Characterizing and Optimizing Emerging Mobile Applications From the System Perspective

Thumbnail Image

Persistent link to this item

View Statistics

Journal Title

Journal ISSN

Volume Title


Characterizing and Optimizing Emerging Mobile Applications From the System Perspective

Published Date




Thesis or Dissertation


The past few years have witnessed exciting mobile technologies advances: high-speed LTE access has become a norm, mobile devices are unprecedentedly powerful, VR/AR has eventually stepped out of the lab, wearable devices are waving computing into our daily lives, etc. Although the above remarkable achievements have boosted the emergence of various novel mobile applications, there still remains three major challenges for researchers, developers as well as the mass mobile users. First, emerging mobile applications are still relatively new to the commercial market, and the research community lacks a thorough understanding of their ecosystem. Second, emerging mobile applications pose stringent requirements for computing and networking capability, leaving their quality of experience (QoE) suffers in the presence of constrained mobile CPU/GPU and challenging network conditions. Third, emerging mobile applications are operating under tight battery budget, however they often incur non-trivial energy overhead. My dissertation is dedicated to address above challenges identified in three representative emerging mobile applications: 360-degree live video, untethered Virtual Reality (VR), and wearable applications. The goal is to uncover their unique characteristics by comprehensive measurements, propose optimizations towards their major inefficiencies, and implement empirical systems to improve the QoE. Specifically, we designed and implemented LIME (LIve video MEasurement platform), a generic and holistic system allowing researchers to conduct crowd-sourced measurements with 360-degree live video streaming. We utilized LIME to perform a first study of personalized 360-degree live video streaming on popular commercial platforms. LIME revealed significant network utilization inefficiencies across different streaming platforms that were previously ignored. Our findings suggested that bringing viewport-adaptive streaming into 360-degree live video could significantly improve viewer side QoE in the presence of challenging network conditions. Inspired by the 360-degree video measurement study, we further studied generic VR. We designed, implemented, and evaluated an untethered multi-user VR system, called Firefly, that supports more than 10 users to simultaneously enjoy high-quality monoscopic VR content using a single commodity server, a single WiFi access point (AP), and commercial off-the-shelf (COTS) mobile devices. Firefly employed a series of innovations to adapt to the stringent networking and rendering requirements of supporting high quality VR on commodity smartphones. Our prototype of Firefly demonstrated, for the first time, the feasibility of supporting 15 mobile VR users at 60 FPS using COTS smartphones and a single AP/server. We also performed in-depth characterization of emerging wearable applications. In particular. We conducted an IRB-approved measurement study involving 27 Android smartwatch users. Using a 106-day dataset collected from our participants, we performed in-depth characterization of three key aspects of smartwatch usage in the wild: usage patterns, energy consumption, and network traffic. Our findings revealed root causes for wearables’ shortened battery life and we proposed optimizations to further improve the energy efficiency of smartwatches.


University of Minnesota Ph.D. dissertation. August 2020. Major: Computer Science. Advisor: Feng Qian. 1 computer file (PDF); xi, 130 pages.

Related to




Series/Report Number

Funding information

Isbn identifier

Doi identifier

Previously Published Citation

Suggested citation

Liu, Xing. (2020). Characterizing and Optimizing Emerging Mobile Applications From the System Perspective. Retrieved from the University Digital Conservancy,

Content distributed via the University Digital Conservancy may be subject to additional license and use restrictions applied by the depositor. By using these files, users agree to the Terms of Use. Materials in the UDC may contain content that is disturbing and/or harmful. For more information, please see our statement on harmful content in digital repositories.