SDN switch node incremental deployment method based on genetic algorithm

Abstract

The invention relates to an SDN switch node incremental deployment method based on a genetic algorithm. The SDN switch node incremental deployment method comprises the following steps: S1, generatingan initialized deployment sequence by adopting a random algorithm; S2, calculating a deployment cost value corresponding to the initialized deployment sequence S; S3, sorting the deployment cost values from small to large; S4, dividing the sorted deployment sequences into three sets according to the deployment cost values of the deployment sequences; S5, selecting parents of which the number is the same as that of elements in the middle class; S6, performing crossover variation to generate a new generation of deployment sequence; S7, updating the initial deployment sequence; S8, calculating adeployment cost value of each deployment sequence in the new deployment sequence population; and S9, recording the deployment sequence with the minimum deployment cost value and the corresponding value, updating the optimal deployment sequence and the value thereof, performing iterative search for a plurality of times, and taking the deployment sequence with the minimum deployment cost value in the finally generated deployment sequence population as the optimal deployment sequence of the switch node.

Description

technical field

[0001] The invention relates to the technical field of the Internet, in particular to a genetic algorithm-based SDN switch node incremental deployment method. Background technique

[0002] With the rapid development of Internet applications, the traffic of Internet Service Provider (ISP) network is exploding. Traditional heuristic routing optimization methods cannot get rid of the restriction that traffic is routed on the shortest path, resulting in frequent congestion in the network, and network performance, user experience, and satisfaction cannot be guaranteed.

[0003] The emergence of software-defined networking (Software Defined Networking, SDN) has brought new ideas to the solution of routing optimization problems. The SDN centralized controller can dynamically and flexibly adjust the routing of traffic in the network, getting rid of the restrictions of weight modification and routing convergence in traditional networks, and can greatly improve routin...

Examples

Embodiment Construction

[0036] The present invention will be further described below in conjunction with the accompanying drawings and embodiments.

[0037] Please refer to figure 1 , the present invention provides a kind of SDN switch node increment deployment method based on genetic algorithm, comprises the following steps:

[0038] Step S1: Use a random algorithm to generate the full permutation sequence of some network topology nodes as the initial deployment sequence.

[0039] Step S2: Calculate the initial deployment sequence , the sum of the corresponding maximum link utilization of the network, as the deployment cost value of the sequence ,specifically:

[0040] Step S2.1: To the nodes in the current deployment sequence , computing deployment nodes increase After that, the maximum link utilization of the network value.

[0041] Step S2.1.1: Traverse every node in the network a , using Dijkstra's algorithm to construct from a The starting shortest path tree to each node. Then tr...

technical field

[0001] The invention relates to the technical field of the Internet, in particular to a genetic algorithm-based SDN switch node incremental deployment method. Background technique

[0002] With the rapid development of Internet applications, the traffic of Internet Service Provider (ISP) network is exploding. Traditional heuristic routing optimization methods cannot get rid of the restriction that traffic is routed on the shortest path, resulting in frequent congestion in the network, and network performance, user experience, and satisfaction cannot be guaranteed.

[0003] The emergence of software-defined networking (Software Defined Networking, SDN) has brought new ideas to the solution of routing optimization problems. The SDN centralized controller can dynamically and flexibly adjust the routing of traffic in the network, getting rid of the restrictions of weight modification and routing convergence in traditional networks, and can greatly improve routin...

Examples

Embodiment Construction

[0036] The present invention will be further described below in conjunction with the accompanying drawings and embodiments.

[0037] Please refer to figure 1 , the present invention provides a kind of SDN switch node increment deployment method based on genetic algorithm, comprises the following steps:

[0038] Step S1: Use a random algorithm to generate the full permutation sequence of some network topology nodes as the initial deployment sequence.

[0039] Step S2: Calculate the initial deployment sequence , the sum of the corresponding maximum link utilization of the network, as the deployment cost value of the sequence ,specifically:

[0040] Step S2.1: To the nodes in the current deployment sequence , computing deployment nodes increase After that, the maximum link utilization of the network value.

[0041] Step S2.1.1: Traverse every node in the network a , using Dijkstra's algorithm to construct from a The starting shortest path tree to each node. Then tr...

Images