Energy-Efficient Distributed Computing – ENEDISC

In recent years, the scientific field of distributed computing has been subject to a revolutionary discovery that goes against the conventional wisdom that saving energy means slowing down calculations: it is actually possible to design algorithms that are both time and energy efficient. In a nutshell, the project ENEDISC is aiming at exploiting this discovery for developing a consistent theory of energy-efficient distributed computing.
The technological advances in network architectures, especially those related to the ability of modern computing devices to turn on and off their communication interfaces at will, to shut down for a prescribed amount of time, or even to wake up other nodes in the system thanks to specific signals, have only been recently considered by the distributed computing community. Tremendous results have already emerged from this recent consideration, which demonstrate the huge potential benefit of this technology in the design of “frugal” algorithms. As a typical conceptual example, Maximal Independent Set (MIS) is arguably one of the most important symmetry-breaking problems in distributed network computing. A series of recent papers, led to the conclusion that MIS can be solved in almost as fast as the fastest known distributed algorithms, but with a significantly smaller number of awaken rounds for each node. Each node is idle almost all time, yet MIS can be computed efficiently!
These results change our perspective on distributed computing, at a time when eco-sustainability has become one of, if not the main concern of mankind. The new perspective is that energy-saving and time-efficiency are not orthogonal, but may well go hand in hand. The objective of ENEDISC is therefore to revisit some of the main achievements of distributed computing by focusing primarily on the energy-complexity of the algorithms, but not giving up on efficiency.