摘要
云平台多处理器的任务调度是解决庞大用户群中庞大任务量和数据量的关键,云平台中任务调度的计算性能影响整个系统的运行效率。提出一种基于任务信息流特征尺度谱分析的开销折减算法,采用可分解特征下的云平台任务同步开销折减算法,通过构建复杂通道下多处理器运行环境下的云平台任务调度整合基础模型,进行任务节点信息表征,使用GSM、TD-SCDMA、TD-LTE和WLAN,实现多处理器和多通道任务调度,计算各任务匹配资源的效率,得到资源相似度,基于任务信息流特征尺度谱分析方法计算每个任务的特征尺度,得到尺度优化的开销折减目标函数。仿真结果表明,采用该算法进行任务调度,具有较高的执行效率,CPU利用率高,网络开销折减幅度较高,提高了数据通信效率。
Multi-processors task scheduling in cloud platform is the key to solve the huge user group in the huge task and data quantity, it influences the running efficiency of the whole system and computing performance of task scheduling in cloud platform. An overhead reduction algorithm based on task information flow characteristic spectrum analysis is pro-posed. And the task scheduling integration synchronization algorithm of the cloud platform multiprocessor is proposed based on feature scale decomposition, The decomposition characteristics of the cloud platform task integration and synchro-nization algorithm, through the cloud platform task scheduling environment construction of complex multi processor chan-nel integration under the base model, task node information representation, using GSM, TD-SCDMA, TD-LTE and WLAN, realizes the multi processor and multi channel scheduling, each task efficiency calculation of matching resources, to obtain the resource similarity, the characteristic scale task information flow characteristic scale is calculated for each task analysis method based on spectrum, obtained the scale optimization overhead reduction target function. The simulation results show that, it has high efficiency and high utilization rate of CPU, the network overhead is less, the overall performance is better than the traditional algorithm.
出处
《科技通报》
北大核心
2014年第10期52-54,共3页
Bulletin of Science and Technology
关键词
任务调度
云平台
特征尺度分解
task scheduling
cloud platform
characteristic scale decomposition
