Abstract

The goal of this research is to validate clock skew based device fingerprinting introduced in 2005 and explore the feasibility of its usage and/or modification to facilitate unique device identification across heterogenous target devices with improved accuracy and reduced errors. Our network consists of 212 devices that include desktops, laptops and handhelds. We conduct a systematic evaluation of the clock-skew fingerprint stability across 3 primary dimensions namely, change in target host environment, configuration and measurement methodology. We also investigate parameters that affect clock skew of a device. Our results indicate that a minimum of 70 packets are required to achieve a stable skew estimate. We also observe a significant difference between desktop and handheld clock-skew behavior with the factors affecting skew estimates being handheld power state and NTP updates. Thus, for a moderate …

Date

2012

Authors

Swati Sharma, Alefiya Hussain, Huzur Saran

Book

Proceedings of the 2012 Workshop on Learning from Authoritative Security Experiment Results

Pages

9-18