Font Size: a A A

The Research Of Multi-controller Load Balancing And Collaborative Technology For SDN Network

Posted on:2018-01-22Degree:MasterType:Thesis
Country:ChinaCandidate:X GuoFull Text:PDF
GTID:2348330542472251Subject:Computer Science and Technology
Abstract/Summary:PDF Full Text Request
The definition of software-defined network solves the problem that the coupling degree of network equipment control layer and data forwarding plane is too high in the traditional network.Compared with the traditional network,SDN network structure is easier to manage and deploy,making the network service update more quickly,and thus received wide attention.For the issue of load balancing,existing load balancing methods are mostly based on static controller deployment optimization or single switch migration method.Static controller deployment optimization method can not meet the requirements of software-defined network dynamic traffic changes,and the single switch migration method would not make the overall control plane load balancing.In view of this situation,this paper first carries on the model to the control plane load balancing question,then proposes a global switch migration method based on genetic algorithm.Through the genetic algorithm while accord to the switch traffic request to regenerate the controller switch mapping,And then migrate the switch so that the entire control plane to achieve load balancing.Experiments show that the load balancing method of this paper can make the global controller utilization more evenly than single switch migration strategy.For the issue of routing reconfiguration,this paper presents a method of multi-controller cooperative route reconfiguration based on domain partition.The method divides the switch into three domains according to the overlap of the switches in the old and new routing paths,and coordinates the domain controllers in the three domains through the designed multi-controller cooperative platform to distribute the rules in order.The method solves the problem of network packet loss during route reconfiguration effectively.Experimental results show that the method proposed in this paper can greatly reduce the packet loss rate as the packet rate increases and the domain sizes expands.Compared to the method based on controller storage has a lot of performance improvement.
Keywords/Search Tags:Software-Defined Network, Multi-Controller, Load Balancing, Genetic Algorithm, Routing Reconfiguration
PDF Full Text Request
Related items