电信科学 ›› 2015, Vol. 31 ›› Issue (3): 142-147.doi: 10.11959/j.issn.1000-0801.2015084

• 电力信息化专栏 • 上一篇    下一篇

基于数据挖掘的电力云资源规划调度

庞松涛   

  1. 浪潮集团 北京 100142
  • 出版日期:2015-03-15 发布日期:2017-02-23

Resource Scheduling for Cloud Data Center Based on Data Mining in Smart Grid

Songtao Peng   

  1. Inspur Group Co., Ltd., Beijing 100142, China
  • Online:2015-03-15 Published:2017-02-23

摘要:

随着虚拟化技术的应用,数据中心的资源利用率已经得到一定程度的提高,但是云资源通常还是根据用户提出的需求预先分配,资源利用率仍然有待提高。为了进一步改进资源利用率,云中心的实际资源需求可分为“周期资源”及“峰期资源”。“周期资源”可以通过历史规律,利用递阶成分负载模型进行分析,预测出数据中心大部分时间的资源需求;“峰期资源”主要是满足短时间内的各种高峰资源需求,这种需求应用了排队论及随机均衡算法模型并根据实际应用的需要动态启用、分配和收回资源。通过使用资源收集与分配守护进程对云中心资源需求进行规划调度实验,效果明显,从而为云中心资源利用率的提升及节能减排提供了一种有效途径。

关键词: 数据挖掘, 智慧电力云, 云资源调度

Abstract:

Since the wide use of virtual technology,the resource use rate in cloud data center has been improved effectively than ever before.However,there is still a large space for improvement due to the resources usually are pre-started and pre-allocated by the user demand rather than the actual needs.In order to allocate available resource more accurately,two algorithms were proposed to meet the needs of the daily use in most of time.The available virtual resources would be arranged according the forecast using the algorithms of hierarchical composition of loading and the peek resources needs would be dynamic allocated using the algorithms of stochastic equilibrium and queuing theory.The results of experiment via the system based upon above theories show that the solution provides a kind of very effective advanced means for the optimal use of resources and energy saves.

Key words: data mining, smart grid cloud, cloud resource scheduling

No Suggested Reading articles found!