SDN multistage virtual network mapping method and device based on reinforcement learning

A technology of virtual network mapping and reinforcement learning is applied in the field of SDN multi-level virtual network mapping based on reinforcement learning, which can solve the problems of reducing algorithm flexibility and request acceptance rate.

Active Publication Date: 2019-10-22
BEIJING UNIV OF POSTS & TELECOMM
View PDF9 Cites 17 Cited by
  • Summary
  • Abstract
  • Description
  • Claims
  • Application Information

AI Technical Summary

Problems solved by technology

However, if the exact virtual network mapping algorithm is directly used to solve the mapping of each level in turn, when the resources requested by the upper

Method used

the structure of the environmentally friendly knitted fabric provided by the present invention; figure 2 Flow chart of the yarn wrapping machine for environmentally friendly knitted fabrics and storage devices; image 3 Is the parameter map of the yarn covering machine
View more

Image

Smart Image Click on the blue labels to locate them in the text.
Viewing Examples
Smart Image
  • SDN multistage virtual network mapping method and device based on reinforcement learning
  • SDN multistage virtual network mapping method and device based on reinforcement learning
  • SDN multistage virtual network mapping method and device based on reinforcement learning

Examples

Experimental program
Comparison scheme
Effect test

Embodiment 1

[0084] Embodiment 1 of the present invention provides an SDN multi-level virtual network mapping method based on reinforcement learning, such as figure 1 As shown, the method includes the following steps:

[0085] Step S100, establishing and training a reinforcement learning mapping model.

[0086] The training process for the reinforcement learning mapping model may include pre-training and / or ad-hoc training. Pre-training means that before the application of the reinforcement learning mapping model (before processing the actual mapping request), the preset or historical actual mapping request data and network resource status data are used as training input in advance, and the reinforcement learning mapping model is repeatedly trained , continuously optimize the model parameters until a model with better solution performance is obtained.

[0087] Step S101, for the current underlying virtual network request, obtain the current resource status information of the physical net...

Embodiment approach

[0094] In the embodiment of the present invention, the training of the reinforcement learning model may include three implementations: one is only pre-training, the second is only temporary training, and the third is pre-training before application and application Temporary training is performed for each mapping request. The first type requires that the pre-trained model is relatively mature, and its performance meets certain requirements. Since it does not require temporary training, this method responds faster; the second type will slow down the response speed, but does not require the previous pre-training process. It is more suitable for occasions that do not require high response speed; the mapping strategy obtained by the model trained in the third method is optimal, and is suitable for occasions that require high mapping strategy.

[0095] Step S103, through the reinforcement learning model, the mapping of the upper layer virtual nodes is sequentially solved.

[0096] ...

Embodiment 2

[0104] Embodiment 2 of the present invention provides a preferred embodiment of an SDN multi-level virtual network mapping method based on reinforcement learning.

[0105] The main process framework of the mapping method provided by Embodiment 2 of the present invention is as follows figure 2 shown. In order to reduce the complexity of the mapping solution, the present invention adopts a two-level-two-step mapping idea. Whether it is processing the bottom layer virtual network request or the upper layer virtual network request, it will not proceed until the mapping solution of all virtual nodes in this layer is completed. The mapping solution of the layer virtual link specifically includes the following steps:

[0106] S200, abstractly represent the request information of the physical network and the virtual network.

[0107] The underlying physical network can be represented as a weighted undirected graph where N S is the set of network nodes, L S is the set of network...

the structure of the environmentally friendly knitted fabric provided by the present invention; figure 2 Flow chart of the yarn wrapping machine for environmentally friendly knitted fabrics and storage devices; image 3 Is the parameter map of the yarn covering machine
Login to view more

PUM

No PUM Login to view more

Abstract

The invention discloses an SDN multistage virtual network mapping method and device based on reinforcement learning. The SDN multistage virtual network mapping method comprises the steps: establishingand training a reinforcement learning mapping model; for an underlying virtual network request, obtaining current resource state information of a physical network and inputting the current resource state information into the reinforcement learning mapping model to perform underlying virtual node mapping; then, carrying out bottom layer virtual link mapping solving; for an upper-layer virtual network request, obtaining current resource state information of a bottom-layer virtual network, inputting the current resource state information into the reinforcement learning mapping model, and performing upper-layer virtual node mapping; then, carrying out upper-layer virtual link mapping solution; and if the mapping fails in any stage, dynamically adjusting the underlying virtual network until all nodes and links are successfully mapped. The SDN multistage virtual network mapping device comprises a reinforcement learning module, a bottom-layer mapping module, an upper-layer mapping module anda dynamic adjustment module. The SDN multistage virtual network mapping method and device are suitable for multistage virtual network mapping, and can improve the overall request acceptance rate.

Description

technical field [0001] The invention relates to the technical field of computer networks, in particular to an SDN multi-level virtual network mapping method and device based on reinforcement learning. Background technique [0002] With the rapid development of cloud computing, the Internet of Things, and 5G, traditional IP architecture networks are increasingly difficult to deploy new network technologies and protocols, and cannot meet the development needs of new services. The combination of Software Defined Network (SDN) and Network Virtualization (NV) technologies is considered to be an effective way to overcome current network rigidity and promote future network innovation. SDN is a new type of network architecture with centralized control. Its core technology, OpenFlow, separates the control plane from the data plane of network equipment and centralizes the logic of the control plane, thereby realizing flexible control of network traffic and enabling the network control...

Claims

the structure of the environmentally friendly knitted fabric provided by the present invention; figure 2 Flow chart of the yarn wrapping machine for environmentally friendly knitted fabrics and storage devices; image 3 Is the parameter map of the yarn covering machine
Login to view more

Application Information

Patent Timeline
no application Login to view more
IPC IPC(8): H04L12/24G06N3/04G06N3/08
CPCH04L41/145H04L41/12G06N3/08G06N3/045
Inventor 卢美莲顾云
Owner BEIJING UNIV OF POSTS & TELECOMM
Who we serve
  • R&D Engineer
  • R&D Manager
  • IP Professional
Why Eureka
  • Industry Leading Data Capabilities
  • Powerful AI technology
  • Patent DNA Extraction
Social media
Try Eureka
PatSnap group products

Browse by: Latest US Patents, China's latest patents, Technical Efficacy Thesaurus, Application Domain, Technology Topic.

© 2024 PatSnap. All rights reserved.Legal|Privacy policy|Modern Slavery Act Transparency Statement|Sitemap