Overlay network routing decision-making method based on deep learning

Abstract

The invention discloses an overlay network routing decision-making method based on deep learning. The method comprises the following specific steps: S1, deploying overlay network routing nodes, and collecting a bandwidth time delay change data set; S2, making a data set for training, wherein the data set comprises a prediction network data set and a classification network data set selected by thecoverage routing node, the label of the prediction network is the time delay and bandwidth data of the next moment, and the label calculation mode of the classification network selected by the coverage routing node is obtained by calculation on the coverage network topology formed by the coverage network routing node by adopting a shortest path algorithm; S3, constructing a bandwidth delay prediction network based on a recurrent neural network LSTM and a coverage routing node selection classification network for coverage routing node selection; S4, training and optimizing a deep learning model; and S5, making a decision through the model. According to the method, the decision response speed can be greatly improved while the routing decision effect is ensured, especially when the scale of coverage network nodes is huge.

Description

technical field [0001] The invention relates to the fields of computer network and video transmission, in particular to a real-time interactive video stream transmission scene with high transmission delay requirements, and in particular to an overlay network routing decision method based on deep learning. Background technique [0002] With the development of the Internet, online entertainment has widely penetrated into people's daily life. Moreover, with the interactive attributes of online social entertainment, the user's quality of experience is highly related to network delay. In a cloud gaming application scenario, if the network delay is greater than 50ms, 70% of gamers will feel an obvious delay. Additionally, the underlying routing path from the hosting game server to the user will experience high network latency if network congestion occurs in the underlying routing path. If the underlying routing connection fails, the cloud game service will experience a long reco...

AI Technical Summary

Problems solved by technology

In cloud game application scenarios, if the network delay is greater than 50ms, 70% of gamers will feel obvious delay

Additionally, the underlying routing path from the hosting game server to the user will experience high network latency if network congestion occurs in the underlying routing path

If the underlying routing connection fails, the cloud game service will experience a long recovery time, which will greatly affect the quality of experience

[0003] At the same time, for the traditional routing decision-making method (bottom routing selection), in the process of routing path selection, first of all, for any node, only the bandwidth and delay of the current node can be observed, and it is impossible to predict the current situation at the next moment. The parameters of the node, so that the optimal path of the current node cannot be selected

Secondly, for the overlay routing decision-making method, due to the design and construction of many overlay routing nodes, if the traditional shortest optimal path algorithm is used to calculate the node selection process, the calculation amount is too large at this time to satisfy the real-time selection of routing. need

Images