降低数据副本的创建次数,提高网格副本利用率并且在节点空间有限的情况下不会影响网格的效率.
[Abstract]:The current replica creation strategy often does not respond to the access mutation when selecting hot data, resulting in grid turbulence where copies are created and deleted frequently. In view of the above situation, this paper introduces the concept of file access trend, and proposes a dynamic replica creation strategy for data grid based on access trend. (ADRS), ADRS calculates the access trend of the file according to the historical data access situation, and then determines the creation and deletion of the data copy according to the access trend of the file. The simulation results show that compared with the recent access maximum weight algorithm (LALW), ADRS, the simulation results show that the creation times of data replicas can be effectively reduced, the utilization of grid replicas can be improved, and the grid efficiency will not be affected when the node space is limited.
【作者单位】： 四川大学计算机学院;攀枝花学院数学与计算机学院;
【基金】：教育部重大项目培育基金(708075) 国家自然科学基金(61173159)
【分类号】：TP393.01
[Abstract]:The current replica creation strategy often does not respond to the access mutation when selecting hot data, resulting in grid turbulence where copies are created and deleted frequently. In view of the above situation, this paper introduces the concept of file access trend, and proposes a dynamic replica creation strategy for data grid based on access trend. (ADRS), ADRS calculates the access trend of the file according to the historical data access situation, and then determines the creation and deletion of the data copy according to the access trend of the file. The simulation results show that compared with the recent access maximum weight algorithm (LALW), ADRS, the simulation results show that the creation times of data replicas can be effectively reduced, the utilization of grid replicas can be improved, and the grid efficiency will not be affected when the node space is limited.
【作者单位】： 四川大学计算机学院;攀枝花学院数学与计算机学院;
【基金】：教育部重大项目培育基金(708075) 国家自然科学基金(61173159)
【分类号】：TP393.01
【参考文献】
相关期刊论文 前2条
1 亓雪冬;李霞;梁鸿;邢常振;仝兆岐;;一种新型的数据网格动态副本策略(英文)[J];北京交通大学学报;2008年06期
2 孙海燕,王晓东,周斌,贾焰,王怀民,邹鹏;基于存储联盟的双层动态副本创建策略—SADDRES[J];电子学报;2005年07期
【共引文献】
相关期刊论文 前10条
1 蒋砚军;马华东;张海e，
本文编号：2507010
本文链接：https://www.wllwen.com/guanlilunwen/ydhl/2507010.html
