The Load Balancing Research Of SDN Based On Ant Colony Algorithm

In recent years, with the development of services of the mobile Internet,e-commerce and the large data, the number of users increases in proportion, and thesize of the Internet and the flow fall into the explosive growth， while thevirtualization and the cloud computing technology develop rapidly, that is not only aconsiderable challenge for the capacity of the network, but also the traditionalnetwork architecture with IP exposes more and more problems.In order to find asolution, Stanford clean slate project team proposed a new network architecture，thatis software-defined networking（SDN）.in the network system, SDN abstracts the layerof control to a layer in traditional network layers, and that separates with the dataforwarding layer,that controls and manages all processing motion of the wholenetwork, and the OpenFlow switch is only responsible for forwarding of the data inthe network.Although Software Defined Network (SDN) starts relatively late, thespeed of development is very fast, and more and more researchers will join this camp.There are a lot of problems in Software Defined Network, and one of the importantproblems is the load balancing problem of the network, and it directly affects thestability and the operating costs of the whole network.In current networkenvironment, the new network architecture (SDN) is certainly the trend of thenetwork technology, so the corresponding research of technology (such as loadbalancing) is imperative in this new network architecture.While the traditional IPnetwork has many mature load balancing algorithms, however, after reading theknowledge about the load balancing of SDN I found that the study of load balancingtechnology of this new network architecture is still exploratory stage, yet relativelymature and stable technology. In this paper, through the study and research about SDN and its relatedtechnologies, and the analysis and summary of the traditional IP network loadbalancing techniques, this paper attempts to apply the ant colony algorithm to thisnew network architecture to achieve load balancing of each chain road. The antcolony algorithm that entomologists found is that the ant colony always search ashortest path between ants’nest and food in nature, and is gradually recognized by theindustry.Because of SDN having the feature that is control layer separating fromforwarding layer, the control layer as the core of network monitoring and managethe entire network. The ant colony algorithm running on the controller calculate thechain road of the minimum load based on real-time network load provided by thecontroller, so the controller provides dynamic data forwarding strategy to switch.Meanwhile, about some of the parameters affecting the ant colony algorithm’sperformance, we passed a large number of experiments to complete the values ofthese parameters, so that the performance can be improved ant colony algorithm tofurther improve the effectiveness of the load balancing of SDN.This paper uses network simulation software Mininet and OpenDayLightcontroller to built and set up the environment of SDN, to generate custom networktopology, through the introduction of the scheme of load balancing of the ant colonyalgorithm to monitor the experiment, obtained a series of experimental data.Byanalysing the experimental data in detail, we found that the ant colony algorithm isintroduced to this new network architecture can achieve the network link loadbalancing, and made choices the corresponding parameters of the experiment, it ispossible to achieve a better load balance of each link of SDN, which met our intendedpurpose.