UC San Diego

Energy-efficiency is a key concern in continuously-running mobile applications, such as those for health and context monitoring. Unfortunately, developers must implement complex and customized power-management policies for each application. Not only does this require a developer to have a strong understanding of hardware and how various operations impact resource usage, but it involves the use of complex primitives and writing error-prone multithreaded code to monitor hardware state. To address this problem, this dissertation presents Annotated Programming for Energy-efficiency (APE), an annotation language and middleware service that eases the development of energy-efficient Android applications. APE annotations are used to demarcate a power-hungry code segment whose execution is deferred until the device enters a state that minimizes the cost of that operation. The execution of power-hungry operations is coordinated across applications by the APE middleware. The APE Policy Generation Engine can automatically identify operations in an application that utilize power-hungry resources, generate an APE-based power-management policy, and provide feedback to the developer regarding available options for policy customization. Various language constructs and static program analyses allow a developer to specify constraints for delay-sensitive operations and data in the application, while runtime support ensures that specified constraints are satisfied. Several case studies of applying APE to real mobile sensing applications demonstrate the expressive power of the APE approach and show that annotations can cleanly specify a power management policy and reduce the complexity of its implementation. An empirical evaluation of the middleware shows that APE introduces negligible overhead and equals hand-tuned code in energy savings