Google's BBR (Bottleneck Bandwidth and Round-trip Propagation Time) approach is used to enhance internet network transmission. It is particularly intended to efficiently handle enormous amounts of data. Traditional TCP (Transmission Control Protocol) algorithms confront the most difficulty in calculating the proper quantity of data to send in order to prevent congestion and bottlenecks. This wastes bandwidth and causes network delays. BBR addresses this issue by adaptively assessing the available bandwidth (also known as bottleneck bandwidth) along the network channel and calculating the round-trip time (RTT) between the sender and receiver. Although when several flows compete for bandwidth, BBR may supply more bandwidth to one flow at the expense of another, resulting in unequal resource distribution. This paper proposes to integrate machine learning with BBR to enhance fairness in resource allocation. This novel strategy can improve bandwidth allocation and provide a more equal distribution of resources among competing flows by using historical BBR data to train an ML model. Further we also implemented a classifier model that is graphic neural network in the congestion control method.

翻译：谷歌提出的BBR（瓶颈带宽与往返传播时间）方法用于增强互联网网络传输，其设计初衷是高效处理海量数据。传统TCP（传输控制协议）算法在计算避免拥塞和瓶颈的适当数据发送量时面临最大困难，这导致带宽浪费和网络延迟。BBR通过自适应评估网络信道上的可用带宽（也称为瓶颈带宽）并计算发送端与接收端之间的往返时间（RTT）来解决该问题。然而当多个数据流竞争带宽时，BBR可能以牺牲某个数据流为代价为另一数据流提供更多带宽，造成资源分配不均。本文提出将机器学习与BBR相结合以增强资源分配的公平性。这种新型策略通过利用历史BBR数据训练机器学习模型，能够优化带宽分配并在竞争数据流间实现更均衡的资源分布。此外，我们还实现了一个基于图神经网络的分类器模型用于拥塞控制方法。