This repository contains the dataset for the paper LENS: A LEO Satellite Network Measurement Dataset published in ACM Multimedia Systems Conference (MMSys'24) Open-source Software and Dataset (ODS) track.
The original dataset snapshot for the published paper can be found at commit c084c11.
You can also check out our MMSys'24 paper Low-Latency Live Video Streaming over a Low-Earth-Orbit Satellite Network with DASH that utilizes this dataset for low-latency live video streaming over Starlink.
For inside-out measurements, the datasets are collected with multiple dishes located in the following regions.
| ID | Location | Dish Generation | Point-of-Presence | Service Tier |
|---|---|---|---|---|
| victoria_active_1 | Victoria, BC, Canada | rev3_proto2 | Seattle | Standard |
| victoria_active_2 | Victoria, BC, Canada | rev3_proto2 | Seattle | Mobile |
| victoria_inactive | Victoria, BC, Canada | rev3_proto2 | Seattle | Inactive Mobile, Roam |
| vancouver | Vancouver, BC, Canada | rev2_proto3 | Seattle | Standard [1] |
| seattle | Seattle, WA, USA | rev3_proto2 | Seattle | Standard |
| seattle_hp | Seattle, WA, USA | hp1_proto1 | Seattle | Priority |
| alaska | Anchorage, AK, USA | rev3_proto2 | Seattle | Mobile |
| calgary | Calgary, AB, Canada | rev3_proto2 | Seattle | Inactive Standard |
| ottawa | Ottawa, ON, Canada | rev3_proto2 | New York | Standard [5] |
| iowa | Iowa City, IA, USA | rev1_pre_production | Chicago | Standard |
| denver | Denver, CO, USA | rev3_proto2 | Denver | Mobile, Roam [6] |
| dallas | Oxford, MS, USA | rev3_proto2 | Dallas | Inactive Standard |
| louvain | Louvain, Belgium | rev3_proto2 | Frankfurt / London [4] | Standard |
| seychelles [3] | Seychelles | rev3_proto2 | Lagos / Frankfurt [2] | Mobile, Roam |
Ref:
Source: https://twitter.com/olegkutkov/status/1742322178320670753/
Note:
- The subscription plan associated with the vancouver dish was paused between 2023/12/29 and 2024/01/09, during which inactive measurements was conducted.
- The PoP associated with the seychelles dish was changed from Lagos to Frankfurt on 2023/12/08.
- The Starlink subscription at this installation has been canceled since March 2024, see 202403 for more details.
- The PoP associated with the louvain dish was changed from Frankfurt to London.
- The Starlink subscription at this installation has been paused since March 2024, see 202403 for more details.
- The Starlink subscription at this installation has been paused since March 2024, see 202403 for more details.
| victoria_active_1 | victoria_active_2 | victoria_inactive |
|---|---|---|
 |
 |
 |
| vancouver | seattle | seattle_hp |
 |
 |
 |
| alaska | ottawa | iowa |
 |
 |
 |
| denver | dallas | louvain |
 |
 |
 |
| seychelles | ||
 |
The dataset is split into monthly snapshots. Each can be retrieved from Zenodo using the links below.
Due to the file size limit on Zenodo, monthly snapshots are created in splitted tar archives using the following command.
e.g.,
tar -I "zstd -T24 -8" -cvf - LENS-2024-01 | split --bytes=40GB - LENS-2024-01.tar.zst.To decompress, make sure Zstd is installed. Download all the splitted tar archives for the same month in the same folder, and decompress using the following command. Make sure you have enough disk space.
e.g.,
cat LENS-2024-01.tar.zst.* | tar --zstd -xf -RAW dataset contains IRTT metrics in .json formats and ping metrics in .txt formats. Examples of both files are shown below.
