Skip to content
Tutorials Center
My learning

Topics

Lesson 27 of 36 ยท Security and Privacy

IoT Threat Modeling: What Can Go Wrong

Learn to think like a defender: map the IoT attack surface, apply STRIDE, study classic failures like default passwords, and run a configuration auditor.

An IoT device is a small computer that lives for ten years, sits in a place nobody checks, speaks to the internet and often controls something physical. That combination makes security harder than for a laptop: devices are cheap, rarely patched, physically reachable by attackers, and compromised devices can open doors, stop pumps or join botnets that take down websites. A famous example is the Mirai botnet of 2016, built mostly from cameras and routers that still used factory-default passwords.

The way to start is threat modeling: a structured look at what you are protecting, who might attack and how, before choosing defenses. This lesson gives you the method; the next lessons cover identity and encryption and updates and privacy.

Everything here is for defending your own systems. The examples audit configuration data; they do not scan or attack anything.

Step 1: Know your assets and boundaries

List what is valuable: sensor data (privacy), control commands (safety), credentials and keys, firmware (intellectual property and integrity), and the device's reputation and availability. Then draw the trust boundaries: places where data passes from one level of trust to another.

flowchart LR
  D["Device<br/>(physically exposed)"] -- "Wi-Fi / cellular" --> G["Gateway"]
  G -- "Internet" --> B["Broker / Cloud"]
  B --> A["Web and mobile app"]
  U(("User")) --> A
  X(("Attacker")) -.-> D
  X -.-> G
  X -.-> B
  X -.-> A

Every arrow is an interface an attacker can listen on, replay on or tamper with.

Step 2: The attack surface

The attack surface is the sum of everything an attacker can touch. For IoT it spans:

Layer Examples of weakness
Hardware open debug ports (UART, JTAG), readable flash, extractable keys
Firmware hard-coded secrets, buffer overflows, no signature check on updates
Network unencrypted protocols, open ports, weak Wi-Fi setup
Cloud/API broken access control, shared credentials, no rate limiting
Mobile/web app insecure pairing, token leaks, weak password reset
People default passwords, phishing, lost phones

Step 3: STRIDE

STRIDE is a checklist of six threat kinds. For each interface ask which could apply.

STRIDE, part 1

Spoofing Pretending to be a device or server
Tampering Changing data or firmware
Repudiation Denying an action, no logs
1 / 2

Example for one interface, a device publishing temperature to the broker:

  • Spoofing: someone connects with the device's client ID and publishes fake readings. Defense: per-device credentials or certificates.
  • Tampering: a man in the middle alters the value. Defense: TLS.
  • Information disclosure: a neighbor subscribes to #. Defense: topic access control lists.
  • Denial of service: a flood of messages fills the broker. Defense: rate limits and quotas.
  • Elevation of privilege: a temperature sensor can also publish commands to the door lock. Defense: least privilege per topic.
  • Repudiation: no log says who unlocked the door. Defense: audit logging with device identity.

Common IoT failures

  1. Default or shared credentials. One password for a whole product line means one leak compromises everything.
  2. Hard-coded secrets in firmware. Anyone who downloads the firmware reads them.
  3. No encryption, or encryption with certificate checks disabled "to make it work".
  4. Unauthenticated services left running (telnet, a debug web server, an open MQTT broker).
  5. No update path, so known bugs stay forever.
  6. Over-collection of data, then a breach exposes it.
  7. Insecure pairing: a setup Wi-Fi network with no password, or a Bluetooth pairing without confirmation.

Auditing a fleet configuration

Many of these failures are visible in configuration data. The program below checks a device inventory against rules and prints a risk score. It is a pure data check, which is how real compliance scanners start.

Node.js A simple IoT configuration auditor

Press Run (or Ctrl + Enter) to see the result here.

Program input ยท what you'd type while the program runs

Type the answers before you press Run, one per line. Each time the program reads a line (with node:readline or process.stdin), it gets the next line from this box.

Real scores use frameworks such as CVSS, but the idea is the same: turn vague worries into a ranked to-do list.

Throttling guesses

One cheap defense against password guessing is to slow the attacker down: count failures per source and lock out temporarily. The simulation below shows exponential back-off.

Node.js Lockout with exponential back-off

Press Run (or Ctrl + Enter) to see the result here.

Program input ยท what you'd type while the program runs

Type the answers before you press Run, one per line. Each time the program reads a line (with node:readline or process.stdin), it gets the next line from this box.

Slowing guesses never replaces strong, unique credentials, but it makes bulk guessing impractical.

Check your understanding

Quick check ยท 4 questions

of 4 answered You scored / 4 ยท perfect!

  1. 1. In STRIDE, an attacker connecting with a stolen client ID and sending fake readings is an example of:

    The attacker pretends to be a legitimate device, which is spoofing identity.
  2. 2. Why are factory-default passwords a particular IoT risk?

    Default credentials are public knowledge, and automated scanners try them against every reachable device.
  3. 3. What is a trust boundary?

    Threat modeling focuses on these crossing points because that is where checks are needed.
  4. 4. A sensor can publish to the door lock's command topic. Which principle is violated?

    Each device should have only the permissions its job requires.

Summary

  • Threat modeling: list assets, draw trust boundaries, map the attack surface, apply STRIDE to each interface.
  • IoT adds physical access, long lifetimes, weak update paths and a real-world impact.
  • Most failures are basic: default credentials, hard-coded secrets, missing encryption, no updates.
  • Audit configurations automatically and throttle guessing, but build strong identity and least privilege in from the start.

Next: Device Identity, TLS and Access Control