A Multi-Controller Deployment Approach for Large-Scale Software-Defined Networks

Abstract

The invention discloses a multi-controller deployment method suitable for large-scale software-defined networks, which uses crossover and mutation sub-algorithms to change the number of controllers to ensure that the number of controllers in the solution concentration is diverse; and based on the Pareto minimum The optimal principle performs genetic optimization on multiple objectives. For a given network, a set of Pareto optimal solutions can be obtained. Among the two performance scales of server load balancing, at least one performance scale has better performance, so as to ensure that the network has better performance in terms of delay and load balancing; at the same time, when the network fails, dynamic deployment control can be realized Device; in the case of some network failures, it can effectively guarantee the performance of the network.

Description

technical field [0001] The invention belongs to the technical field of multi-controller deployment, and in particular relates to a multi-controller deployment method suitable for large-scale software-defined networks. Background technique [0002] In recent years, with the development of network technology, a decoupled network architecture - Software Defined Network (SDN) has been proposed. With the in-depth research on SDN and the development of its technology, SDN technology is being continuously applied in various network fields. With the development and expansion of SDN applications, the scale of SDN networks is also increasing. When SDN is applied in a large-scale network, a single controller cannot efficiently grasp and control the traffic of the entire network. Based on considerations of control plane performance and network performance, multiple controllers need to be deployed in the network to reduce and balance the load on the controllers, optimize overall networ...

AI Technical Summary

Problems solved by technology

One is to optimize the expected failure rate of the link between the controller and the switch when solving the CPP problem, and ensure the robustness of the network by reducing the expected failure rate. At present, there are many research contents in this direction, but when the network occurs In the event of a failure, the dynamic deployment of the controller cannot be realized

The second is to design a dynamic deployment scheme. When the network fails, the dynamic deployment scheme is used to quickly solve the CPP problem and deploy the controller to deal with network failures, improving the robustness of the network. Researchers such as Bari M and Roy A proposed a dynamic deployment program, and currently there are few research contents in this direction

Images