What happens when engineering students get a hold of 38 Raspberry Pis

<< Back

Efficient Monitoring of CPS and IoT Systems: A Deployment Guide for Empirical Evaluations.
Jan Körner, Samuel Bach, Albin Karlsson, Linus Sundkvist, Romaric Duvignau

During my studies, I took a course on "Data-driven support for cyber-physical systems," where I, along with three classmates, explored different monitoring algorithms and compared their performance. To carry out our comparisons, we utilized 38 Raspberry Pi devices connected in a LAN, with a master computer acting as the sink and orchestrator, in order to simulate a larger-scale deployment.

We evaluated various monitoring algorithms based on metrics such as CPU utilization, memory usage, network traffic, energy consumption, and monitoring accuracy.

Our work examined the trade-offs between these algorithms, focusing on how factors such as energy efficiency and processing power influenced their performance. Additionally, we developed a deployment guide to assist others in reproducing similar experiments.

The findings from this work were published in the conference paper titled "Efficient Monitoring of CPS and IoT Systems: A Deployment Guide for Empirical Evaluations."

The setup

img

TODO

Topology

img

TODO

How Deplyment worked

img

TODO