APP下载

大数据网络的均衡调度平台设计与改进

2016-05-14蒋本立张小平

现代电子技术 2016年6期

蒋本立 张小平

摘 要: 当前的数据调度平台以数据完成时间实现网络资源的调度,将任务完成时间当成网络均衡调度的基础,未考虑大数据网络的能耗指标,无法实现真正的网络均衡调度。提出基于能量评判的大数据网络均衡调度平台,平台由服务模块、控制模块、虚拟化模块和物理资源模块组成,包括资源定义模块、资源监测模块以及资源调度模块三大功能模块。设计一种基于能量评判的网络资源调度模型,在一定的时间区间内,以能量消耗最佳为约束条件,全面分析大数据网络资源调度中的能量评判问题,给出调度平台采集数据的关键代码。实验结果表明,所设计平台的平均能耗率、吞吐率以及CPU利用率指标都较优,具有较高的负载均衡效果。

关键词: 大数据网络; 均衡调度平台; 网络资源调度; 能量评判

中图分类号: TN915?34; TP315 文献标识码: A 文章编号: 1004?373X(2016)06?0062?04

Design and improvement of balanced scheduling platform for big data network

JIANG Benli1, ZHANG Xiaoping2

(1. Department of Electronics and Information Engineering, Hunan Vocational College of Science & Technology, Changsha 410004, China;

2. Information School, Hunan University of Science & Technology, Xiangtan 411201, China)

Abstract: The current data scheduling platform realizes the network resource scheduling according to data completion time, and takes the task completion time as the basis of balanced scheduling network, in which the energy consumption index of big data network is not considered. That is why it can realize the real network balanced scheduling. Therefore, a big data network equilibrium scheduling platform based on energy evaluation is put forward in this paper. The platform is composed of the service module, control module, virtualization module and the physical resource module, including the resource definition module, resource monitoring module and resource scheduling module. A network resource scheduling model based on energy evaluation was designed, which takes optimal energy consumption in a certain time interval as its constraint condition. The energy evaluation in the big data network resource scheduling is analyzed comprehensively. The key code for data acquisition of the scheduling platform is given. Experiment result shows that the average energy consumption rate, throughput and CPU utilization rate of the designed platform are better, and it has the high load balancing effect.

Keywords: big data network; balanced scheduling platform; network resource scheduling; energy evaluation

随着网络通信技术的不断发展,网络中的数据规模逐渐增加,使得大数据网络的均衡调度问题日益突出[1?3]。当前的大数据网络存在资源使用率低、网络负载失衡的缺陷。因此,针对大数据网络资源环境下的资源调度问题,寻求高效调度策略,对于增强大数据网络资源利用率以及负载均衡具有重要作用。大数据网络的能量是有限的,网络中大规模数据的调度会消耗大量的能量,导致资源调度时间增加。传统的资源调度算法,通常基于数据完成时间实现网络资源的调度,将任务完成时间当成调度目标,未对大数据网络的能耗进行综合分析,无法真正实现网络均衡调度[4?5]。……

登录APP查看全文