Research On Path Optimization Method Based On Link State Awareness In SDN Environment

With the development of Internet services,the variety and number of business traffic has increased dramatically,which makes the traditional distributed network architectures faces challenges.The emergence of Software Defined Network(SDN)provides a new idea for routing policy formulation in multi-service traffic scenarios.This topic is mainly aimed at the scenario where the multi-service traffic coexisted in the current network.By using the advantages of centralized control and programming of the SDN network,the unnatural ant colony algorithm(UACA)is designed to meet the business.A UACA path optimization system enables to performs traffic Quality of Service(QoS)requirements and load balancing of traffic on the network.The research content of this topic includes two aspects.Firstly,this research proposes the unnatural ant colony load balancing algorithm in SDN environment,which regards the data packets in the network as ants,correlates the number of data packets passing through the link with the concentration of pheromone,thus achieving load balancing.The candidate path of the unnatural ant colony algorithm is filtered by Top-K optimal path algorithm.At the same time,the link information collection mechanism in the SDN environment is designed to obtain the link state parameters,the link state parameters are associated with corresponding link pheromones and the pheromone evaporation rate is correlated to optimize the algorithm model.Secondly,based on the proposed algorithm,this research implements the UACA path optimization system by developing six modules in the application layer.Each module is a QoS requirement registration module,a traffic identification module,a pheromone management module,a link status sensing module,an unnatural ant colony algorithm module,and a flow table generation delivery module,wherein the unnatural ant colony algorithm module is designed based on the subject.The non-natural ant colony algorithm is implemented to realize the path optimization model combining deterministic and probabilistic routing.This research simulates the common multi-service traffic scenario network through the Mininet simulation platform,simulates the traffic flow using the Iperf command,and conducts multiple sets of comparison experiments.The experimental results show that compared with Dijkstra's shortest hops single path selection algorithm and equivalent multipath routing algorithm,the path optimization system designed in this research can better guarantee the QoS attribute requirements of service traffic,has lower packet loss rate and smaller end-to-end delay in a network with large traffic network and fluctuating link.