计算机科学 ›› 2011, Vol. 38 ›› Issue (Z10): 157-160.

• CRSSC-CWI-CGrC2015 • 上一篇    下一篇

云计算中的集群资源模糊聚类划分模型

刘伯成,陈庆奎   

  1. (上海理工大学光电信息与计算机工程学院 上海200093)
  • 出版日期:2018-11-16 发布日期:2018-11-16
  • 基金资助:
    本文受上海信息技术领域重点科技攻关(09517507000),上海重点科技项目(09220502800)资助。

Fuzzy Clustering Partition Model for Computer Cluster in Cloud Computing

LIU Bo-cheng,CHEN Qing-kui   

  • Online:2018-11-16 Published:2018-11-16

摘要: 随着云计算应用的开展,计算机集群的作用越来越重要,但集群中计算机的性能良芬不齐,虽然能互联且能共享资源,但是很有可能因为某些集群内的计算机的性能不均衡或集群性能与并行任务资源需求不匹配而造成任务低效执行的后果。如何把物理集群(普通局域网互联的计算机构成)分为若千个性能均衡的逻辑集群是集群调度的关键。通过对计算机资源的模糊聚类来划分集群中的计算机,引进任务资源需求向量和最低误差容忍向量机制,把物理集群划分为若千个性能均衡或与并行任务资源匹配的逻辑计算机集群,使集群更易管理调度。对物联云运用此算法划分了网关集群、数据库集群和服务集群,验证了本算法,这种划分方法适合云计算应用。

关键词: 云计算,计算机集群,资源,模糊聚类

Abstract: With the development of cloud computing applications, the computer cluster becomes more and more important, but the computer performance in cluster is so different. Even though they can connect and can share their resources each other, but when the computer performance is not balance and the cluster performance is not to match with the parallcl tasks, it can cause the low efficiency for parallel task. So, it is an important problem how to partition and management computer cluster. The paper gave a way to solve it. Using the fuzzy clustering method according to the computer resources in computer cluster and through the vector of resource demand and the vector of lowest inaccuracy tolerance,we can divide the computer cluster into several classes(logical computer cluster) and make the every computer performance in one class to be more similar. It makes the computer cluster management to become easier. We partition the computers in lab into network gateway cluster, data base cluster, and service cluster for the Internet of thing cloud using this model.This model is fit for the cloud computing applications.

Key words: Cloud computing,Computer cluster,Resource,Fuzzy clustering

No related articles found!
Viewed
Full text


Abstract

Cited

  Shared   
  Discussed   
No Suggested Reading articles found!