摘要
基于云任务的动态延迟到达,到达时间具有随机性,进行任务分配到虚拟机执行的仿真模拟。仿真任务是基于排队论的指数分布到达方式,虚拟机的处理速度存在差异。实验思路是高优先级尽可能早的执行,同时避免低优先级任务饿死的思路,并定义了一个消费量作为对比点,对比基于一定等待的调度与贪心调度的差异。通过设置动态等待时间的设置来进行实验,同时测试多组输入数据减少样本的偶然性。对比分析各种数据,验证实验的有效性。并通过CloudSim进行云仿真实验模拟。
Dynamic delay arrival based on cloud tasks, random arrival time, simulation of task assignment to virtual machine execution. The simulation task is based on the arrival of the exponential distribution of the queuing theory, and the processing speed of the virtual machine is different. The experimental idea is to execute the high priority as early as possible, while avoiding the idea of starvation of low priority tasks, and define a consumption as a comparison point, comparing the difference between scheduling based on certain waiting and greedy scheduling. Experiments are performed by setting the dynamic wait time setting, while testing multiple sets of input data reduces the contingency of the sample. Compare and analyze various data to verify the effectiveness of the experiment. And cloud simulation experiment simulation through CloudSim.
作者
花嵘
张友裕
HUA Rong;ZHANG You-yu(College of Computer Science and Engineering, Shandong University of Science and Technology, Qingdao 266510, China)
出处
《软件》
2019年第10期139-141,190,共4页
Software
关键词
云计算
任务调度
延迟任务
Cloud computing
Task scheduling
Delayed task
