Research On SDN Path Enhancement Algorithm

In order to keep up with the development of Internet technology,network technology is also in urgent need of change.Nowadays,the scale of the network is becoming larger and larger,the network structure tends to be complex,and it is necessary to carry a large number of different types of traffic services.The routing algorithms of traditional networks are all in distributed routing protocols.It is particularly difficult to calculate the path that meets the requirements of various complex QoS traffic services.At this time,the emergence of SDN technology,its centralized control and programmable features can effectively solve many problems faced by traditional networks.The path calculation element in the SDN controller can calculate the path through the routing algorithm combined with the global topology.However,due to the inefficiency of its own routing algorithm and the lack of network multi-constraint calculation,and the lack of load balancing.So,an in-depth reseacher of the path planning algorithm is needed.This thesis proposes a genetic routing algorithm for SDN by improving the traditional genetic algorithm and combining the specific constraint path computation problem model.The encoding of path is adopted,and the initial population is obtained by the improved minimum hop-first algorithm.The path planning constraints is solved before the algorithm iteration,the performance of the traditional algorithm is enhanced and optimized.The experimental verify that the algorithm can calculate the relative optimal solution while satisfying the multiple constraints and achieve the goal of path optimization.To solve the problem of network link load balancing,we first regard it as a multi-objective path allocation problem,which is essentially a multi-objective optimization problem(MOP).There is no precise mathematical method that can be solved in polynomial time in academia.Therefore,based on the idea of heuristic algorithm,thought of the multi-objective particle swarm optimization algorithm,a link load balancing method based on multi-path allocation scheduling strategy is designed.Finally,the method demonstrates the feasibility and performance superiority of the simulation test.