摘要
针对SDN多控制器负载均衡过程中交换机迁移僵化和迁移冲突问题,对现有的交换机迁移模型实施优化,提出了一种阶段式控制器负载均衡策略。在阶段1,综合控制器的各类开销代价,基于遗传算法优化迁移目标,选取交换机迁入和迁出域;在阶段2,通过设定子域迁移度和交换机有效期,实现了多交换机的协调迁移。仿真结果表明,与现有的负载均衡策略相比,改善了迁移域的选取过程,优化了交换机到控制器的动态部署,控制器负载均衡率提升了约21%。
Aiming at rigid switch migrating and conflicting migration during SDN multi-controller load balancing, this paper optimized the existing switch migration model and proposed a stage controller load balancing strategy(SCLB). In phase one ,by synthesizing various types of controller overhead and optimizing the migration target based on genetic algorithm, it selected im- migration domain and emigration domain of switch. In phase two, it completed the coordination of multi-switch migration through setting subdomain migration degree and switch validity. The simulation results show that, compared with the existing load balancing policies, SCLB improves the process of selecting migration domain and optimizes dynamic deployment between switch and controller. The controller load balancing level is increased about 21%.
出处
《计算机应用研究》
CSCD
北大核心
2018年第2期559-563,共5页
Application Research of Computers
基金
国家自然科学基金创新研究群体科学基金资助项目(61521003)
国家自然科学基金资助项目(61372121
61521003)
国家科技支撑计划课题(2014BAH30B01)
国家"973"计划资助项目(2012CB315901
2013CB329104)
国家"863"计划资助项目(2015AA016102
2013AA013505)