Network flow type prediction method based on deep learning

Abstract

The invention discloses a network flow type prediction method based on deep learning. A multistage prediction scheme of ''edge pre-classification + center fine classification'' is adopted, that is, pre-classification is performed at first and then fine classification is performed, and deep learning models of pre-classification and fine classification are respectively constructed on an SDN switch and an SDN controller of a network edge, wherein the network function virtualization NFV technology is adopted, computing resources of switches in the SDN network and a distributed deep learning network constructed by links are used as hardware resources required for the pre-classification model, and the SDN controller is used as the hardware resource required for the fine classification model; andthe pre-classification model uses four joint features, and the fine classification model uses ten joint features. By adoption of the multistage prediction scheme in the network flow type prediction method disclosed by the invention, the communication overhead of the switch to the controller can be reduced, and the load of the controller can also be alleviated; the prediction is achieved by usinga capsule network method as early as possible; and meanwhile, the deep learning model is periodically trained by using an autonomously updated training data set to improve the prediction accuracy.

Description

technical field [0001] The invention belongs to the technical field of Internet data processing, and in particular relates to a network flow type prediction method based on deep learning. Background technique [0002] Studies have shown that the size and length of flows in the data center network show the differentiation characteristics of elephant flows and mouse flows: the proportion of elephant flows is less than 1%, and the proportion of traffic is greater than 90%; on the contrary, the number of mouse flows accounts for more than 99% , traffic accounted for no more than 10%. Moreover, the duration of the mouse flow is extremely short, most of which do not exceed 100ms. In a software-defined data center network, the SDN controller needs to frequently create a flow table for the mouse flow. The life cycle of the flow table from formulation to distribution is very short, and the efficiency will be very low. Consumption such as available secure control channels, etc.) wil...

AI Technical Summary

Problems solved by technology

However, many current methods only use thresholds to identify elephant flows, while ignoring the duration of flows and the real demand for resources

Images