Title: Minimizing energy consumption of real-time workflows in failure-prone environments
Directors: Frédéric VIVIEN and HAN Li
Discipline: Computer Science
Status: Ongoing Project
Starting date: 2022
Directors
Summary
Low energy consumption and high reliability are widely identified as increasingly relevant issues in real-time systems. The tasks are replicated to ensure a prescribed reliability threshold, which leads to higher energy consumption. Furthermore, energy management methods may conflict with the deadlines and/or reliability goal. Finally, real-time workflows arising from complex applications typically involve many precedence constraints between tasks, which makes the problem even harder. This project aims at designing multi-criteria optimization strategies for real-time workflows to minimize the expected energy consumption while enforcing the reliability threshold and meeting all task deadlines.