๐Ÿ‡จ๐Ÿ‡ณStalecollected in 15m

New side-channel attack tracks browsing via SSD activity

New side-channel attack tracks browsing via SSD activity
PostLinkedIn
๐Ÿ‡จ๐Ÿ‡ณRead original on cnBeta (Full RSS)

๐Ÿ’กA new hardware-level side-channel attack that bypasses standard browser privacy protections.

โšก 30-Second TL;DR

What Changed

Uses SSD data request timing to infer user activity

Why It Matters

This vulnerability poses a significant risk to privacy-focused applications and browser security, necessitating potential OS or firmware-level mitigations.

What To Do Next

Review your browser's security settings and consider disabling high-precision timers if you are building privacy-sensitive web applications.

Who should care:Researchers & Academics

Key Points

  • โ€ขUses SSD data request timing to infer user activity
  • โ€ขOperates via browser-based code without traditional tracking scripts
  • โ€ขExposes a new hardware-level privacy vulnerability in modern storage

๐Ÿง  Deep Insight

Web-grounded analysis with 11 cited sources.

๐Ÿ”‘ Enhanced Key Takeaways

  • โ€ขThe attack, dubbed FROST (Fingerprinting Remotely using OPFS-based SSD Timing), was developed by researchers at Graz University of Technology in Austria.
  • โ€ขFROST exploits the Origin Private File System (OPFS) API, a browser feature that allows websites to create and store large files on a user's local disk without requiring explicit permissions.
  • โ€ขThe technique demonstrated high accuracy on a test Mac, correctly identifying visited websites with approximately 89% accuracy and running applications with roughly 96% accuracy.
  • โ€ขThe attack works across different browsers (e.g., a malicious page in Chrome can monitor activity in Safari) because the contention occurs at the shared physical storage hardware level.
  • โ€ขBrowser vendors, including Google, Apple, and Mozilla, were notified of the findings; however, Google reportedly does not classify fingerprinting as a security vulnerability, and Apple deemed the attack 'currently out of scope,' suggesting immediate browser-level fixes are unlikely.

๐Ÿ› ๏ธ Technical Deep Dive

  • The attack initiates by a malicious website creating a large file (up to 60% of total disk space, potentially over 150GB on a 256GB drive) within its Origin Private File System (OPFS) sandbox.
  • To bypass the operating system's page cache, this OPFS file must be larger than the system's available RAM, forcing continuous random 4KB reads to directly access the physical SSD.
  • When other applications or browser tabs perform disk I/O, they compete for SSD access, causing measurable latency spikes in the attacker's ongoing read operations.
  • To achieve high-resolution timing measurements, the attack configures Cross-Origin Opener Policy (COOP) and Cross-Origin Embedder Policy (COEP) HTTP headers, which unlock nanosecond-level timers (e.g., performance.now()) in the browser, bypassing default timer-fuzzing protections.
  • The collected timing patterns, representing I/O signatures, are then fed into a pre-trained convolutional neural network (CNN) to classify and identify specific websites or applications running on the victim's system.
  • Researchers also demonstrated a covert channel using FROST, achieving data exfiltration throughputs of 661.63 bit/s on Linux and 891.77 bit/s on macOS.
  • The full fingerprinting attack was primarily tested on an M2 Mac Mini with 8GB of RAM and a 256GB SSD, and its effectiveness requires the OPFS file to reside on the same physical SSD as the monitored activity.

๐Ÿ”ฎ Future ImplicationsAI analysis grounded in cited sources

Browser privacy models will require re-evaluation to incorporate hardware-level side-channel threats.
Existing browser anti-fingerprinting mechanisms are ineffective against this hardware-based attack, necessitating a fundamental shift in how privacy threats are modeled.
Browser vendors will face a significant dilemma between user privacy and web application performance.
Restricting the precision of storage API timing to mitigate FROST could degrade the performance of legitimate web applications that rely on such measurements.
Hardware manufacturers may introduce SSDs with built-in timing obfuscation as a privacy countermeasure.
The emergence of this attack could spur the development of new hardware features designed to randomize or obscure SSD timing signatures, creating a new market for privacy-enhanced storage.

โณ Timeline

1990s
Discovery of Differential Power Analysis and early timing attacks on cryptographic algorithms.
2017
Meltdown and Spectre CPU vulnerabilities highlight cache-based side-channel attacks.
2026-05-28
FROST (Fingerprinting Remotely using OPFS-based SSD Timing) attack, developed by Graz University of Technology, is publicly disclosed.

๐Ÿ“Ž Sources (11)

Factual claims are grounded in the sources below. Forward-looking analysis is AI-generated interpretation.

  1. tomshardware.com
  2. hannesweissteiner.com
  3. gbhackers.com
  4. techradar.com
  5. techspot.com
  6. slashdot.org
  7. easternherald.com
  8. cyberpress.org
  9. daringfireball.net
  10. aiweekly.co
  11. hothardware.com
๐Ÿ“ฐ

Weekly AI Recap

Read this week's curated digest of top AI events โ†’

๐Ÿ‘‰Related Updates

AI-curated news aggregator. All content rights belong to original publishers.
Original source: cnBeta (Full RSS) โ†—