France-based computing solutions provider ViFiB has deployed a new resource allocation algorithm that reportedly reduces energy usage of cloud computing by 30%.
According to the company, thanks to this new algorithm, ViFiB net power usage effectiveness (PUE) reached a new low of 20%.
PUE is a measure of how efficiently a computer data centre uses energy; specifically, how much energy is used by the computing equipment, in contrast to cooling and other equipment. The PUE metric is widely believed to be the most popular method of calculating energy efficiency.
A recent study by consultancy firm Accenture found, for large deployments, cloud solutions can reduce energy use and carbon emissions by more than 30%. It adds the benefits are even more impressive for small deployments, as energy use and emissions can be reduced by more than 90% with a shared cloud service.
Accenture notes cloud computing reduces wasted computing resources through better matching of server capacity with actual demand, and flattens relative peak loads by serving large numbers of organisations and users on a shared infrastructure.
Cloud enables organisations to operate servers at higher utilisation rates, Accenture states, adding it also utilises advanced data centre infrastructure designs that reduce power loss through improved cooling as well as power conditioning.
Jean-Paul Smets, CEO of ViFiB, explains PUE measures the efficiency of energy usage in data centres by comparing how much power is used by computers in an IT facility and how much net power is used by the IT facility as a whole.
"In traditional data centres, a lot of energy is wasted on cooling, resulting in PUE of 150% or more. By re-using heat generated by computers for household heating and using natural cooling, some data centres with a PUE of 150% achieve net PUE of less than 50%," says Smets.
Rafael Monnerat, ViFiB COO, adds: "ViFiB innovation consists of reducing net PUE by allocating cloud computing resources in priority to, first, computers that contribute to household heating; second, to computers with loads greater than 25%; and third to computers with loads less than 25%.
"This simple allocation process optimises power efficiency by leveraging the non-linear relationship between computer load and power usage. The more a computer is used, the less additional energy it requires to be further used."
Monnerat continues: "In other words, running two processes of 40% load each in the same computer consumes less energy than running each process alone in separate computers."
Share