IRTT Example
{
"version": {
"irtt": "0.9.1-clarkzjw",
"protocol": 1,
"json_format": 1
},
"system_info": {
"os": "linux",
"cpus": 16,
"go_version": "go1.21.6",
"hostname": "REDACTED"
},
"config": {
"local_address": "REDACTED",
"remote_address": "REDACTED",
"open_timeouts": "1s,2s,4s,8s",
"params": {
"proto_version": 1,
"duration": 5000000000,
"interval": 1000000000,
"length": 60,
"received_stats": "both",
"stamp_at": "both",
"clock": "both",
"dscp": 0,
"server_fill": ""
},
"loose": false,
"ip_version": "IPv4",
"df": 0,
"ttl": 0,
"timer": "comp",
"time_source": "go",
"waiter": "3x4s",
"filler": "none",
"fill_one": false,
"server_fill": "",
"thread_lock": false,
"supplied": {
"local_address": "REDACTED",
"remote_address": "REDACTED",
"open_timeouts": "1s,2s,4s,8s",
"params": {
"proto_version": 1,
"duration": 5000000000,
"interval": 1000000000,
"length": 0,
"received_stats": "both",
"stamp_at": "both",
"clock": "both",
"dscp": 0,
"server_fill": ""
},
"loose": false,
"ip_version": "IPv4",
"df": 0,
"ttl": 0,
"timer": "comp",
"time_source": "go",
"waiter": "3x4s",
"filler": "none",
"fill_one": false,
"server_fill": "",
"thread_lock": false
}
},
"stats": {
"start_time": {
"wall": 1707277440772833268,
"monotonic": 22210666
},
"send_call": {
"total": 127256,
"n": 5,
"min": 12052,
"max": 30262,
"mean": 25451,
"stddev": 7639,
"variance": 58362885
},
"timer_error": {
"total": 1299408,
"n": 4,
"min": 42505,
"max": 643010,
"mean": 324852,
"stddev": 265465,
"variance": 70472098752
},
"rtt": {
"total": 83084018,
"n": 5,
"min": 16064572,
"max": 17291970,
"mean": 16616803,
"median": 16597560,
"stddev": 488290,
"variance": 238427848599
},
"send_delay": {
"total": 52436439042,
"n": 5,
"min": 10486437967,
"max": 10488213200,
"mean": 10487287808,
"median": 10487235754,
"stddev": 703446,
"variance": 494837607693
},
"receive_delay": {
"total": -52353355157,
"n": 5,
"min": -10470976733,
"max": -10470232546,
"mean": -10470671031,
"median": -10470851192,
"stddev": 342554,
"variance": 117343485897
},
"server_packets_received": 5,
"bytes_sent": 300,
"bytes_received": 300,
"duplicates": 0,
"late_packets": 0,
"wait": 51875910,
"duration": 4053373029,
"packets_sent": 5,
"packets_received": 5,
"packet_loss_percent": 0,
"upstream_loss_percent": 0,
"downstream_loss_percent": 0,
"duplicate_percent": 0,
"late_packets_percent": 0,
"ipdv_send": {
"total": 2978734,
"n": 4,
"min": 405684,
"max": 1284073,
"mean": 744683,
"median": 644488,
"stddev": 397052,
"variance": 157650351613
},
"ipdv_receive": {
"total": 1895362,
"n": 4,
"min": 70066,
"max": 744191,
"mean": 473840,
"median": 540552,
"stddev": 290347,
"variance": 84301497643
},
"ipdv_round_trip": {
"total": 1760386,
"n": 4,
"min": 194481,
"max": 806291,
"mean": 440096,
"median": 379807,
"stddev": 261474,
"variance": 68368804430
},
"server_processing_time": {
"total": 17180,
"n": 5,
"min": 1780,
"max": 4380,
"mean": 3436,
"stddev": 987,
"variance": 976030
},
"timer_err_percent": 0.0324852,
"timer_misses": 0,
"timer_miss_percent": 0,
"send_rate": {
"bps": 599,
"string": "599 bps"
},
"receive_rate": {
"bps": 600,
"string": "600 bps"
}
},
"round_trips": [
{
"seqno": 0,
"lost": "false",
"timestamps": {
"client": {
"receive": {
"wall": 1707277440790127432,
"monotonic": 39504855
},
"send": {
"wall": 1707277440772833721,
"monotonic": 22211105
}
},
"server": {
"receive": {
"wall": 1707277451261046921,
"monotonic": 8380840759811810
},
"send": {
"wall": 1707277451261048701,
"monotonic": 8380840759813590
}
},
"Ecn": 0
},
"delay": {
"receive": -10470921269,
"rtt": 17291970,
"send": 10488213200
},
"ipdv": {}
},
{
"seqno": 1,
"lost": "false",
"timestamps": {
"client": {
"receive": {
"wall": 1707277441790353996,
"monotonic": 1039731408
},
"send": {
"wall": 1707277441773479264,
"monotonic": 1022856655
}
},
"server": {
"receive": {
"wall": 1707277452261201298,
"monotonic": 8380841759966187
},
"send": {
"wall": 1707277452261205188,
"monotonic": 8380841759970077
}
},
"Ecn": 0
},
"delay": {
"receive": -10470851192,
"rtt": 16870863,
"send": 10487722034
},
"ipdv": {
"receive": 70066,
"rtt": -421107,
"send": -491173
}
},
{
"seqno": 2,
"lost": "false",
"timestamps": {
"client": {
"receive": {
"wall": 1707277442788720025,
"monotonic": 2038097444
},
"send": {
"wall": 1707277442772652035,
"monotonic": 2022029432
}
},
"server": {
"receive": {
