| Software-defined network development has been no longer limited to single-control experimental networks and campus networks.Distributed control plane cluster deployment has become the deployment trend and application requirements of future networks.Therefore,the optimization problem of controller deployment has become a hot topic in the academic and industrial circles in recent years.There have been many studies on controller deployment scenarios for software-defined networks,but these studies are often based on single-objective optimization studies.In the actual use of the network,network delay,network reliability,network deployment overhead,and operation and maintenance overhead are all issues worth considering.Based on the above points,the paper studies the algorithm and mechanism of controller deployment based on online computing in software-defined networks.Firstly,the paper proposes a performance scale model based on multiple optimization objectives for the distributed multi-controller deployment architecture in the current software-defined network research hotspot.Meanwhile,based on the requirements of multi-obj ective optimization and online computing,combined with the fast convergence of simulated annealing algorithm and the stability of immune optimization algorithm,the paper designs a multi-controller deployment algorithm based on online computing.At the same time,this thesis builds a simulation verification platform that can verify the controller deployment algorithm.We use virtualized container technology to build a multi-controller software-defined network.Finally,the simulation is verified to verify that the proposed algorithm is compared to the traditional single optimization algorithm.The algorithm has higher computational efficiency and lower deployment overhead while maintaining network latency and network reliability in compliance with performance specifications. |
PDF Full Text Request