Uplink Bandwidth Allocation Method Based on Neural Network Prediction in 50g-pon Network

Abstract

The invention provides an uplink bandwidth allocation method based on neural network prediction in a 50G-PON network, comprising: training an LSTM neural network; all ONUs send report messages to the OLT to report the queue length; and the OLT prioritizes the services of the ONUs according to the report messages Divide, and sort all ONUs and report queue lengths according to RTT; allocate all services according to the sorting results and LSTM neural network; calculate and allocate the low-priority services of the last ONU to ensure the end of the polling cycle At the same time, the allocation result is obtained; the OLT packages the allocation result and publishes it to all ONUs; each ONU sends information in turn according to the received allocation result. The uplink bandwidth allocation method adopts the method of neural network prediction, which ensures that the data can be forwarded in time and effectively reduces the transmission delay; at the same time, this scheme takes into account the characteristics of bandwidth and delay requirements of different services, and reasonably allocates the transmission order. , to ensure the quality of network service.

Description

technical field [0001] The invention relates to the technical field of optical access network communication, and more particularly, to a method for allocating uplink bandwidth based on neural network prediction in a 50G-PON network. Background technique [0002] PON network system is one of the most widely used network structures at present, and it is the best solution to solve the "last mile" in optical fiber communication. Its structure is a point-to-multipoint tree structure, and one optical line terminal (OLT) is connected with multiple optical network units (ONUs). In its upstream communication, multiple ONUs share an upstream path, which requires reasonable allocation of upstream paths to prevent conflicts when each ONU transmits information. Bandwidth allocation algorithms can be divided into static bandwidth allocation and dynamic bandwidth allocation. The static bandwidth allocation algorithm allocates a fixed bandwidth for each ONU in advance. This algorithm will...

AI Technical Summary

Problems solved by technology

[0004] The traditional offline bandwidth allocation method needs to collect the bandwidth information of all ONUs for unified allocation and scheduling; this method will lead to a long time interval from the ONU sending the bandwidth request information to the actual data sending, and during this time The newly added data cannot be sent in time, and it is necessary to wait for the scheduling of the next cycle, causing delays. Although there are some methods to solve these problems, these methods are based on linear models, which are very obvious for the nonlinear characteristics of network traffic. The effect of the model is poor, which is reflected in the inability to accurately predict the sudden change of traffic

[0005] The Chinese invention patent application with the publication number CN110213679A disclosed a passive optical network system and its implementation method on September 6, 2019, including: the optical line terminal OLT determines a fixed value T, according to the characteristics of the services carried by the PON system Allocate bandwidth for the optical network unit ONU, and the interval between any two adjacent bandwidths does not exceed the fixed value T; although this method makes bandwidth allocation no longer dependent on the bandwidth request of the ONU or the monitoring of the ONU traffic by the OLT, shortening The delay of ONU sending data is reduced, but the time for shortening the transmission delay is very limited, and it is not suitable for the latest standard of IEEE 802.3ca

Images