Logo
ScienceStack
Table of Contents
Fetching ...
Sign In

Low-Latency Video Conferencing via Optimized Packet Routing and Reordering

Yao Xiao, Sitian Chen, Amelie Chi Zhou, Shuhao Zhang, Yi Wang, Rui Mao, Xuan Yang

TL;DR

VCRoute and WMJitter address end-to-end latency challenges in geo-distributed video conferencing by jointly optimizing routing decisions and jitter management. VCRoute reframes per-packet routing as a Multi-Armed Bandit problem and uses Thompson Sampling with a DNN-based latency predictor to balance transmitting latency and reordering delay, while WMJitter employs watermark-based Out-of-Order Processing to reduce reordering latency at the receiver. Trace-driven evaluations on WonderProxy and Tencent Cloud show up to 44% end-to-end latency reductions with overhead under 1%, and WMJitter provides additional gains in highly dynamic networks. These results demonstrate practical benefits for global video conferencing deployments, enabling lower latency and more stable QoE across diverse geo-distributed data-center topologies.

Abstract

In the face of rising global demand for video meetings, managing traffic across geographically distributed (geo-distributed) data centers presents a significant challenge due to the dynamic and limited nature of inter-DC network performance. Facing these issues, this paper introduces two novel techniques, VCRoute and WMJitter, to optimize the performance of geo-distributed video conferencing systems. VCRoute is a routing method designed for audio data packets of video conferences. It treats the routing problem as a Multi-Armed Bandit issue, and utilizes a tailored Thompson Sampling algorithm for resolution. Unlike traditional approaches, VCRoute considers transmitting latency and its variance simultaneously by using Thompson Sampling algorithm, which leads to effective end-to-end latency optimization. In conjunction with VCRoute, we present WMJitter, a watermark-based mechanism for managing network jitter, which can further reduce the end-to-end delay and keep an improved balance between latency and loss rate. Evaluations based on real geo-distributed network performance demonstrate the effectiveness and scalability of VCRoute and WMJitter, offering robust solutions for optimizing video conferencing systems in geo-distributed settings.

Low-Latency Video Conferencing via Optimized Packet Routing and Reordering

TL;DR

VCRoute and WMJitter address end-to-end latency challenges in geo-distributed video conferencing by jointly optimizing routing decisions and jitter management. VCRoute reframes per-packet routing as a Multi-Armed Bandit problem and uses Thompson Sampling with a DNN-based latency predictor to balance transmitting latency and reordering delay, while WMJitter employs watermark-based Out-of-Order Processing to reduce reordering latency at the receiver. Trace-driven evaluations on WonderProxy and Tencent Cloud show up to 44% end-to-end latency reductions with overhead under 1%, and WMJitter provides additional gains in highly dynamic networks. These results demonstrate practical benefits for global video conferencing deployments, enabling lower latency and more stable QoE across diverse geo-distributed data-center topologies.

Abstract

In the face of rising global demand for video meetings, managing traffic across geographically distributed (geo-distributed) data centers presents a significant challenge due to the dynamic and limited nature of inter-DC network performance. Facing these issues, this paper introduces two novel techniques, VCRoute and WMJitter, to optimize the performance of geo-distributed video conferencing systems. VCRoute is a routing method designed for audio data packets of video conferences. It treats the routing problem as a Multi-Armed Bandit issue, and utilizes a tailored Thompson Sampling algorithm for resolution. Unlike traditional approaches, VCRoute considers transmitting latency and its variance simultaneously by using Thompson Sampling algorithm, which leads to effective end-to-end latency optimization. In conjunction with VCRoute, we present WMJitter, a watermark-based mechanism for managing network jitter, which can further reduce the end-to-end delay and keep an improved balance between latency and loss rate. Evaluations based on real geo-distributed network performance demonstrate the effectiveness and scalability of VCRoute and WMJitter, offering robust solutions for optimizing video conferencing systems in geo-distributed settings.
Paper Structure (26 sections, 3 equations, 14 figures, 1 table, 1 algorithm)

This paper contains 26 sections, 3 equations, 14 figures, 1 table, 1 algorithm.

Table of Contents

  1. Introduction
  2. Background and Motivation
  3. Geo-Distributed Network Features
  4. video Conferencing Systems
  5. Motivations
  6. System Overview
  7. VCRoute Scheduler
  8. Reducing overhead
  9. Making good routing decisions
  10. Thompson Sampling algorithm
  11. Watermark-based Jitter Manager
  12. Incorporating Watermarks in Jitter Management
  13. Balancing Latency with Packet Loss
  14. Evaluation
  15. Experimental Setup
  16. ...and 11 more sections

Figures (14)

  • Figure 1: video conferencing system service architectures.
  • Figure 2: RTT variation over time between different Tencent cloud DCs. We measure RTT once per minute and plot the average of every ten minutes.
  • Figure 3: Jitter Buffer Manager in webRTC
  • Figure 4: Latencies of 30,000 packets communicated between two ends in a video conferencing system. Average latency of every 50 packets is plotted.
  • Figure 5: System Overview
  • ...and 9 more figures