Starlink Measurement in the Wild

Starlink.md

Required programs: curl, nslookup, grpcurl, traceroute, ping, mtr

Note: On some Linux distributions, you need to manually install iputils-ping, instead of inetutils-ping.

sudo apt-get install -y iputils-ping

(1). What is the public IPv4 address? Does it have a DNS PTR record?

curl -4 ipconfig.io
nslookup $(curl -s -4 ipconfig.io)

The expected output of the DNS PTR record has the following format:

customer.*.pop.starlinkisp.net.

while * indicates the associated Starlink Point-of-Presence (PoP), e.g., customer.sttlwax1.pop.starlinkisp.net., where sttlwax1 represents the Starlink PoP in Seattle.

(2). Does it support IPv6? What is the public IPv6 address and its DNS PTR records?

curl -6 ipconfig.io
nslookup $(curl -s -6 ipconfig.io)

(3). Can you run traceroute to 1.1.1.1?

traceroute 1.1.1.1

and

sudo mtr 1.1.1.1 -b -c 10 -i 0.5 -r -w

(4). What is the assigned LAN IP address?

(5). Can you reach 192.168.100.1 with ping?

If not, can you find the gateway of your LAN and try to add routing rules to 192.168.100.1 via the gateway as the next hop? On Linux, it's similar to

sudo ip r a 192.168.100.1 dev <iface>

or

sudo ip r a 192.168.100.1 via <gateway>

On macOS, you can try something like: https://superuser.com/questions/756134/how-to-direct-ip-route-through-specific-interface-in-os-x

sudo route -t -n add -net 192.168.100.1/32 <gateway>

(6). Can you reach 100.64.0.1 with ping?

Note: if you are already running Tailscale locally on your machine, on Linux, you can set up with sudo tailscale up --netfilter-mode=off. On macOS or Windows, you can temporarily disable/quit Tailscale.

If not, can you follow similar steps as in (5) to add routing rules for 100.64.0.1?

(7). If you can reach 100.64.0.1 with ping, can you run:

ping -D -i 0.01 -c 60000 100.64.0.1 > ping-100.64.0.1-`date "+%y%m%d-%H%M%S"`.txt

If you cannot reach 100.64.0.1, can you ping 1.1.1.1?

ping -D -i 0.01 -c 60000 1.1.1.1 > ping-1.1.1.1-`date "+%y%m%d-%H%M%S"`.txt

Note: You may need to add sudo in order to run ping with -i 0.01.

On macOS, you have to replace -D with --apple-time, e.g.:

ping --apple-time -i 0.01 -c 60000 100.64.0.1 > ping-100.64.0.1-`date "+%y%m%d-%H%M%S"`.txt

ping --apple-time -i 0.01 -c 60000 1.1.1.1 > ping-1.1.1.1-`date "+%y%m%d-%H%M%S"`.txt

(8) If you can reach 192.168.100.1, can you run grpcurl with the following:

grpcurl -plaintext -d {\"get_status\":{}} 192.168.100.1:9200 SpaceX.API.Device.Device/Handle

grpcurl -plaintext -d {\"get_diagnostics\":{}} 192.168.100.1:9200 SpaceX.API.Device.Device/Handle

(9) Can you run the following traceroute script?

Note: This script will spawn many traceroute processes and create ~30 text files in the current directory, it would take a while (around 20-30 minutes) to finish.

#!/bin/bash

for i in $(seq 108 109)
  do
  for j in $(seq 0 255)
    do
    traceroute -enm 18 -w 1 149.19.$i.$j >> tr-sl-bb-ip-$i.txt 2>&1
    done &
  done &

for i in $(seq 64 95)
  do
  for j in $(seq 0 255)
    do
    traceroute -enm 18 -w 1 206.224.$i.$j >> tr-sl-bb-ip-$i.txt 2>&1
    done &
  done &

or with mtr

#!/bin/bash

for i in $(seq 108 109)
  do
  for j in $(seq 0 255)
    do
    echo 149.19.$i.$j >> mtr-sl-bb.ip-$i.txt
    sudo mtr 149.19.$i.$j -m 18 --mpls -r -w -b -c 3 >> mtr-sl-bb.ip-$i.txt 2>&1
    done &
  done &
  
for i in $(seq 64 95)
  do
  for j in $(seq 0 255)
    do
    echo 206.224.$i.$j >> mtr-sl-bb-ip-$i.txt
    sudo mtr 206.224.$i.$j -m 18 --mpls -r -w -b -c 3 >> mtr-sl-bb-ip-$i.txt 2>&1
    done &
  done &

The expected output looks like:

head tr-sl-bb-ip-109.txt
traceroute to 149.19.109.0 (149.19.109.0), 18 hops max, 60 byte packets
 1  * * *
 2  100.64.0.1  24.071 ms  24.106 ms  24.054 ms
 3  172.16.252.40  23.958 ms  23.993 ms  23.905 ms
 4  * * *
 5  206.224.64.19  39.331 ms  39.485 ms 206.224.64.45  39.279 ms
 6  149.19.108.109  92.913 ms  88.453 ms  88.930 ms
 7  149.19.108.66  99.491 ms  99.282 ms  99.580 ms
 8  149.19.108.185  168.231 ms  170.564 ms  159.923 ms
 9  149.19.109.77  181.158 ms  174.520 ms  189.733 ms

while it's normal some targets cannot be reached:

head tr-sl-bb-ip-108.txt
traceroute to 149.19.108.0 (149.19.108.0), 18 hops max, 60 byte packets
 1  * * *
 2  100.64.0.1  42.131 ms  42.098 ms  42.069 ms
 3  * * *
 4  * * *
 5  * * *
 6  * * *
 7  * * *
 8  * * *
 9  * * *

(10). Can you repeat steps (1), (2) and (7) multiple times at different times of the day and record the date and time of the measurement?

Thanks!