Network rerouting method based on multilayer mask attention mechanism

By introducing a network rerouting method with a multi-layer mask attention mechanism in the software-defined network (SDN), identifying and rerouting key traffic, the negative impact of traditional SDN management methods on network service quality is solved, and more efficient network resource utilization and service quality are achieved.

CN119728518APending Publication Date: 2025-03-28BEIJING UNIV OF POSTS & TELECOMM
0 Cites 2 Cited by

Patent Information

Application Number
CN202411567887.2
Authority / Receiving Office
CN · China
Patent Type
Applications(China)
Current Assignee / Owner
Filing Date
2024-11-05
Publication Date
2025-03-28

AI Technical Summary

Technical Problem

Traditional software-defined network (SDN) management methods often rerout large amounts of traffic, will have a negative impact on the network's quality of service (QoS) and existing reinforcement learning methods have little research on key traffic selection, and have failed to effectively identify and prioritize the critical flows that have the greatest impact on network performance.

Method used

A network rerouting method based on multi-layer mask attention mechanism is proposed. By inputting the topological characteristics and spatiotemporal characteristics of the network into a pre-trained neural network model, the traffic information related to network bottlenecks and key links is identified and centrally processed, key traffic that is more influential to the entire network is selected, and rerouted.

Benefits of technology

By intelligently selecting and rerouting key traffic in the network, it can more effectively respond to rapid changes in network status and maximize network resource utilization efficiency and quality of service (QoS).

✦ Generated by Eureka AI based on patent content.
Patent Text Reader

Abstract

The invention discloses a network rerouting method based on a multi-layer mask attention mechanism. The method comprises the following steps: inputting current network topology information and network flow information of a network into a pre-trained neural network model obtained by combining a multi-layer mask self-attention mechanism and an Actor-Critic actor commentator network; the neural network model selects key flow information from the network flow information; performing network rerouting based on a transmission path of the key traffic determined by the key traffic information; wherein the mask matrix of the mask self-attention mechanism is determined according to the network topology information. According to the invention, the key traffic which has more influence on the whole network can be selected, and the key traffic can respond to the rapid change of the network state more effectively through the rerouting of the key traffic.
Need to check novelty before this filing date? Find Prior Art