As UK and US smart power grids find crowd sourced energy storage  a potential route to flexibility, the Cloud Data community are now considering savings that could be reaped if servers were to take over the dual role of house or office furnace as well as data cruncher say researchers Jie Liu, Michel Goraczko, Sean James, Christian Belady (Microsoft Research) with Jiakang Lu, Kamin Whitehouse
(Computer Science Department, University of Virginia) in their paper "The Data Furnace: Heating Up with Cloud Computing."

From the home owner’s perspective, reports Randall Stross in the New York Times, a DF is equivalent to a typical heating system: a metal cabinet is shipped to the home and added to the ductwork or hot water pipes. Technically DFs create new opportunities for both lower cost and improved quality of service, if cloud computing applications can exploit the differences in the cost structure and resource profile between Data Furances and conventional data centres.

The conventional data centre must invest about $400 a year to run each server, or about $16,000 for a cabinet filled with 40 of them calculate researchers. (This includes the costs of building a bricks-and-mortar centre as well as the cost of cooling the machines.)

Having homes host the units could reduce the need for a company to build new data centres. The company’s cost to operate the same cabinet in a home would be less than $3,600 a year — leaving a smaller carbon footprint, too. The company Data Centre would thus cover a homeowner’s electricity costs for servers,  and still come out a good way ahead financially.

Machines would remain under remote control of the company’s centralised data centre, and their workings remain opaque. Network traffic and data would have to be encrypted. Sensors would warn if the cabinet was opened. If a server failed, its tasks would automatically reassign to another as software is built with the expectation that an individual machine can break at any time.

The Data Furnace would be best suited for computing tasks that aren’t time-sensitive and can be broken into chunks performed by thousands of machines — say, for scientific research.

It's an idea that waits for one big-name Internet company to experiment, allocating prospective usersenough financial incentive to have servers replace basement furnaces.

Co-author (right) Kamin Whitehouse, assistant professor of computer science at the University of Virginia said there had been "a very strong response" to the paper."

" We heard from several people who are already heating their homes with computer systems, which shows that it works. Our contribution is to show that the data furnace also has lower cost and lower energy than the conventional data centre.”