Research On Load Balancing Routing Algorithm In Software Defined Network

In recent years,With the rapid development of cloud computing technology and big data,a variety of new Internet applications have emerged.Due to the increasing demand for data storage and cloud computing services,the network needs to deal with the traffic increasingly.The closed structure of the traditional TCP/IP network exists some problems such as low bandwidth utilization,which hinder the development of load balancing technology.How to maximize the use of limited cyber source is one of problems that the current Internet technology to deal with.The emergence of the Software Defined Network breaks the traditional network's closure and increases the flexibility of the network to meet changing network requirements.SDN separates the control plane and the forwarding plane,the controller can obtain the state information in real time to manage the network and conveniently realize the network load balance better.In this paper,we propose a load balancing algorithm based on traffic scheduling,which is combined with the local and dynamic characteristics in the data center network.First of all,this algorithm sets a weight which considers the link critical value and the link congestion degree for all paths in the network according to the basic information of the monitoring network controller,and we select the best path based on weight to transmit data streams.Then,when the network triggers the load balance threshold,the link with the maximum load is selected and the large flow which satisfies the constraint condition is selected to schedule.This paper design related modules on RYU controller to achieve the proposed algorithm.Finally,the performance of proposed algorithm through the Mininet platform is analyzed,compared with ECMP algorithm and DLB algorithm in the average bandwidth utilization,the average delay and the traffic load distribution.The effectiveness of the proposed load balancing strategy based on flow scheduling is verified.At the same time,we also consider the limitation of the link capacity and the limited capacity of the switch flow table.In order to solve these problems,this paper builds the load balancing model and uses ant colony algorithm to find the solution.This paper uses ant colony algorithm to find the optimal path through two kinds of heuristic information to calculate the state transition probability,and the algorithm use the appropriate rules to update the pheromone.The feasibility and effectiveness of the algorithm proposed in this paper are verified by simulation.Compared the proposed ACLB algorithm in the paper with GDMR algorithm and the PDMF algorithm in the transmission rate,the maximum link utilization and the number of table entry in the switch,the results show that the performance of the ACLB algorithm is better than others two algorithms.