Mariner

RTT-based User Device & State Identification

---

Idea

Research Paper

This project is inspired by recent academic research that shows Round Trip Time (RTT) patterns can be used to infer user-side properties without accessing device internals.

Primary reference:

• “Inferring Mobile Device States from Network RTT Measurements” (arXiv:2411.11194)

The paper demonstrates that subtle variations in RTT distributions correlate with:

• Device type (mobile vs desktop / WhatsApp Web)
• Mobile operating system
• App and device state (locked, unlocked, foreground app)

This project attempts to replicate and simplify those findings in a practical, application-level setup.

---

Showcase:

1. Fill the form with target's whatsapp no in international format.

   • Silent payload : This would not alert the target by sending messages without producing user-facing notifications or visible prompts.
   • Count : Number of messages to send. It is limmted to 10 to prevent unethical use.

2. After sending messages A RTT vs Count graph is plotted

3. The probality is callucated based on RTTs bands of differnt devices and there states

---

Purpose

The goal of this project is to:

• Identify whether a user is on mobile or web
• Infer mobile OS (Android / iOS)
• Estimate device state:
  • Locked
  • Unlocked
  • WhatsApp in foreground

All of this is done using only RTT samples, without:

• Device permissions
• OS-level access
• Invasive fingerprinting

This can be useful for:

• Security research
• Network behavior analysis
• Academic replication
• Anti-abuse & anomaly detection (research-only)

---

Setup

Architecture

• Frontend: React (RTT visualization & control)
• Backend: Go
• Messaging Layer: WhatsApp (via whatsmeow)
• Data: RTT samples collected per message/receipt event

Steps :

1. clone the repository
2. go to ./app and run npm i and then npm run dev
3. go to ./whatsapp_client and run go mod tidy after it run go run , in the same folder and connect via your WhatsApp locally

Flow

1. Send a controlled batch of messages
2. Measure RTT for each send/ack event
3. Collect RTT samples into arrays
4. Apply probability-based classification
5. Output:
   • Most probable device/state
   • Confidence percentages for each category

Requirements

• Node.js + React
• Go (whatsmeow)
• Stable network for controlled measurements

---

Disclaimer

This project is for educational and research purposes only.
No personal data is stored, and no user is tracked beyond anonymous RTT metrics.