Clock skew and crystal timing fingerprint
Measures tiny clock speed differences between a device's timers.
In one line
Can compare your device clocks over many seconds to tell it apart
Measures tiny clock speed differences between a device's timers.
Each device has slightly different oscillator frequencies. Comparing performance.now, Date, requestAnimationFrame and server timestamps over many seconds yields a device-specific drift signature, researched as time-based fingerprinting.
What it can and can't do
What the user does: Nothing. It happens while the page stays open for some seconds
It can
- Compare several clocks, such as the page timer, the date and screen refresh
- Compare them with the site's server time
- Build a device number from tiny speed differences
It cannot
- Work in the first moments of a visit, since it needs many seconds
- Ask permission first
- See your files
- Know who you are by itself
Browser support
| Browser | Status | First version | Released | Note |
|---|---|---|---|---|
| Chrome | Supported | 1 | 2008-12-11 | None |
| Edge | Supported | 79 | 2020-01-15 | None |
| Firefox | Supported | 1 | 2004-11-09 | None |
| Safari | Supported | 1 | 2003-06-23 | None |
| Safari on iPhone | Supported | 1 | 2007-06-29 | None |
| Chrome on Android | Supported | 1 | Not listed | None |
| Samsung Internet | Supported | 5.0 | 2016-12-15 | None |
How a site asks for it
| API | performance.now(); Date.now(); requestAnimationFrame; WebSocket timestamps |
| How access is granted | none |
| What the user sees | Not listed |
| Needs a secure context (HTTPS) | No |
| Needs a user gesture | No |
| How long it lasts | Not listed |
| How to take it back | Not listed |
| Inside an iframe | Not listed |
Typical flow
- Sample timers repeatedly
- Compute drift ratio against server clock
- Cluster across visits
Data you get
What comes back: Drift in parts per million
Shape: float
| Field | Type | Unit | Range | Precision |
|---|---|---|---|---|
skew | float | ppm | -100 to 100 | needs minutes of samples |
Technical limits
- Needs seconds to minutes and low noise, less reliable on mobile
- Timer coarsening reduces accuracy
The story
What it is
A script measures tiny speed differences between your device's clocks to make a hardware ID.
For you
You never see it. Every device has tiny flaws in its clock chips. A page can measure those flaws over a few seconds and get a number that stays about the same for your device.
For makers
It works even when software signals are hidden or randomized, since it comes from the hardware itself. Researchers have shown it can separate nearly identical phones. Browsers fight it by blurring timers.
Where you see it
| Type | Example | Apps |
|---|---|---|
| security | A research team shows that two phones of the same model can be told apart by clock skew. | None named |
| ads | An ad tracker could use clock skew to recognize one device across apps and sites. | None named |
| banking | A bank looks for a hardware trait to link a login to a stolen or cloned device. | None named |
| games | A game with strict anti cheat checks for devices reused by a banned player. | None named |
| analytics | A cookie free counter uses clock skew to count unique devices. | None named |
| developer | A browser maker adds noise to timers so these tricks lose accuracy. | Chrome, Firefox, Safari |
| shopping | A ticket site tries to find a scalper using one laptop with many fake browsers. | None named |
Risks
- Hard to hide a hardware trait
- Mostly research so far
Good uses and abuse
Legitimate uses
- None
Known abuse
- Device-bound identification not reset by clearing cookies
Scores
| Score | Value | Meaning |
|---|---|---|
| Sensitivity | 3 of 5 | How revealing the data is |
| Utility | 1 of 5 | How useful it is to honest apps |
| Friction | 0 of 5 | How much the user has to do |
| Fingerprint bits | 5 | Rough identifying entropy, when known |
History
- 2018-10, paper: Sanchez-Rola et al. Clock Around the Clock (CCS 2018)
Related capabilities
Related
- CPU benchmarking and device class. Can run a short speed test to guess how fast your device is
Sources
- Sanchez-Rola et al., Clock Around the Clock (CCS 2018)
- Kohno et al., Remote Physical Device Fingerprinting (2005)
How well checked: Based on official documentation.