Zoom校园遥测数据集和WebRTC-网络5G遥测数据集
收藏资源简介:
本文研究的两个数据集分别是校园范围内的低速率Zoom遥测数据集,记录了长达数日的数据,以及高速度、相关性的WebRTC-网络5G遥测数据集,记录了数小时的数据。这些数据集通过详细测量私有CBRS和商业运营商蜂窝网络的动态,捕捉物理层和链路层事件,并将它们与网络层和传输层的效果以及视频会议应用本身的效果相关联。研究团队基于这些数据,追溯性能异常的根源,并构建了自动化工具Domino,用于识别导致5G视频会议性能下降的因果事件链。
The two datasets investigated in this study are, respectively, a campus-wide low-rate Zoom telemetry dataset collected over several consecutive days, and a high-speed and correlated WebRTC-network 5G telemetry dataset collected over several hours. These datasets capture physical-layer and link-layer events through detailed measurements of the dynamics of private Citizens Broadband Radio Service (CBRS) and commercial carrier cellular networks, and correlate these events with performance impacts at the network layer, transport layer, and the video conferencing application itself. Leveraging these datasets, the research team traced the root causes of performance anomalies and developed an automated tool named Domino, which is designed to identify causal event chains that lead to degraded performance of 5G video conferencing.