"wall": 1707277453259090002,
"monotonic": 8380842757854891
},
"send": {
"wall": 1707277453259093442,
"monotonic": 8380842757858331
}
},
"Ecn": 0
},
"delay": {
"receive": -10470373417,
"rtt": 16064572,
"send": 10486437967
},
"ipdv": {
"receive": 477782,
"rtt": -806291,
"send": -1284073
}
},
{
"seqno": 3,
"lost": "false",
"timestamps": {
"client": {
"receive": {
"wall": 1707277443789142691,
"monotonic": 3038520113
},
"send": {
"wall": 1707277443772879290,
"monotonic": 3022256680
}
},
"server": {
"receive": {
"wall": 1707277454260115044,
"monotonic": 8380843758879943
},
"send": {
"wall": 1707277454260119424,
"monotonic": 8380843758884323
}
},
"Ecn": 0
},
"delay": {
"receive": -10470976733,
"rtt": 16259053,
"send": 10487235754
},
"ipdv": {
"receive": -603323,
"rtt": 194481,
"send": 797804
}
},
{
"seqno": 4,
"lost": "false",
"timestamps": {
"client": {
"receive": {
"wall": 1707277444789867766,
"monotonic": 4039245182
},
"send": {
"wall": 1707277444773266535,
"monotonic": 4022643932
}
},
"server": {
"receive": {
"wall": 1707277455260096622,
"monotonic": 8380844758861511
},
"send": {
"wall": 1707277455260100312,
"monotonic": 8380844758865201
}
},
"Ecn": 0
},
"delay": {
"receive": -10470232546,
"rtt": 16597560,
"send": 10486830087
},
"ipdv": {
"receive": 744191,
"rtt": 338507,
"send": -405684
}
}
]
}Also see IRTT-CLIENT (1) for details.
Ping Example
[1700182800.410606] 64 bytes from 2605:59c8:1000:962f::1: icmp_seq=1 ttl=63 time=60.6 ms
[1700182800.421290] 64 bytes from 2605:59c8:1000:962f::1: icmp_seq=2 ttl=63 time=56.9 ms
[1700182800.442474] 64 bytes from 2605:59c8:1000:962f::1: icmp_seq=3 ttl=63 time=62.0 ms
[1700182800.465254] 64 bytes from 2605:59c8:1000:962f::1: icmp_seq=4 ttl=63 time=68.8 ms
[1700182800.475936] 64 bytes from 2605:59c8:1000:962f::1: icmp_seq=5 ttl=63 time=65.1 ms
Processed CSV dataset only contains timestamps and necessary latency metrics.
For IRTT metrics, an example of the processed *.csv is shown below.
IRTT CSV Example
timestamp,rtt,uplink,downlink
1705744800616897349,0,0,-1
1705744800626565141,120.126123,72.439368,47.68675
1705744800636737402,109.976237,62.295097,47.681199
1705744800646307744,100.411714,52.731855,47.679937
1705744800656691526,90.032887,42.357473,47.67541
1705744800666731717,0,0,-1
1705744800676416449,0,0,-1
1705744800686681812,97.297734,49.362602,47.935125
1705744800696496401,87.506136,39.592273,47.913945
Note, for rtt, uplink, downlink, a value of -1 represents packet loss.
Per IRTT documentation,
lost the lost status of the packet, which can be one of false, true, true_down or true_up. The true_down and true_up values are only possible if the ReceivedStats parameter includes ReceivedStatsWindow (irtt client --stats flag). Even then, if it could not be determined whether the packet was lost upstream or downstream, the value true is used.
When converting from *.json to *.csv, we assign -1 to rtt and 0 to others, when the lost status is true; assign -1 to uplink and 0 to others, when the lost status is true_up; assign -1 to downlink and 0 to others, when the lost status is true_down;
For Ping metrics, an example of the processed *.csv is shown below.
Ping CSV Example
timestamp,rtt
1705744800.533641,91.8
1705744800.55492,96.5
1705744800.602846,111.0
1705744800.61361,106.0
1705744800.613665,88.3
1705744800.677437,106.0
This repository is licensed under GPL-3.0.
The dataset files on Zenodo are released under CC BY-SA 4.0.
If you use this dataset in your research, please cite the following paper:
@inproceedings{10.1145/3625468.3652170,
author = {Zhao, Jinwei and Pan, Jianping},
title = {LENS: A LEO Satellite Network Measurement Dataset},
year = {2024},
isbn = {9798400704123},
publisher = {Association for Computing Machinery},
address = {New York, NY, USA},
url = {https://doi.org/10.1145/3625468.3652170},
doi = {10.1145/3625468.3652170},
booktitle = {Proceedings of the 15th ACM Multimedia Systems Conference},
pages = {278–284},
numpages = {7},
keywords = {Dataset, Inter-Satellite Links, LEO, Latency, Network Measurement},
location = {Bari, Italy},
series = {MMSys '24}
}
This work is not possible without our alumni and their students who hosted our Starlink dishes, and other Starlink users who allowed us to access their dishes remotely.
React
Typescript
Django
Laravel
Facebook
Microsoft
Google
Alibaba
D3
Tencent