TOWARDS ENERGY-EFFICIENT DATA TRANSMISSION ON MOBILE DEVICES

The Hong Kong University of Science and Technology
Department of Computer Science and Engineering


PhD Thesis Defence

Title: "TOWARDS ENERGY-EFFICIENT DATA TRANSMISSION ON MOBILE DEVICES"

By

Mr. Yi ZHANG


Abstract

The evolving mobile network, which provides fast and flexible access to 
the Internet, has greatly changed our daily life. Various Apps and 
services are now supported on mobile devices and thus, facilitates our 
research, work and so on. However, constrained to the current 
technologies, the energy-inefficient mobile data transmission often 
hinders people from fully enjoying modern mobile techniques. In order to 
address this issue, in this thesis, we aim to solve some tightly-related 
challenges to achieve energy-efficient mobile data transmission and thus, 
narrow down the gap between our daily life and fast evolving mobile 
technologies.

First, we study the energy issues of uncontrolled App network activities 
in mobile phones. As more and more applications being installed, their 
competition for network resources incurs serious problems to battery life 
and thus, degrade users’ normal experiences. To address this issue, we 
make comprehensive measurements on users habit and propose a novel 
approach to orchestrate network activities of smartphone applications, 
based on users habit. The proposed algorithm is proven to approximates the 
online optimal solution with a competitive ratio of (1−epsilon)/2. We 
implement the algorithm as a middleware service and it achieves over 70% 
energy savings in network activities.

Then we move our focus onto a more essential factor, the variation of WiFi 
link quality. Nowadays, offloading mobile traffic from cellular to WiFi is 
widely recognized as a viable solution to improve the energy efficiency on 
mobile devices. However, through extensive field experiments, we find WiFi 
offloading is not always energy efficient and even consumes more energy 
than cellular network due to link quality variation. On the other hand, we 
observe from our past experiences that practical data transmission 
deadline requirement and link utilization allows scheduling of data 
traffic to time periods with good link quality. Accordingly, we propose 
Q-offload, the first attempt towards energy efficient WiFi offloading with 
link dynamics. In Q-offload, we propose an iterative framework to achieve 
energy efficient WiFi offloading by exploiting good link quality while not 
affecting user experience. The results from extensive experiments show 
that Q-offload can achieve 33.5%∼55.7% energy efficiency improvement, 
compared with state-of-the-arts under different conditions.

Enlightened by the results of our prior works, we raise our focus onto 
users’ lifestyles and this constitutes our third work. In this work, we 
propose a context-based link quality estimation scheme for achieving 
energy-efficient data transmission in WiFi/mobile networks. Through 
analyzing millions of user traces, we exploit that link quality is highly 
relevant with users’ lifestyle and can be extracted as fingerprints. 
Following this observation, we propose an individual-oriented system, 
Furion, for exploiting beneficial WiFi/mobile links based on users’ 
contexts. In Furion, we introduce a context-based link quality 
discrimination scheme and design a more practical probabilistic model to 
predict the energy efficiency of links. As a result, the accuracy of link 
quality estimation is further improved given limited hardware on mobile 
devices and it can also be extended to different environments. The 
prototype of Furion is implemented on Android platform and the results 
demonstrate that Furion achieves a significant performance improve 
compared with the state-of-the-arts.


Date:			Thursday, 6 April 2017

Time:			2:30pm - 4:30pm

Venue:			Room 4472
 			Lifts 25/26

Chairman:		Prof. Jie Yuan (ECE)

Committee Members:	Prof. Bo Li (Supervisor)
 			Prof. Kai Chen
 			Prof. Wei Wang
 			Prof. Chin-Tau Lea (ECE)
 			Prof. Xiaohua Jia (Comp. Sci., City U)


**** ALL are Welcome ****