Table of Contents
Fetching ...

Enhancing QoE in HTTP/3 using EPS Framework

Abhinav Gupta, Radim Bartos

TL;DR

This paper addresses QoE improvements for HTTP/3 by integrating the Extensible Prioritization Scheme (EPS) with a mixed scheduling strategy that simultaneously handles Non-Incremental ($N$) and Incremental ($I$) resource delivery under EPS urgencies. It contributes an EPS-based priority mapping derived from Chromium resource types, and a two-queue scheduling mechanism that reorders $I$ deliveries by urgency while serving $N$ deliveries by their urgency. The evaluation, conducted on eight popular websites with Lighthouse measurements, shows consistent FCP/LCP and TBT/TTI gains on larger sites, with CLS remaining stable below 0.1 and notable improvements for Amazon and Etsy. The results suggest that combining EPS with a mixed $N$/$I$ delivery strategy can significantly enhance web performance in HTTP/3, motivating deployment in other QUIC implementations and further adaptive priority schemes.

Abstract

HTTP/3, the latest evolution of the Hypertext Transfer Protocol, utilizes QUIC, a new transport protocol leveraging UDP to overcome limitations such as connection time and head-of-line blocking prevalent in HTTP/2. This advancement is enhanced by the Extensible Prioritization Scheme (EPS), which introduces a flexible prioritization framework for improving website resource delivery. This paper proposes a mixed scheduling mechanism that delivers using mixed incremental and non-incremental resource delivery and adheres to EPS urgency levels to improve the QoE. Additionally, we propose an EPS priority mapping to enhance the QoE further. This mapping is based on the priority indicated by the Chromium browser and the resource type. The result of the experimental evaluation indicates that the proposed mechanism and mapping improve commonly-used website performance measures for sites featuring a comparatively large number and size of resources.

Enhancing QoE in HTTP/3 using EPS Framework

TL;DR

This paper addresses QoE improvements for HTTP/3 by integrating the Extensible Prioritization Scheme (EPS) with a mixed scheduling strategy that simultaneously handles Non-Incremental () and Incremental () resource delivery under EPS urgencies. It contributes an EPS-based priority mapping derived from Chromium resource types, and a two-queue scheduling mechanism that reorders deliveries by urgency while serving deliveries by their urgency. The evaluation, conducted on eight popular websites with Lighthouse measurements, shows consistent FCP/LCP and TBT/TTI gains on larger sites, with CLS remaining stable below 0.1 and notable improvements for Amazon and Etsy. The results suggest that combining EPS with a mixed / delivery strategy can significantly enhance web performance in HTTP/3, motivating deployment in other QUIC implementations and further adaptive priority schemes.

Abstract

HTTP/3, the latest evolution of the Hypertext Transfer Protocol, utilizes QUIC, a new transport protocol leveraging UDP to overcome limitations such as connection time and head-of-line blocking prevalent in HTTP/2. This advancement is enhanced by the Extensible Prioritization Scheme (EPS), which introduces a flexible prioritization framework for improving website resource delivery. This paper proposes a mixed scheduling mechanism that delivers using mixed incremental and non-incremental resource delivery and adheres to EPS urgency levels to improve the QoE. Additionally, we propose an EPS priority mapping to enhance the QoE further. This mapping is based on the priority indicated by the Chromium browser and the resource type. The result of the experimental evaluation indicates that the proposed mechanism and mapping improve commonly-used website performance measures for sites featuring a comparatively large number and size of resources.
Paper Structure (12 sections, 5 figures, 1 table, 1 algorithm)

This paper contains 12 sections, 5 figures, 1 table, 1 algorithm.

Figures (5)

  • Figure 1: Proposed Mixed Scheduling Mechanism
  • Figure 2: Example of Request Scheduling
  • Figure 3: Websites used in Experimental Evaluation
  • Figure 4: Relative Performance Improvement of Mixed over Standard Sequential
  • Figure 5: Relative Performance Improvement