Doctoral Dissertation Proposal, Chinmaey Shende

Title: Data Saving and QoE Improvement for Adaptive Bitrate Streaming in Cellular Networks

Ph.D. Candidate: Chinmaey Shende

Major Advisor: Dr. Bing Wang
Associate Advisors:  Dr. Song Han, Dr. Minmie Wang
Committee Members:  Dr. Wei Zhang, Dr. Jerry Zhijie Shi

Date/Time: Monday, February 26th, 2024, 10:00 am- 11:00 am

Location: HBL Instruction 1102

WebEx:

https://uconn-cmr.webex.com/uconn-cmr/j.php?MTID=m83549450b0d02d8c2045c5c7c4e6c55f

Monday, February 26, 2024 10:00 AM | 1 hour | (UTC-05:00) Eastern Time (US & Canada)

Meeting number: 2632 764 9375
Password: HaakjJiT288

Abstract

Adaptive bitrate (ABR) streaming is the current de facto video streaming technology in the Internet. It accounts for the majority of traffic in cellular networks, and also places a heavy demand on users’ limited monthly cellular data budgets. Despite much research on ABR streaming in cellular networks, reducing data usage and improving quality of experience (QoE) remain a significant challenge. In this dissertation, we explore two directions in resolving the challenge: (1) developing quality-aware ABR streaming strategies guided by data budget to reduce data usage, while minimizing the impact on QoE, and (2) developing accurate cross-layer bandwidth estimation for low-latency live (LLL) ABR streaming to improve QoE for end users.

In the first part of the dissertation, we develop an approach, DataPlanner, that uses data budget information to better manage the data usage of mobile video streaming, while minimizing the impact on users’ QoE. Specifically, we propose a novel framework for quality-aware ABR streaming involving a per-session data budget constraint. Under the framework, we develop two planning based strategies, one for the case where fine-grained perceptual quality information is known to the planning scheme, and another for the case where such information is not available. Using evaluations under a wide range of network conditions using different videos covering a variety of content types and encodings, we demonstrate that both these strategies use much less data compared to state-of-the-art ABR schemes, while still providing comparable QoE. Our proposed approach is designed to work in conjunction with existing ABR streaming workflows, enabling ease of adoption.

LLL ABR streaming relies critically on accurate bandwidth estimation. While existing studies have proposed bandwidth estimation techniques for LLL ABR streaming, these approaches are at the application level, and their accuracy is limited by the distorted timing information observed at the application level. We propose a new cross-layer approach that combines application-level semantics and fine-grained kernel-level packet capture to achieve higher accuracy in bandwidth estimation. The more accurate bandwidth estimation can lead to better bandwidth prediction and significantly better QoE for end users.