An Energy Efficient Scheduling Manager for Cloud-Based Mobile Navigation Applications

2. Background

Pathak et al. (2012) presented an energy profiler for applications on modern smartphones which is called ‘eprof’. Its architecture has three components: code instrumentation and logging, power modeling and energy accounting, and profile presentation. Firstly, eprof needs to collect processes, threads, subroutines and application system calls. Secondly it tracks logs and draws back energy activities to smartphone hardware components. Thirdly, it draws the energy activities which matched the entities responsible for the activities. Then eprof was used in case studies to understand the energy spent in mobile applications. A result of using eprof shows that free applications use 65%-75% of energy in modules for third-party advertisement.

The measurement study conducted in (Pathak et al. 2012) has found that the tail energy phenomenon in wireless communications causes the 3G, GSM or WiFi-based devices to consume a large amount of energy. Tail energy is energy spent to keep a network interface in the same power state after a transfer is completed. However, if a data transfer occurs within tail time, there will be no energy spent. Therefore, TailEnder was developed based on a measurement study to reduce energy consumption while still meeting user expectations. With E-mail applications, for example, a data transfer is scheduled by TailEnder to minimize tail energy.