Device Fingerprint Matching with Adaptive Change Indicators
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing device identification solutions are vulnerable to fraudsters impersonating legitimate devices due to the dynamic nature of device fingerprints, which change with software updates and other modifications, making it difficult to accurately verify device identities and prevent fraudulent transactions.
Innovation Solution
Implementing enhanced device fingerprint matching algorithms that incorporate adaptive, self-learning techniques to generate metrics accounting for expected changes in device fingerprints, performing deep inspections, and using unsupervised learning mechanisms to differentiate between plausible and implausible fingerprints, thereby improving the accuracy of device identification and reducing false positives and negatives.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Reliability
If device fingerprints are used for device identification, then device identity can be verified, but the dynamic nature of device fingerprints causes identification accuracy to deteriorate over time
Solution Approach 1:
The patent implements dynamic fingerprint matching by generating an expected current fingerprint from a historical fingerprint using device change indicators. This allows the system to adapt to legitimate device changes (software updates, OS upgrades) while maintaining identification accuracy. The matching algorithm dynamically adjusts the expected fingerprint based on the type and extent of device changes, rather than using a static historical fingerprint for comparison.
Solution Approach 2:
The system changes the parameters used in fingerprint comparison by introducing device change indicators that quantify the extent of device modifications. Different parameter thresholds are applied based on the type of change detected - for example, software version changes are treated differently from hardware changes. This allows the system to accommodate legitimate parameter changes while detecting fraudulent alterations.
2Reliability
If strict fingerprint matching is used to prevent fraud, then security is improved, but false positives increase causing legitimate transactions to be blocked
Solution Approach 1:
The patent applies different matching thresholds and parameter weights based on the type of device changes detected. For legitimate changes like OS updates or browser version changes, the system uses more permissive matching parameters. For suspicious changes like MAC address alterations or hardware modifications, stricter parameters are applied. This dynamic parameter adjustment reduces false positives while maintaining fraud prevention effectiveness.
Solution Approach 2:
The matching algorithm dynamically adjusts its strictness based on the device change indicators. When changes are consistent with normal device evolution (software updates, security patches), the system applies lenient matching. When changes suggest potential fraud (sudden hardware changes, inconsistent modification patterns), the system increases matching strictness. This dynamic approach balances security with user convenience.
3Measurement precision
If device fingerprints are updated frequently to maintain accuracy, then identification precision is improved, but device complexity and processing requirements increase
Solution Approach 1:
Instead of completely regenerating and storing full device fingerprints after every change, the system performs partial updates by generating change indicators that represent specific aspects of device evolution. The matching algorithm then applies only the necessary adjustments to the expected fingerprint based on these partial change indicators. This reduces processing complexity while maintaining accuracy.
Solution Approach 2:
The patent segments the fingerprint matching process into two independent components: (1) generating device change indicators from historical and current device states, and (2) applying these indicators to adjust the expected fingerprint for comparison. This segmentation allows each component to be optimized independently and reduces overall processing complexity while maintaining high accuracy.
Data Source
AI summary
A method of identifying a device includes receiving a device transaction request from a remote device, receiving a first device fingerprint of the remote device, and receiving a second device fingerprint of a known device. The first device fingerprint is compared with the second device fingerprint and a first metric indicative of a similarity of the first device fingerprint and the second device fingerprint is generated. A third device fingerprint corresponding to an expected current value of the second device fingerprint is generated, and the first device fingerprint is compared with the third device fingerprint to generate a second metric indicative of a similarity of the first device fingerprint and the third device fingerprint. A response to the transaction request is formulated based on the first metric and the second metric.