计算机应用 ›› 2016, Vol. 36 ›› Issue (7): 1784-1788.DOI: 10.11772/j.issn.1001-9081.2016.07.1784

• 网络与通信 • 上一篇    下一篇

软件定义网络架构下基于流调度代价的数据中心网络拥塞控制路由算法

孙三山1,2, 汪帅3, 樊自甫3   

  1. 1. 通信抗干扰技术国家级重点实验室(电子科技大学), 成都 611731;
    2. 四川师范大学 物理与电子工程学院, 成都 610101;
    3. 重庆邮电大学 通信与信息工程学院, 重庆 400065
  • 收稿日期:2015-12-24 修回日期:2016-04-06 出版日期:2016-07-10 发布日期:2016-07-14
  • 通讯作者: 孙三山
  • 作者简介:孙三山(1983-),男,四川达州人,博士研究生,主要研究方向:无线资源管理、网络资源虚拟化;汪帅(1991-),男,江西宜春人,硕士研究生,主要研究方向:下一代网络技术;樊自甫(1977-),男,安徽舒城人,副教授,硕士,主要研究方向:下一代网络技术、电信组织和运营管理。
  • 基金资助:
    四川省教育厅基金资助项目(13ZB0159)。

Flow scheduling cost based congestion control routing algorithm for data center network on software defined network architecture

SUN Sanshan1,2, WANG Shuai3, FAN Zifu3   

  1. 1. National Key Laboratory of Science and Technology on Communications (University of Electronic Science and Technology of China), Chengdu Sichuan 611731, China;
    2. School of Physics and Electronic Engineering, Sichuan Normal University, Chengdu Sichuan 610101, China;
    3. School of Communication and Information Engineering, Chongqing University of Posts and Telecommunications, Chongqing 400065, China
  • Received:2015-12-24 Revised:2016-04-06 Online:2016-07-10 Published:2016-07-14
  • Supported by:
    This work is partially supported by the Scientific Research Fund of Sichuan Provincial Education Department (13ZB0159).

摘要: 针对传统数据中心网络极易发生拥塞的问题,提出了在软件定义网络(SDN)的架构下设计基于流调度代价的拥塞控制路由算法加以解决。首先,进行拥塞链路上的大小流区分,并对所有大流的各条等价路径进行路径开销权重的计算,选择权重最小的路径作为可用调度路径;然后,使用调度后路径开销变化量和流占用带宽比例来共同定义流调度代价;最终选择调度代价最小的流进行调度。仿真结果表明,所提算法能在网络发生拥塞时降低了拥塞链路上的负荷,并且与仅进行流路径选择的拥塞控制算法相比,提高了链路利用率,减少了流传输时间,使得网络链路资源得到更好的利用。

关键词: 数据中心网络, 软件定义网络, 拥塞控制, 链路负荷, 流调度代价

Abstract: To alleviate the congestion of traditional data center network, the flow scheduling cost based congestion control routing algorithm on Software Defined Network (SDN) architecture was proposed. Firstly, the maximum flows and minimum flows on congestion links were differentiated. Secondly, the cost of each equivalent route for maximum flows was calculated, and the routes with minimum cost were selected as available scheduling routes for maximum flows. Then, the flow scheduling cost of each available scheduling route was defined by considering the cost change of rerouting operation and the occupancy ratio of bandwidth together. Finally, the maximum flows with minimum scheduling cost were scheduled to related available scheduling routes. The experimental results show that, the proposed algorithm is qualified to reduce the load of congestion links when network congestion occurs. Moreover, compared with the previous congestion control algorithm which taking into account the flow route selection only, the proposed algorithm improves the link utilization and reduces the transmission time of flow, which make the network link resources to be better used.

Key words: data center network, Software Defined Network (SDN), congestion control, link load, flow scheduling cost

中图分类号: