Development of User-Centric HTTP Adaptive Video Streaming Technique

Abstract

User-perceived Quality of Experience (QoE) is vital in Internet video applications, influencing revenue for content providers and delivery systems. Due to limited network support for QoE optimization, bottlenecks may arise anywhere in the delivery system. Consequently, a robust bitrate adaptation algorithm in client-side players is critical to ensure good user experience. This study proposes a client-side buffer control model for mobile video streaming to enhance QoE under dynamic network conditions, supporting Dynamic Adaptive Streaming over HTTP (DASH). The model employs a Double Leaky Bucket (DLB) traffic shaping scheme with a First-In- First-Out scheduler. A DLB-based algorithm was developed in C++ and simulated using Network Simulator 3 (NS-3) with a trace-based dataset of a 4420-second video comprising 221 segments across 8 representations. Simulation parameters included segment duration, buffer size, packet size, and number of packets. Performance was evaluated using Packet Loss, Jitter, and Throughput metrics, comparing the DLB model to the existing Single Leaky Bucket (SLB) approach. Results showed DLB reduced packet loss from 60.8% to 39.2%, jitter from 53.4% to 46.6%, and increased throughput from 33.7% to 66.3%. These improvements demonstrate that the DLB algorithm significantly enhances mobile video streaming QoE by ensuring smoother playback under fluctuating network conditions.