University project aims to limit power consumption

Researchers at the University of Sydney have developed a new scheduling algorithm designed to reduce data centre energy consumption without disrupting operations.

The Energy Conscious Scheduling algorithm (ECS) has been patented by Young Choon Lee and Albert Zomaya at the university's Centre for Distributed and High Performance Computing.

Lee and Zomaya are now developing an ECS prototype, with a view to commercialising the research by late 2010.

Zomaya said the ECS software will be a suite of algorithms (written in C and C++) acting as 'middleware' that can see the operating system and hardware and then decide what to do with different tasks.

"In doing so it makes sure whatever decisions are made are energy-conscious," he said, adding the software stack still "gives you what you want".

"The hope is that you don't have to alter your hardware, but use a middleware layer that is portable. We want it to be as seamless as possible and it could eventually be integrated with the operating system."

According to the university, power consumption by servers has more than doubled since 2000, with estimates of electricity use for servers worldwide costing about £4.4bn in 2005.

Lee and Zomaya have tested ECS with a number of benchmarks that "reflect the behaviour of many applications".

"For example, we have used ECS for solving equations and other computing-intensive applications. They are drawn from real-world examples," Zomaya said.

The team is hoping to have the prototype ready for early next year with a product available by the end of 2010.

Zomaya said the growing popularity of cloud computing will help IT managers reduce data centre energy costs, but more research needs to be done before large organisations can fully trust their "precious data" to cloud technology. "ECS will work well on cloud systems as they mature, but will give a more immediate means for organisations to reduce their data centre's energy consumption," he said.

ECS uses a processor's dynamic voltage scaling (DVS) capability to map computational tasks to minimise completion time and energy use.

"Computations are typically comprised of interdependent tasks, so the need to wait for a parent task to complete can create slack and therefore wastage," Zomaya said.

"When ECS is employed with the help of DVS capability, mapping decisions between processors, supply voltages, and tasks are streamlined to significantly lower the amount of energy required at any given time."

The researchers claim ECS can more than halve the energy consumption of processors in data centres at "little or no operational cost".

"This reduces slack and creates energy savings of between 10 to 60 per cent," Zomaya said.

The researchers are also looking at situations where ECS can be used to manage energy consumption by storage and networking units for a complete data centre solution. "Bringing this into the model will make it more comprehensive," Zomaya said.