Resource Management on Mobile Platforms

Smart phones are becoming increasingly powerful and popular. However, mobile phones are resource constrained embedded devices - with scare resources being compute capability, energy (battery), memory, and network bandwidth. With the increasing usage of smartphones, the demands on these scarce resources become more severe and could be the Achilles’ heel impeding further progress. In this project, our goal is to develop a resource management framework for smart phones to maximize user experience. We address two unique issues:

- Prioritize resource usage, e.g., background data synchronization should not jeopardize future voice communication, and helping another device should not jeopardize its own usage.

- The explicit usage of user profile, i.e., resource management should not be one-size-fit-all.

As the preliminary step, we developed an optimization scheme that dynamically determines the frequency of background data synchronization (e.g., email) based on the remaining energy on the phone, recharging time, and user profile (i.e., how often the user has phone calls and how long they last). Because of the dynamic nature of the adaptive synchronization, the phone performs more synchronization in days with light voice traffic and receives more phone calls in days with heavy voice traffic, compared to the static synchronization scheme.

We also developed PACE: power aware collaborative execution for mobile phones. Collaborative execution is a way to pool resources together to give the abstraction to the end user as if the application is running on the local phone. The pool of resources could be garnered from other phones or from the cloud (or a local server). For collaborative execution to be practical it is important that the helper phone is protected from jeopardizing its battery life for its own usage until its recharging time. To address this issue, we develop a Markov Decision Process (MDP) framework that enables collaborative execution while protecting the interests of the helper phone and taking into account the user call profiles and preferences such as when they expect to recharge their phones.

Graduate students:

Eric Jung (ECE)

Frank Maker III (ECE)

Yichuan Wang (CS)

Undergraduate students:

Sophia Tang

Early contributors:

Tang Lung Cheung (CS)

Kari Okamoto (CS)

Iuri Prilepov

Jonathan Crussell

Publications:

UPDATE: User-Profile-Driven Adaptive Transfer, [pdf] 

Yichuan Wang and Xin Liu.

PACE yourself – Power Aware Collaborative Execution on Mobile Phones. Under submission. [pdf]
Eric Jung, Yichuan Wang, Iuri Preilpov, Frank Maker, Xin Liu, and Venkatesh Akella.

Markov Decision Process (MDP) Framework for Optimizing Software on Mobile Phones. Tang Lung Cheung, Kari Okamoto, Frank Maker, Xin Liu, and Venkatesh Akella, the International Conference on Embedded Software, Oct. 12-16, Grenoble, France. [pdf]