北京邮电大学学报

  • EI核心期刊

北京邮电大学学报 ›› 2015, Vol. 38 ›› Issue (2): 78-82.doi: 10.13190/j.jbupt.2015.02.014

• 论文 • 上一篇    下一篇

SDN架构下数据中心内存能耗优化策略

彭宏玉1,2, 陈刚1,2, 张英海1,2, 王朝炜1,2   

  1. 1. 北京邮电大学 电子工程学院, 北京 100876;
    2. 北京邮电大学 泛网无线通信教育部重点实验室, 北京 100876
  • 收稿日期:2014-11-03 出版日期:2015-04-28 发布日期:2015-05-14
  • 作者简介:彭宏玉(1977—), 男, 博士生, E-mail: penghongyu@bupt.edu.cn; 张英海(1951—), 男, 教授, 博士生导师.
  • 基金资助:

    北京市教委共建项目

A New Data Center Memory Energy Consumption Optimization Stragegy in SDN

PENG Hong-yu1,2, CHEN Gang1,2, ZHANG Ying-hai1,2, WANG Chao-wei1,2   

  1. 1. College of Electronic Engieering, Beijing University of Posts and Telecommunications, Beijing 100876, China;
    2. Key Laboratory of Universal Wireless Communications, Ministry of Education, Beijing University of Posts and Telecommunications, Beijing 100876, China
  • Received:2014-11-03 Online:2015-04-28 Published:2015-05-14

摘要:

提出一种基于内存共享机制数据中心进程优化调度策略,并基于软件定义网络 (SDN)架构建立数据中心负载模型及内存节点能耗模型,给出2个用于降低内存能耗的启发式进程调度算法. 这2个算法以实际内存节点为研究对象,在保证数据中心负载均衡的基础上,通过合理进程调度实现内存共享率最大,使得处于活动态的内存节点数最少,从而达到数据中心内存能耗优化目的. 仿真结果表明,提出的启发式进程调度算法能有效降低内存能耗.

关键词: 内存能耗, 软件定义网络, 进程调度算法

Abstract:

A new heuristic process scheduling strategy was proposed on the basis of a memory sharing mechanism. Based on software defined networking (SDN) architecture, A workload model and a memory energy consumption model were designed. Two heuristic process scheduling algorithms were designed to save the memory energy consumption. The algorithms come from physical memory nodes. The key idea of the above proposed algorithms is to achieve the largest sharing ratio of memory which assures workload balance through scheduling processes reasonably. The least number of active memory nodes means the least memory energy consumption of data center. Simulations verify that these strategies can reduce the energy consumption effectively.

Key words: memory energy consumption, software defined networking, process sheduling algorithm

中图分类号: