The Controller Deployment In SDN

With the continuous expansion of the network scale and the rapid growth of network traffic.The network management and control capabilities have been weakened,increasing the operator's network operation and maintenance costs.At present,the closeness of network equipment greatly limits the flexibility and scalability of the network,making new protocols and new services difficult to experiment and deploy in real network environments.In this context,Software-Defined Networking(SDN)is an innovative network architecture that separates logic control from data forwarding and provides a flexible programming interface for applications.This significantly improves the flexibility of the network and allows it to be controlled.Sexuality and programmability are considered to be one of the most promising development directions in the future network field,and have become a research hotspot in the network field in the future.Logically centralized control plane is prone to arise the challenge of performance bottleneck in software defined network architecture.Generally,multiple controllers should be deployed to handle all flow requests from switches in the network.Thus,it is an important issue that how to deploy multiple controllers reasonably in practical deployments of software defined networks.Compared with the research on controller deployment by other scholars proposed,we proposes a general-purpose controller deployment multi-objective model that considers network real-time,service quality,and reliability factors.The studies of this paper are as follows:(1)This paper proposes a controller deployment model based on delay and load for large-scale network scenarios.The model firstly combines the characteristics of network bearer services,considers the network performance requirements of the SDN network in the end-to-end delay,models the average flow installation delay and controller load variance of the switch,and proposes a multi-objective optimization model for controller deployment.Based on the particle swarm optimization algorithm and simulated annealing algorithm,the optimal controller deployment solution was solved,and the balanced deployment of stream installation delay and controller load was realized.Experimental results indicate that our proposed model is better with smaller average flow setup delay and controller load variance when deploying a same number of controllers for a given network topology.(2)This paper proposes a QoS-based controller deployment model for large-scale actual production networks.The model first analyzes the actual factors affecting QoS.At the same time,taking into account the reliability of the entire network and the priority characteristics of the flow,based on the theory of queuing theory and reliability theory,it deduces the modeling of the response delay and network reliability of the SDN controller.An optimization model for controller deployment was proposed,and the optimal controller deployment solution was solved based on the k-means algorithm to implement SDN controller deployment that satisfies QoS requirements,thereby effectively guaranteeing network reliability and service quality.The experimental results show that when deploying the same number of controllers in the network,the response time obtained by the model proposed in this paper is smaller and the network reliability is higher.