云环境下基于通信开销的并行任务调度策略
|
摘要
云环境下,任务的执行效率受限于任务间的通信时间和计算时间,通信时间是由于数据跨数据中心传输产生的,计算时间与任务所在集群的计算能力有关,有效减少任务因等待数据的到来而产生的时间开销可提高任务的执行效率,进而降低用户租赁云资源的费用,提出基于任务复制的调度策略,以提高任务的并行性。经性能分析,该策略在提高任务的执行效率方面有一定的贡献。
In the cloud environment, the execution efficiency of tasks is limited by the communication time and computing time between tasks. Communication time is caused by data transmission across data centers. The computing time is related to the computing power of the cluster where the task is located. Effectively reduce the time overhead incurred by the task waiting for the arrival of data, which can improve the execution efficiency of the task. In turn, it can reduce the cost for users to rent cloud resources. A scheduling strategy based on task replication is proposed to improve the parallelism of tasks. Through performance analysis, this strategy has a certain contribution to improve the efficiency of task execution.
In the cloud environment, the execution efficiency of tasks is limited by the communication time and computing time between tasks. Communication time is caused by data transmission across data centers. The computing time is related to the computing power of the cluster where the task is located. Effectively reduce the time overhead incurred by the task waiting for the arrival of data, which can improve the execution efficiency of the task. In turn, it can reduce the cost for users to rent cloud resources. A scheduling strategy based on task replication is proposed to improve the parallelism of tasks. Through performance analysis, this strategy has a certain contribution to improve the efficiency of task execution.
引文
[1]CHEN W,ALTINTAS I,WANG J,et al.Enhancing smart rerun of Kepler scientific workflows based on near optimum provenance caching in cloud[C]//Services(SERVICES),2014IEEE World Congress on.IEEE,2014:378-384.
[2]段菊,陈旺虎,王润平,et al.云环境下基于聚簇的科学工作流执行优化策略[J].计算机应用,2015,35(6):1580-1584.
[3]陈旺虎,段菊,俞茂义.允许违反局部时间约束的科学工作流调度策略[J].计算机工程与科学,2016(11):2165-2171.
[4]段菊,于治国.云环境下基于相关性的并行任务调度策略[J].计算机技术与发展,2018,28(06):178-183.
[5]喻莉,阮文涛.负载均衡技术的研究与实现[J].计算机技术与发展,2007,17(8):120-122.
[6]刘少伟,孔令梅,任开军,等.云环境下优化科学工作流执行性能的两阶段数据放置与任务调度策略[J].计算机学报2011,34(11):002121-2130.
[7]贾丽云,张向利,张红梅.分布式系统下的启发式任务调度算法[J].计算机工程与应用,2017,53(12):63-69.
[8]周超群,周亦敏.一种改进的基于复制的异构多核任务调度算法[J].电子科技,2017(06):63-68.
[9]何琨,赵勇,黄文奇.基于任务复制的分簇与调度算法[J].计算机学报,2008,31(5):733-740.
[10]PARK C I,CHOE T Y.An optimal scheduling algorithm based on task duplication[C]//Parallel and Distributed Systems,2001.ICPADS 2001.Proceedings.Eighth International Conference on.IEEE,2001:9-14.
[11]李静梅,尤晓非,韩启龙.基于任务复制的多关键路径任务调度算法[J].计算机工程与设计,2014,35(5):1639-1645.
[2]段菊,陈旺虎,王润平,et al.云环境下基于聚簇的科学工作流执行优化策略[J].计算机应用,2015,35(6):1580-1584.
[3]陈旺虎,段菊,俞茂义.允许违反局部时间约束的科学工作流调度策略[J].计算机工程与科学,2016(11):2165-2171.
[4]段菊,于治国.云环境下基于相关性的并行任务调度策略[J].计算机技术与发展,2018,28(06):178-183.
[5]喻莉,阮文涛.负载均衡技术的研究与实现[J].计算机技术与发展,2007,17(8):120-122.
[6]刘少伟,孔令梅,任开军,等.云环境下优化科学工作流执行性能的两阶段数据放置与任务调度策略[J].计算机学报2011,34(11):002121-2130.
[7]贾丽云,张向利,张红梅.分布式系统下的启发式任务调度算法[J].计算机工程与应用,2017,53(12):63-69.
[8]周超群,周亦敏.一种改进的基于复制的异构多核任务调度算法[J].电子科技,2017(06):63-68.
[9]何琨,赵勇,黄文奇.基于任务复制的分簇与调度算法[J].计算机学报,2008,31(5):733-740.
[10]PARK C I,CHOE T Y.An optimal scheduling algorithm based on task duplication[C]//Parallel and Distributed Systems,2001.ICPADS 2001.Proceedings.Eighth International Conference on.IEEE,2001:9-14.
[11]李静梅,尤晓非,韩启龙.基于任务复制的多关键路径任务调度算法[J].计算机工程与设计,2014,35(5):1639-1645.