摘要
云计算环境中的集群资源性能各不相同,为了避免在资源调度时随机调度资源而造成任务执行低效,提出了一种基于模糊聚类的集群资源调度方法;首先,给出了集群资源的综合评价模型,模型中综合考虑性能指标、服务能力和服务质量等指标,然后将任务和资源组合成混合矩阵进行数据格式化并构造出对应的模糊相似度矩阵进行聚类,最后,定义了基于模糊聚类和云计算的集群资源调度具体算法;采用CloudSim云计算仿真工具对实例进行仿真,当φ为0.825时,F值最大为7.14,此时集群资源可以分为三类,实验结果表明文中方法能高效地实现云计算环境下的集群资源调度,且与其它方法比较,具有较低的时间复杂度。
The performance of computers in computer cluster are different in cloud computing, so in order to avoid the ineffective execu- ting for task resulting from random scheduling the computer for task, a method for computer cluster scheduling based on fuzzy clustering was proposed. Firstly, the comprehensive model for computer cluster was given considering the performance indictors, the service capability and service quality etc, then the task and computer vector were grouped to the compound matrix to be formatted and then the relating fuzzy simi- larity matrix was constructed. Finally, the specific algorism for computer cluster scheduling based on fuzzy clustering was defined. Using the CloudSim as the simulation tool, when the value of 90 is 0. 825, the max value of F is 7.14, and the computer cluster can be divided to three categories. The experiment result shows the method in this paper can realize scheduling for computer cluster effectively, compared with the other methods, it has the lower time complexity.
出处
《计算机测量与控制》
北大核心
2013年第9期2529-2531,共3页
Computer Measurement &Control
基金
嵌入式系统虚拟化技术研究与应用(XY12KJ09)
关键词
云计算
调度
集群资源
聚类
cloud computing
scheduling
computer cluster
clustering
