«上一篇/Previous Article|本期目录/Table of Contents|下一篇/Next Article»

《计算机技术与发展》[ISSN:1006-6977/CN:61-1281/TN]

卷:

30

期数:

2020年04期

页码:

63-68

栏目:

智能、算法、系统工程

出版日期:

2020-04-10

文章信息/Info

Title:

Modeling of Health Evaluation of Container Cloud System Based on Rough Set

文章编号:

1673-629X(2020)04-0063-06

作者:

张可颖¹ ; 龙士工²; 3 ; 吕尚青⁴ ; 吕晓丹³

1. 贵州大学 大数据与信息工程学院,贵州 贵阳 550025; 2. 贵州大学 贵州省公共大数据重点实验室,贵州 贵阳 550025; 3. 贵州大学 计算机科学与技术学院,贵州 贵阳 550025; 4. 北京邮电大学 信息与通信工程学院,北京 100000

Author(s):

ZHANG Ke-ying¹ ; LONG Shi-gong²; 3 ; LYU Shang-qing⁴ ; LYU Xiao-dan³

1. School of Big Data and Information Engineering,Guizhou University,Guiyang 550025,China; 2. Guizhou Provincial Key Laboratory of Public Big Data,Guizhou University,Guiyang 550025,China; 3. School of Computer Science and Technology,Guizhou University,Gui

关键词:

系统健康度; 粗糙集; 信息熵; 云平台

Keywords:

system health; rough set; information entropy; cloud platform

分类号:

TP319

DOI:

10. 3969 / j. issn. 1673-629X. 2020. 04. 012

摘要:

现在容器云平台容器数目日益增加,相关监控数据爆炸式增长,而现有的运行在容器内的微服务监控软件监控指标不仅种类繁多,配置繁琐,并且往往只是直接给出监控数据,没有根据得到的监控指标对系统的健康度进行度量。 针对该问题,提出了一种新的基于粗糙集的容器云系统健康度评价模型。 通过建立的粗糙集云系统健康度评价模型,可以直观地反映整个集群的健康程度。 首先通过信息熵对监控到的连续属性进行断点分割,离散化处理,然后利用粗糙集理论实现对监控数据进行知识约简、一致性检查和决策表建立,从而建立了基于粗糙集和信息熵的集群健康度指标模型。 最后,通过 Kubernetes 容器云平台分别进行计算密集负载和网络密集负载仿真实验,实验结果表明,该模型能够反映集群的性能和对异常进行检测。

Abstract:

Now the number of containers of container cloud platform is increasing, and the related monitoring data is exploding. The existing monitoring indicators of micro-service monitoring software running in the container are not only diverse and complicated to configure,but also often just give monitoring data directly without measuring the health of the system based on the monitoring indicators obtained. For this,a new rough set based container cloud system health evaluation model is proposed,by which the health of the entire cluster can be intuitively reflected. Firstly, through the information entropy, the monitored continuous attributes are segmented and discretized. And then the knowledge reduction,consistency check and decision table establishment of monitoring data are completed by using rough set theory,so as to establish a cluster health indicator model based on rough set and information entropy. Finally,the computational intensive load and network intensive load simulation experiments are carried out through the Kubernetes container cloud platform,which show that this model can reflect the performance of the cluster and detect the anomaly.

常用功能

更新日期/Last Update: 2020-04-10