Fides at PDP 2026: dstributed AI for energy monitoring

At the Euromicro PDP 2026 international conference, focused on distributed systems, parallel computing, and network technologies, Fides presented the paper "Exploring Distributed Energy Learning: A Federated, Self-Adapting Framework for Sensor-Free Power Estimation at Scale".

Developed within the ERA – Explore, Redesign, Accelerate research project, the study introduces MIRU, a framework based on Federated Learning and Long Short-Term Memory (LSTM) designed to estimate the energy consumption of PCs and servers without the need for physical sensors.

The solution aims to make energy monitoring more efficient, scalable, and sustainable, reducing dependence on dedicated hardware while enabling smarter management of IT infrastructures.

The project was developed in collaboration with Luiss Guido Carli University, Prof. Fabio Angeletti, Secure Network, and Time Vision.

Fides' participation in PDP 2026 reinforces its commitment to applied research and the development of innovative solutions that drive efficiency, scalability, and sustainability in modern IT environments.