Deep-learning-based SDN (Software Defined network) flow forecasting method

Abstract

The invention discloses a deep-learning-based SDN (Software Defined network) flow forecasting method, which comprises the following steps: building an SDN flow forecasting model, and respectively adding corresponding modules into all layers of an SDN; acquiring a network flow of a forwarding layer to a control layer for forecasting analysis to realize overall optimization of the application layer;and putting a network flow forecasting strategy into a whole SDN system in a penetrating manner to guarantee stable and efficient operation of a network and improve the quality of service of the SDNsystem. The invention provides an SDN flow forecasting model and designs a complete forecasting mechanism on the basis of the forecasting model. All the modules work coordinately to complete a networkflow forecasting function together, and a forecast result can be used for link switching of a lower layer and application analysis of an upper layer to realize jam control and load balancing of the network; and the distribution of network resources and the networking design are reasonably optimized to improve the operation efficiency of the SDN.

Description

technical field [0001] The invention belongs to the technical field of network communication, and in particular relates to a deep learning-based SDN network traffic prediction method. Background technique [0002] With the rapid development of modern computer network technology, the number of network applications is also increasing rapidly, and the traffic in the network is also increasing exponentially, which puts forward higher requirements for the performance of network equipment. However, it is impractical to solve the problems existing in the current network simply by improving the performance of the equipment. [0003] Because, the ubiquitous problem is that the distribution of traffic is often in an unbalanced state. In areas with dense traffic, devices are often overwhelmed and fail or even paralyzed; while in areas with sparse traffic, devices often cannot get optimal utilization. Therefore, it is very important to optimize the transmission of traffic. This can n...

AI Technical Summary

Problems solved by technology

[0010] Aiming at the deficiencies of the prior art, the present invention proposes a deep learning-based SDN network traffic prediction method, which provides a traffic prediction function for the unbalanced distribution of SDN network traffic, obtains traffic at the forwarding layer, and performs prediction and analysis at the control layer to realize The overall optimization of the application layer; the network traffic prediction strategy runs through the entire SDN network system to ensure the stable and efficient operation of the network and improve the service quality of the SDN network system; specifically, the following steps are included:

Images