Observing energy in software distributed ecosystems

Overview & Synopsis

The major transformations in our societies and industries are opening the way for the expansion of distributed computing through cyber-physical systems. These systems include sensing, computation, control and networking capabilities between the physical and the cyber worlds. However, measuring and monitoring their energy consumption is challenging, as these ecosystems are inherently heterogeneous: from the large selection of hardware components, devices and architectures, to the important variability in software versions and configurations, and to the various human actors involved. Monitoring such systems with their heterogenous nature requires building approaches and models to measure energy at various granularity levels (from hardware to software source code), and provide these measurements to the various actors (developers, end users). We present our vision on the major transformations and challenges towards modeling energy of connected devices, hardware components, software, and source code.