• Synopsis
    • CAREER: Improving Mobile Video Delivery for Emerging Contents and Networks (NSF Award 1915122)
    Streaming videos wirelessly on mobile devices is an increasingly important application. The objective of this project is to bring innovations to mobile video delivery for new content types and over emerging networks. Specifically, the project investigates three aspects: (1) 360-degree immersive video delivery, (2) video streaming over multiple network paths (multipath), and (3) video streaming over millimeter-wave (mmWave) links. These are expected to be the key building blocks of next-generation video streaming services. First, 360-degree videos provide users with unique panoramic viewing experience; however, 360-degree video content delivery is much more challenging compared to regular videos. Second, multiple network interfaces have become a norm on off-the-shelf mobile devices but their potential is far from being fully exploited. Third, mmWave is a key technology that will be integrated into 5G wireless networks; but adapting video streaming to mmWave largely remains an uncharted territory. The proposed solutions will benefit the society by enhancing the user experience and reducing the resource consumption for next-generation immersive video services. The research will also be integrated with an education plan that seeks to prepare computer science students with the knowledge of new technological trends in networking and systems, and stimulate the general public interest in Science, Technology, Engineering, and Mathematics.
  • Project Publications
    • SIGCOMM21 A Variegated Look at 5G in the Wild: Performance, Power, and QoE Implicationsdownload
    In ACM SIGCOMM 2021.
    WWW21 DeepVista: 16K Panoramic Cinema on Your Mobile Device download
    Wenxiao Zhang, Feng Qian, Bo Han, and Pan Hui.
    In the Web Conference (WWW) 2021.
    IMC20 Lumos5G: Mapping and Predicting Commercial mmWave 5G Throughputdownload
    Arvind Narayanan, Eman Ramadan, Rishabh Mehta, Xinyue Hu, Qingxu Liu, Udhaya Kumar Dayalan, Rostand Fezeu, Saurabh Verma, Peiqi Ji, Tao Li, Feng Qian, and Zhi-Li Zhang.
    In ACM Internet Measurement Conference (IMC) 2020, Pittsburgh, PA.
    ATC20 Firefly: Untethered Multi-user VR for Commodity Mobile Devicesdownload
    Xing Liu, Christina Vlachou, Feng Qian, Chendong Wang, and Kyu-Han Kim.
    In USENIX ATC 2020, Boston, MA.
    MobiSys20 MPBond: Efficient Network-level Collaboration Among Personal Mobile Devicesdownload
    Xiao Zhu, Jiachen Sun, Xumiao Zhang, Yihua Guo, Feng Qian, and Z. Morley Mao.
    In MobiSys 2020, Toronto, Canada.
    CoNEXT19 Analyzing Viewport Prediction Under Different VR Interactionsdownload
    Tan Xu, Bo Han, and Feng Qian.
    In ACM CoNEXT 2019, Orlando, FL.
    WWW20 A First Measurement Study of Commercial mmWave 5G Performance on Smartphonesdownload
    Arvind Narayanan, Eman Ramadan, Jason Carpenter, Qingxu Liu, Yu Liu, Feng Qian, and Zhi-Li Zhang.
    In the Web Conference (WWW) 2020, Taipei, Taiwan.
    MMSys19 LIME: Understanding Commercial 360-degree Live Video Streaming Servicesdownload
    Xing Liu, Bo Han, Feng Qian, and Matteo Varvello.
    To appear in ACM MMSys 2019, Amherst, MA.
    MMSys19 Quality-aware Strategies for Optimizing ABR Video Streaming QoE and Reducing Data Usagedownload
    Yanyuan Qin, Shuai Hao, Krishna Pattipati, Feng Qian, Subhabrata Sen, Bing Wang, and Chaoqun Yue.
    To appear in ACM MMSys 2019, Amherst, MA.
    MobiCom18 (Demo) Demo: Tile-Based Viewport-Adaptive Panoramic Video Streaming on Smartphonesdownload
    Feng Qian, Bo Han, Qingyang Xiao, and Vijay Gopalakrishnan.
    In ACM MobiCom 2018, New Delhi, India.
    MobiCom18 Flare: Practical Viewport-Adaptive 360-Degree Video Streaming for Mobile Devicesdownload
    Feng Qian, Bo Han, Qingyang Xiao, and Vijay Gopalakrishnan.
    In ACM MobiCom 2018, New Delhi, India.
  • Personnel
    • PI: Feng Qian (fengqian at umn.edu)
    Students: Xing Liu (Ph.D.), Yu Liu (Ph.D.), Arvind Narayanan (Ph.D.), Runsheng Ma (undergraduate)
  • Collaborators
    • AT&T Labs Research, HP Labs, University of Connecticut, George Mason University, University of Michigan
  • Data and Code
    • (1) User Head Movement Traces of 360 Videos
    The dataset consists of 5 users' head movement traces when watching 4 360-degree videos. The data was collected by us, and used in our All Things Cellular paper. [Download Dataset, 9.8MB]
    (2) LIME Measurement Data (coming soon)
    The dataset consists of the viewing statistics of 548 paid Amazon Mechanical Turk viewers from 35 countries who have watched more than 4,000 minutes of 360-degree live videos. The data was analyzed in our MMSys 2019 paper.
    (3) 5G Measurement Data
    We release the data collected by us in the WWW20 paper. This data is important in providing a first impression of the world's very first commercial 5G rollouts, and serves as an important baseline of 5G performance. We conduct several experiments to evaluate 5G performance, including but not limited to throughput performance, latency measurements, impact of mobility and obstructions, handoff analysis among many others. Our experiments also illustrate the pros and cons of the different 5G technologies. [Goto download page]
    (4) Lumos5G Data
    We release the data collected by us in the IMC20 paper. The emerging 5G services offer numerous new opportunities for networked applications. In this study, we seek to answer two key questions: (i) is the throughput of mmWave 5G predictable, and (ii) can we build "good" machine learning models for 5G throughput prediction? To this end, we conduct a measurement study of commercial mmWave 5G services in a major U.S. city, focusing on the throughput as perceived by applications running on user equipment (UE). Through extensive experiments and statistical analysis, we identify key UE-side factors that affect 5G performance and quantify to what extent the 5G throughput can be predicted. We then propose Lumos5G -- a composable machine learning (ML) framework that judiciously considers features and their combinations, and apply state-of-the-art ML techniques for making context-aware 5G throughput predictions. Our work can be viewed as a feasibility study for building what we envisage as a dynamic 5G throughput map (akin to Google traffic map) that serves as a fundamental building block for future 5G-aware apps. [Goto download page]
  • Educational Activities
    • Courses Related to the Project:
    Fall 2021: CSCI 8980 Topics in Mobile Computing
    Fall 2020: CSCI 4211 Introduction to Computer Networks
    Fall 2019: CSCI 4211 Introduction to Computer Networks
    Spring 2019: CSCI 8980 Topics in Mobile Computing
  • Outreach Activities
    • Lumos5G ACM IMC 2020 Technical Talk Video
    Firefly USENIX ATC 2020 Technical Talk Video
    MPBond ACM MobiSys 2020 Technical Talk Video
    Flare ACM MobiCom 2018 Technical Talk Video
    Flare ACM MobiCom 2018 Demo Video 1, Video 2