Power to the computers

Author: John Monczunski

The era of all-pervasive computing is rapidly approaching—some would argue it already has arrived—in which countless computers run unnoticed in the background of everyday life, making it more convenient and manageable.

In the not-too-distant future, for instance, clothing may have sensors embedded for medical monitoring. Automobiles may be equipped with computing systems that find the timeliest route to your destination, steering you there based on traffic patterns monitored by remote sensors.

Driven by such trends plus the explosion of mobile computing and smart phones, it’s predicted the number of wireless devices will exceed wired devices by 2010, Christian Poellabauer notes. To ease the coming strain on battery technology from that increase in wireless computing, Poellabauer and other computer scientists are devising strategies to more efficiently use battery power.

If your computing device senses the task it is asked to do would deplete its battery too quickly, the ND engineering professor says it may engage in “cyberforaging,” enlisting other nearby devices to complete the task. Another power-saving technique already being used, Poellabauer says, is to slow down the processor in your computer for tasks that do not require the highest speed.

“In general, we are looking at each resource in a device, such as the processor memory, storage and networking, and trying to see what techniques we can apply to more carefully manage that resource,” he says.

One of the main thrusts of Poellabauer’s current research is attempting to integrate various individual power-saving techniques. “We have all these individual solutions and you would expect a synergistic effect, but actually energy consumption can be worse because they contradict each other. The processor might be trying to save energy, but consequently the network card must consume more to satisfy the processor.”

To solve the problem Poellabauer and his colleagues are attempting to write computer code for a “super scheduler” that would coordinate the various power-saving strategies.

John Monczunski is an associate editor of this magazine.