Quantifying Multimedia Streaming Quality: A Practical Analysis using PIE and Flow Queue PIE
Hemendra M. Naik
TL;DR
This paper addresses the QoE of MPEG-DASH streaming under variable network conditions by integrating multipath transport (MPTCP) with two queue-management strategies: PIE and Flow Queue PIE (FQ-PIE). Using NeST, a Linux-based network emulator, and a two-endpoint, multi-homed topology, the authors evaluate MPEG-DASH performance with background traffic across two paths, measuring bitrate stability, throughput, RTT, and buffers. The key contribution is the empirical demonstration that FQ-PIE’s flow isolation markedly improves QoE for DASH over MPTCP, delivering higher throughput and fewer bitrate switches compared to PIE, albeit with some queue-delay oscillations due to multi-queue scheduling. These findings provide practical guidance for deploying flow-isolation-aware AQMs in multipath streaming scenarios and motivate future work under harsher network conditions and broader datasets.
Abstract
The exponential growth of multimedia streaming services over the Internet emphasizes the increasing significance of ensuring a seamless and high-quality streaming experience for users. Dynamic Adaptive Streaming over HTTP (DASH) has emerged as a popular solution for delivering multimedia content over variable network conditions. However, challenges such as network congestion, intermittent packet losses, and varying network load continue to impact the Quality of Experience (QoE) perceived by the users. In this work, the main goal is to evaluate the effectiveness of using queue management and flow isolation techniques in terms of improving the overall QoE for DASH based multimedia streaming applications. Proportional Integral controller Enhanced (PIE) and Flow Queue PIE (FQ-PIE) are used as queue management and flow isolation mechanisms, respectively. The most distinctive aspect of this work is our assessment of QoE for multimedia streaming applications when multipath transport protocols, like Multipath TCP (MPTCP), are employed. Network Stack Tester (NeST), a Python based network emulator built on top of Linux network namespaces, has been used to perform the experiments. The parameters used for evaluating the QoE include bitrate, bitrate switches, throughput, Round Trip Time (RTT), and application buffer level. We observe that flow isolation techniques, combined with queue management and multipath transport, significantly improve the QoE for multimedia applications.