基于改进帕累托法则的CPU时钟指纹模型

evice fingerprints based on software information, such as browser fingerprints, have the characteristics of easy modification and poor robustness. In order to make up for the shortcomings of device fingerprints based on software information, this paper proposes hardware fingerprints based on CPU clock characteristics. The internal clock of the CPU uses quartz crystals to make oscillators. Small changes in these crystals can cause small but measurable differences in the clock frequency. Use the fCFMAP: CPU Clock Fingerprint Model Based on Advanced Pareto Principleunction execution time as the basic expression of the CPU clock fingerprint, and its distribution follows the Pareto Principle, under low CPU load. Proposes an advanced Pareto Principle that is adaptively adjusted according to the CPU load, and constructs a more robust CPU clock fingerprint model, CFMAP (CPU Clock Fingerprint Model Based on Advanced Pareto). CFMAP solves the problem of instability of existing CPU clock fingerprints under high load. Experiments show that the average False Rejection Rate (FRR) is 1.32%, and the False Acceptance Rate (FAR) is 2.52%.