通信学报 ›› 2019, Vol. 40 ›› Issue (8): 143-150.doi: 10.11959/j.issn.1000-436x.2019172

• 学术论文 • 上一篇    下一篇

基于三次指数平滑法和时间卷积网络的云资源预测模型

谢晓兰1,2,张征征2(),王建伟2,程晓春3   

  1. 1 桂林理工大学广西嵌入式技术与智能系统重点实验室,广西 桂林 541004
    2 桂林理工大学信息科学与工程学院,广西 桂林 541004
    3 密德萨斯大学计算机科学系,伦敦 NW4 4BT
  • 修回日期:2019-07-13 出版日期:2019-08-25 发布日期:2019-08-30
  • 作者简介:谢晓兰(1974- ),女,广西桂林人,博士,桂林理工大学教授、博士生导师,主要研究方向为云计算、并行计算、大数据、地球物理勘查与信息技术。|张征征(1994- ),女,山东临沂人,桂林理工大学硕士生,主要研究方向为云计算、大数据。|王建伟(1993- ),男,河南周口人,桂林理工大学硕士生,主要研究方向为云计算、大数据。|程晓春(1973- ),男,吉林长春人,博士,密德萨斯大学计算机研究项目管理员,主要研究方向为智能计算、通信数据分析管理、通信安全。
  • 基金资助:
    国家自然科学基金资助项目(61762031);广西创新驱动重大专项基金资助项目(2018AA32003);广西重点研发计划基金资助项目(AB17195029);广西重点研发计划基金资助项目(AB18126006);广西硕士研究生创新基金资助项目(YCSW2017156);广西硕士研究生创新基金资助项目(YCSW2018157);广西中青年教师基础能力提升基金资助项目(KY2016YB184)

Cloud resource prediction model based on triple exponential smoothing method and temporal convolutional network

Xiaolan XIE1,2,Zhengzheng ZHANG2(),Jianwei WANG2,Xiaochun GHENG3   

  1. 1 Guangxi Key Laboratory of Embedded Technology and Intelligent System,Guilin University of Technology,Guilin 541004,China
    2 College of Information Science and Engineering,Guilin University of Technology,Guilin 541004,China
    3 Department of Computer Science,Middlesex University,London NW4 4BT,UK
  • Revised:2019-07-13 Online:2019-08-25 Published:2019-08-30
  • Supported by:
    The National Natural Science Foundation of China(61762031);The Science and Technology Major Project of Guangxi(2018AA32003);The Key Research and Development Program of Guangxi(AB17195029);The Key Research and Development Program of Guangxi(AB18126006);Innovation Project of Guangxi Graduate Education(YCSW2017156);Innovation Project of Guangxi Graduate Education(YCSW2018157);Subsidies for the Project of Promoting the Ability of Young and Middle-Aged Scientific Research in Universities and Colleges of Guangxi(KY2016YB184)

摘要:

以Docker和Kubernetes为代表的容器云具有额外的资源开销更小、启动销毁时间更短等优点,但它仍然存在过度供应和供应不足等资源管理问题。为了使Kubernetes集群对部署在其上的应用资源使用量能“提前”响应,并根据预测值为应用及时、准确、动态地调度和分配资源,提出了一种基于三次指数平滑法和时间卷积网络的云资源预测模型,根据历史数据预测未来的资源需求。为了找到参数的最优组合,使用TPOT调参思想对参数进行优化。对Google数据集CPU和内存的预测实验表明,所提模型与其他模型相比具有更好的预测性能。

关键词: 资源预测, Kubernetes, 指数平滑法, 时间卷积网络

Abstract:

The container cloud represented by Docker and Kubernetes has the advantages of less additional resource overhead and shorter start-up and destruction time.However there are still resource management issues such as over-supply and under-supply.In order to allow the Kubernetes cluster to respond “in advance” to the resource usage of the applications deployed on it,and then to schedule and allocate resources in a timely,accurate and dynamic manner based on the predicted value,a cloud resource prediction model based on triple exponential smoothing method and temporal convolutional network was proposed,based on historical data to predict future demand for resources.To find the optimal combination of parameters,the parameters were optimized using TPOT thought.Experiments on the CPU and memory of the Google dataset show that the model has better prediction performance than other models.

Key words: resource prediction, Kubernetes, exponential smoothing method, temporal convolutional network

中图分类号: 

No Suggested Reading articles found!