Researchers at the Lawrence Berkeley National Laboratory have come up with a model designed to show the full extent of cloud’s energy efficiency.
The report, put together by a research team comprising McCormick School of Engineering Northwestern University and Lawrence Berkeley and part funded by Google, unveils the CLEER (Cloud Energy and Emissions Research) model, described as the ‘first ever open-access, fully transparent systems model for energy analysis of cloud systems by the research community’.
What’s more, the model predicts an 87% reduction in primary energy use if companies moved their business software – comprising CRM, productivity and email tools – to the cloud.
The report rightly notes that not all modelling attempts are 100% accurate, but adds: “Despite uncertainties, the energy savings potential of cloud-based software is likely to be substantial on a national scale given the vast differences between the energy efficiencies of local and cloud data centres.”
It’s worth noting that part of these ‘uncertainties’ will revolve around the researchers only having access to publicly available data from a variety of sources – the word ‘estimate’ in its various guises was used 126 times in the report.
Yet the methodology seems more than sound. In this US case study, the researchers worked out that 86.7m of US workers used computers, and of that number, all used email, whilst 58m used productivity tools and 8m CRM software.
Using that data, the academics then analysed present day software data centre characteristics, client IT device characteristics, current data centre and network power use, current cloud-based software data centre characteristics, before comparing the two.
The overall results, particularly the 87% energy decrease, are shown helpfully here. In specific terms, it will mean a 47 petajoule (PJ) output per year in the US, as opposed to 373 PJ in a traditional data centre: