Towards End-to-End QoE Guarantees for Timely Video Delivery in Adversarial Multi-Hop Battlefield Networks
Abstract
Major Goals: The objective of this project is to establish a new framework for developing network algorithms that ensure both end-to-end delay and end-to-end QoE for video delivery in battlefield networks in the presence of event-driven IoT applications. Our framework will be built upon two critical components: One is an analytical model that precisely calculates a range of QoE metrics, such as the frequency and duration of video interruptions and video quality, based on the traces of packet deliveries. This enables us to translate the perceived video quality, which is a subjective matter, into measurable, and hence optimizable, network performance metrics. The other is a network model that explicitly addresses several important features of battlefield networks, including unreliable wireless transmissions, multi-hop transmissions, end-to-end delay bounds, and unpredictable, adversarial even, packet generations from event-driven IoT applications. By combining these two components, our framework can serve as the foundation to analyze the actual QoE performance of network algorithms under realistic battlefield networks. Based on our framework, we will develop a rich suite of network algorithms, including packet scheduling, multi-hop routing, and dynamic video rate control, that optimize end-to-end QoE under end-to-end delay constraints. Our algorithms will jointly consider video flows and event-driven IoT applications, and offer provably optimal QoE performance for video flows without sacrificing the reliability of event-driven IoT applications. We will also study implementation issues of our algorithms within the framework of software-defined networking (SDN).
Document Details
- Document Type
- Technical Report
- Publication Date
- Apr 19, 2023
- Accession Number
- AD1225029
Entities
People
- I-Hong Hou
Organizations
- Texas Engineering Experiment Station