U-M and AT&T researchers develop energy efficiency profiling technology for mobile platforms

A team of researchers at the University of Michigan and AT&T Labs Research, including Assoc. Prof. Z. Morley Mao, PhD student Feng Qian, and PhD student Zhaoguang Wang, has completed a research project aimed at identifying, quantifying, and remediating issues related to the development of energy efficient applications for mobile platforms. The U-M researchers are now launching a study under which CoE students can apply to receive an AT&T 3G phone and wireless plan for free for 6 months so that the researchers can collect data on the phone’s performance.

battery icon Enlarge
Why do some apps sip power while others burn it?

About the Research

Because mobile platforms are a relatively new device type and are architecturally different from traditional laptop and desktop computers in fundamental ways, it is common to find that application developers, device makers, and radio network engineers are knowledgeable about their own design domains but not the others. This explains why users experience instances in which apps that perform similar functions often produce different results in terms of battery usage and page rendering performance.

To resolve these issues, the researchers undertook an in-depth, comprehensive investigation of mobile device wireless communications protocols and infrastructure, ultimately discovering the source of the problems to be in the complex interactions that take place between the device and the cellular network. These interactions are especially hard to see, given the layered nature of the network architecture that intentionally hides lower-level protocols from developers working in the application layer.

Based on their findings, the researchers built a new technology that makes those interactions visible, allowing other researchers and designers to diagnose the specific inefficiencies of individual apps. One popular app was found to be using 40% of its power consumption to transmit 0.2% of its data. In discovering these inefficiencies and their causes, the researchers also identified solutions that turn out to be surprisingly easy to implement.

The app profiling technology, which has already been shared with some app developers, is being productized by AT&T and is currently in prototype form. Anyone wishing to learn more can contact AT&T through their Developer Program Contact form. In particular, AT&T Research is starting to contact developers of popular apps, including Facebook and Pandora. The feedback has been encouragingly positive as the provided technique greatly helps developers identify resource usage inefficiencies and improve their apps.

Students Sought for Usage Study at U-M

The research team is now taking a further step by launching a user study at U-M. Under the study, 20 participating students will each be provided with a 3G smartphone with a wireless data plan at no cost. Smartphone usage data will be collected as the participants use the phones over a six-months period, and will be analyzed to better understand the complex interaction among the user activity, application behavior, and the wireless network. Interested U-M CoE students can apply here to participate in the user study.