Department of Electronic, Electrical and Computer Engineering, Sungkyunkwan University, (16419) 2066, Seobu-Ro, Jangan-Gu, Suwon-Si, Gyeonggi-Do, Korea.
College of Software, Sungkyunkwan University, (16419) 2066, Seobu-Ro, Jangan-Gu, Suwon-Si, Gyeonggi-Do, Korea.
Sensors (Basel). 2019 Jan 14;19(2):311. doi: 10.3390/s19020311.
Load Balancing (LB) is one of the most important tasks required to maximize network performance, scalability and robustness. Nowadays, with the emergence of Software-Defined Networking (SDN), LB for SDN has become a very important issue. SDN decouples the control plane from the data forwarding plane to implement centralized control of the whole network. LB assigns the network traffic to the resources in such a way that no one resource is overloaded and therefore the overall performance is maximized. The Ant Colony Optimization (ACO) algorithm has been recognized to be effective for LB of SDN among several existing optimization algorithms. The convergence latency and searching optimal solution are the key criteria of ACO. In this paper, a novel dynamic LB scheme that integrates genetic algorithm (GA) with ACO for further enhancing the performance of SDN is proposed. It capitalizes the merit of fast global search of GA and efficient search of an optimal solution of ACO. Computer simulation results show that the proposed scheme substantially improves the Round Robin and ACO algorithm in terms of the rate of searching optimal path, round trip time, and packet loss rate.
负载均衡（LB）是最大化网络性能、可扩展性和鲁棒性所需的最重要任务之一。如今，随着软件定义网络（SDN）的出现，SDN 的负载均衡已成为一个非常重要的问题。SDN 将控制平面与数据转发平面解耦，以实现对整个网络的集中控制。LB 通过将网络流量分配到资源中，从而避免任何资源过载，从而最大限度地提高整体性能。蚁群优化（ACO）算法已被公认为是几种现有优化算法中用于 SDN 负载均衡的有效算法。收敛延迟和搜索最优解是 ACO 的关键标准。在本文中，提出了一种将遗传算法（GA）与 ACO 集成的新型动态 LB 方案，以进一步提高 SDN 的性能。它利用了 GA 的快速全局搜索和 ACO 的最优解的高效搜索的优点。计算机仿真结果表明，与轮询和 ACO 算法相比，所提出的方案在搜索最优路径的速率、往返时间和丢包率方面有显著的提高。
