Since the global COVID-19 pandemic, a shift has occurred to remote education and work. This is enabled by various video conferencing systems like Zoom, Google Meet, Microsoft Teams which has consequently increased the pressure on digital infrastructure. Surprisingly, no prior research has been undertaken to understand the performance of various video conferencing systems.In this study, we aim to evaluate and compare the performance of video conferencing systems in the context of distributed systems. To achieve this objective, we designed experiments to compare the performance and scalability of Zoom, Microsoft Teams, and Jitsi. Furthermore, we designed and implemented an experimentation tool that automatically hosts and joins video conferences, and measures the network bandwidth, CPU usage, and memory usage.We observed that the video conferencing tools have a stable performance after the initial warmup time for up to 6 clients. Except for the high memory usage of Zoom web, which grows linearly in time until a certain upper bound. Hence, we do not recommend Zoom web for a small number of clients.

No abstract

No abstract