A drop-in RIPE Atlas probe,
written in Rust.

01What it is

RIPE Atlas is a global measurement network: thousands of probes run scheduled ping, traceroute, DNS, HTTP, TLS, and NTP measurements from their vantage points and ship the results back to the RIPE NCC. The reference probe is a daemon originally written in C, intended for embedded hardware.

starla is the same daemon, rewritten as one Rust binary. It connects to the same controller infrastructure, accepts the same CRONLINE and ONEOFF commands, and produces RESULT lines that are byte-compatible with the reference probe's output (field order, spacing, numeric precision included).

Same wire protocol. Different runtime — one binary, no OpenSSH, no subprocesses, no listening ports.

02Features

03Quick start

Pick a runtime. Each ships with sensible defaults; configuration is optional.

shelldocker · podman
# persistent state, raw sockets for ping/trace
docker run -d --name starla \
  -v starla-state:/state \
  --cap-add NET_RAW \
  ghcr.io/ananthb/starla:latest

nixflake input + module
inputs.starla.url = "github:ananthb/starla";

# in your nixos config:
imports = [ starla.nixosModules.default ];
services.starla.enable = true;

shellapt
curl -LO https://github.com/ananthb/starla/releases/latest/download/starla_amd64.deb
sudo dpkg -i starla_*.deb
sudo systemctl enable --now starla

shelldnf
curl -LO https://github.com/ananthb/starla/releases/latest/download/starla.x86_64.rpm
sudo dnf install ./starla*.rpm
sudo systemctl enable --now starla

shellhomebrew
brew install ananthb/starla/starla
brew services start starla

# optional GUI tray applet
brew install --cask ananthb/starla/starla-tray

shellrelease tarball
curl -LO https://github.com/ananthb/starla/releases/latest/download/starla-amd64.tar.gz
tar xzf starla-amd64.tar.gz
sudo ./starla/starla

Full options, configuration, and signature verification are in the installation guide.

04How it works

On first run, starla generates an SSH key pair and registers with a RIPE Atlas registration server. After registration it opens a single long-lived SSH connection to the assigned controller and receives measurement schedules over a reverse-tunneled telnet stream. Results are uploaded back through SSH direct-tcpip channels.

Probe state (SSH keys, probe ID, known hosts) is persisted in the state directory and survives restarts. Pending result uploads live in a bounded in-memory queue and do not.

See architecture for the data flow, or protocol for the wire-level details.

05Register your probe

After first start, copy the public key from the log output (or probe_key.pub in the state directory) and paste it into the application form at atlas.ripe.net/apply/swprobe. starla retries in the background; once approved, it connects and begins running measurements.