preprints

UVic computer science
pan
Posts: 54855
Joined: Mon Sep 11, 2006 10:23 am

preprints

Post by pan »

to appear at ieee icc 2024
Measuring the Satellite Links of a LEO Network

Jianping Pan, Jinwei Zhao and Lin Cai
University of Victoria, BC, Canada

Abstract—Low-earth-orbit (LEO) satellite networks have be- come very popular in recent years, exemplified by Starlink, OneWeb, Kuiper and others, due to the dramatically reduced launch cost and increased demand for connectivity anytime, anywhere. After an exploration of Starlink access, core and backbone networks, in this paper we focus on the satellite access network (SAN) of Starlink around the world. Particularly, we measure the access performance in terms of one-way delay and round-trip time from user terminal (UT) to ground station (GS) and point-of-presence (PoP), both inside-out and outside-in, and even on inactive dishes. It reveals the unique characteristics of Starlink SAN in terms of satellite-GS scheduling, media access control and user contention, and sheds light on the challenges and opportunities for network protocols and applications. The paper will be complemented by public dataset release and conference on-site demo for the research and industry community.

Index Terms—LEO satellite networks, Starlink, measurement
feedback welcome at the camera-ready stage
Attachments
starlink_icc (6).pdf
(863.29 KiB) Downloaded 299 times
pan
Posts: 54855
Joined: Mon Sep 11, 2006 10:23 am

Re: preprints

Post by pan »

to appear in acm mmsys 2024
LENS: A LEO Satellite Network Measurement Dataset https://github.com/clarkzjw/LENS

Jinwei Zhao University of Victoria Victoria, BC, Canada clarkzjw@uvic.ca
Jianping Pan University of Victoria Victoria, BC, Canada pan@uvic.ca

ABSTRACT Low-Earth-Orbit (LEO) satellite constellations are narrowing the performance gap between satellite networks and the terrestrial Internet. Low-latency satellite Internet offered by Starlink enables functionalities that are otherwise unachievable with the traditional geosynchronous equatorial orbit (GEO) satellite networks, including low-latency live video streaming, cloud gaming and real-time video conferencing. The absence of a comprehensive and long-term network measurement dataset with a global perspective poses significant challenges for researchers to evaluate the application performance over Starlink networks. In this paper, we introduce LENS, which is a LEO satellite network measurement dataset, collected from 13 Starlink dishes, associated with 7 Point-of-Presence (PoP) locations across 3 continents. The dataset currently consists of network latency traces from Starlink dishes with different hardware revisions, various service subscriptions and distinct sky obstruction ratios. We provide a high-level overview and analysis of the latency performance using the dataset and discuss various use cases. This dataset is useful for researchers who wish to understand the long-term network performance of Starlink and to evaluate and optimize the performance of multimedia applications over satellite networks.
feedback welcome during the camera-ready stage
20240314 (1).pdf
(5.41 MiB) Downloaded 260 times
lens = low earth network of satellites ;-)
Post Reply

Return to “UVic CS”