Research On Elephant Flow Scheduling In SDN Data Center Network Based On Genetic-ant Colony Hybrid Algorithm

The rapid development and wide application of cloud computing,big data and Internet of things technologies make the data center network more and more large in scale,more complex in structure and more and more difficult in internal resource management.The traffic pattern within the data center network has also changed a lot,with more and more traffic between servers.Using the traditional static traffic scheduling algorithm,the network equipment and link load will be uneven,and even equipment downtime and link congestion.SDN technology provides a new opportunity to solve the problem of data center network.Therefore,aiming at the traffic scheduling problem in data center network,this paper firstly establishes the mathematical model of integer linear programming,and proposes a traffic scheduling algorithm GA-AC O based on SDN technology to solve the mathematical model and obtain the optimal traffic scheduling scheme.This algorithm first uses ECMP to dispatch the newly arrived data flow,and at the same time uses Sflow technology to monitor the link state information and elephant flow in real time.For the elephant flow on the congestion link,in combination with the current network state,GA-ACO first uses the genetic algorithm to calculate multiple available paths,and then selects the optimal path through the ant colony algorithm to re-route the elephant flow.Compared with ECMP algorithm and ant colony optimization algorithm ACO-SDN,GA-ACO has achieved better results in reducing the maximum link utilization rate and increasing the average bisection bandwidth.