E-Loupe: End-to-End Energy Management for Mobile Devices

Energy drain in mobile devices is well recognized to be a serious problem. One solution is to provide tools and guidelines to enable application writers build more energy efficient programs. This project explores an alternative that mitigates the ill-effects of an energy hungry application. Our system, E-Loupe, offers a finer-grained approach to ensure predictable energy drain in mobile devices.

Our goal:

The goal of our system is to provide fine-grained, on-demand control over energy hungry apps to ensure predictable energy drain.  If a user expects his mobile device's battery to last say 7 days, E-Loupe ensures that the expectation is met without significantly compromising the device's functionality.

E-Loupe collects coarse grained battery consumption data from mobile devices, analyzes them in the cloud to determine the likely culprit process, or group of processes, and then limits the energy drain caused by these processes -- thereby preventing unexpected spikes in energy drain.

We use a technique called energy sandboxing to control when energy hungry apps are scheduled, and how much resources they are allowed to consume. This is unlike current mobile operating systems that provide relatively coarse-grained control over apps.

System Architecture:

 E-Loupe has three components:

  • The first component runs on each user's device and collects energy and resource usage samples.  The collected data is uploaded to a service in a datacenter.
  • A second component running in the data center uses statistical inferencing techniques to diagnose the cause of an energy anomaly, and determines kernel policies to isolate and contain it.  Statistical analyses of energy data from a largepopulation yields sign cant insights.
  • A final component, running in the device kernel implements the policies to limit the impact of the errant app by restricting its resource usage to meet a desired energy goal.

Early Results:

We implemented our system for Windows 8 mobile devices. Using data collected from over 73,000 devices, we show that E-Loupe can successfully isolate the causes of high energy drain for 87.5-92.3% reported energy spikes. The energy sandboxing mechanism is able to reduce the average power consumption by 5-6 times from the peak. Finally, our trace based simulation suggests that E-Loupe can reduce the energy discharge rate by 3 times for real user reported energy spikes.