03

Browser Signal Stability

Evaluating Browser Signal Stability for Device Identification: A Controlled Laboratory Study Motivated by Nepali Mobile-Web Payments

Submitted for publication · 2025

Researcher - Experimental Data Collection & Analysis
Browser fingerprintingDevice identificationMeasurement methodologySystems securityData collection
Overview

The original research direction came primarily from another member of the research team. The study investigated whether browser-derived signals could provide sufficiently stable device identification under controlled laboratory conditions, motivated by the needs of Nepali mobile-web payments.

Research Question

Are browser-derived signals stable enough to serve as a basis for device identification in a controlled laboratory setting, and how consistent are repeated measurements across different browsers and machines?

My Contribution
  • Participated in the experimental implementation and evaluation of the browser/device-identification system.
  • Led significant portions of the physical data-collection process, including conducting experiments across college computer laboratories.
  • Collected the repeated browser/device measurements used in the experimental dataset.
  • Participated in cleaning, analyzing, and interpreting the collected measurements.
  • Helped translate experimental results into figures, findings, and conclusions for the paper.
  • Took substantial responsibility for the final organization, technical consistency, and presentation of the manuscript.
  • Contributed to the final review and refinement of the paper before submission.
Methodology
  • Designed a controlled laboratory study in which browser-derived signals were collected repeatedly across multiple machines and browsers.
  • Conducted data-collection sessions across college computer laboratories to build the experimental dataset.
  • Analyzed the consistency and stability of the collected signals over repeated measurements.
  • Assessed whether the signals were stable enough to support reliable device identification.
Key Findings
  • Browser-derived signals showed measurable variation across machines and browsers in the controlled laboratory setting.
  • The stability of the signals depended on the browser and environment, informing how much these signals can be relied on for device identification.
  • The results clarify the practical limits of using browser signals for identifying devices in mobile-web payment contexts.
What I Learned
  • How to design and run a controlled measurement study.
  • The realities of browser fingerprinting and device identification.
  • How to turn raw measurements into figures, findings, and defensible conclusions.
Limitations
  • The study was conducted in a controlled laboratory environment, which may not fully reflect real-world conditions.
  • The original research direction came from another team member, and the findings reflect the controlled setting rather than a deployed system.
Paper / Code / Artifacts

Paper, code, and other artifacts are not linked yet.