航空学报 > 2018, Vol. 39 Issue (7): 321961-321961   doi: 10.7527/S1000-6893.2018.21961

一种面向航空集群的集中控制式网络部署方法

吕娜, 刘创, 陈柯帆, 曹芳波   

  1. 空军工程大学 信息与导航学院, 西安 710077
  • 收稿日期:2017-12-21 修回日期:2018-04-23 出版日期:2018-07-15 发布日期:2018-04-23
  • 通讯作者: 刘创 E-mail:lc515626@foxmail.com
  • 基金资助:
    国家自然科学基金(61472443)

A method for centralized control network deployment of aeronautic swarm

LYU Na, LIU Chuang, CHEN Kefan, CAO Fangbo   

  1. School of Informational and Navigation, Air Force Engineering University, Xi'an 710077, China
  • Received:2017-12-21 Revised:2018-04-23 Online:2018-07-15 Published:2018-04-23
  • Supported by:
    National Natural Science Foundation of China (61472443)

摘要: 软件定义网络(SDN)的发展为机载网络构建提供了全新的设计思想。针对航空集群环境下控制平面的可扩展性问题,研究了混合层次式架构下的多控制器部署问题。首先,为实现对底层传输节点的弹性管控,通过扩展控制层级定义了本地控制器资源池。然后,为减少时间复杂度,实现本地控制器的快速部署,将传统的多控制器直接部署转化为子域划分和域内部署两个步骤,提出了基于节点密度排序的子域划分算法和基于改进多目标模拟退火的域内控制器部署算法。实验结果表明:与已有算法相比,所提算法在保证控制器的负载均衡基础上,可以有效减少网络控制链路的平均时延和平均失连率,同时时间复杂度更低,能够适用于大规模及动态网络环境下的控制器部署场景。

关键词: 航空集群, 机载网络, 软件定义网络, 控制平面架构, 控制器部署

Abstract: The development of Software Defined Networking (SDN) provides a completely new design concept for construction of airborne network. For the problem of scalability of control plane in aeronautic swarm, the problem of multi-controller deployment under hybrid hierarchy is studied. First, to realize flexible control of the underlying transport nodes, the local controller resource pool is defined by extending the control hierarchy. Then, to reduce the time complexity and realize rapid deployment of local controllers, the traditional controller deployment is transformed into two steps:sub-domain partitioning and intra-domain deployment. A sub-domain partitioning algorithm based on node density sorting and an in-domain controller deployment algorithm based on improved Pareto simulated annealing are developed. Experimental results show that the proposed algorithm can guarantee load balancing of the controller, and effectively reduce the average delay and average loss rate of network control link. Compared with existing algorithms, the algorithm is more suitable for aeronautic swarm controller deployment in large-scale and dynamic topology environment with less time complexity.

Key words: aeronautic swarm, airborne network, software defined networking, control plane architecture, controller deployment

中图分类号: