The ERODS team of University of Grenoble (UGA) will contribute to the management of Edge-to-Cloud Continua by designing and implementing solutions for the detection and the prediction of anomalies at the hardware and at the software level. The goal is to help improving the reliability of such platforms. By anomalies, we refer to crashes as well as to performance bugs. Those are especially challenging to detect and predict in Edge to Cloud continua because of the scale of the infrastructures and of the heterogeneity of the hardware.
In the context of the project, we will study how data coming from different levels of the software stack can help better detecting anomalies. We will also study how to build accurate tools while limiting the amount of collected data and the amount of communication between distributed entities.