Edge Device Fingerprinting for Low-Latency Network Control

Resolve Bottlenecks,
Find Innovative SolutionsGenerate Solutions

Solution Overview

Problem

Existing device fingerprinting methods in local networks rely heavily on cloud-based services, leading to increased latency, bandwidth usage, and costs, which can compromise network performance and security, especially with the rise of MAC randomization making cache-based policies ineffective.

Innovation Solution

Implementing a fingerprint determination model within the network management device to determine device characteristics locally, reducing reliance on cloud services by using machine learning classifiers like SVMs and ANNs to process device message characteristics and control network functions.

Engineering Contradictions & Design Principles

VSEngineering Contradiction Analysis

1Measurement precision

If cloud-based device fingerprinting services are used, then device identification capability is improved, but latency increases and bandwidth usage increases

Engineering Contradiction:

Improvedevice identification accuracyVSAvoidfingerprinting latency

Core Design Contradiction:

Measurement precisionVSLoss of time

Solution Approach 1:

The patent segments the device fingerprinting process by deploying local fingerprinting agents on network devices (routers, switches, access points) that independently collect and analyze device characteristics. This segmentation allows fingerprinting to occur locally without requiring constant cloud communication, thereby reducing latency while maintaining identification accuracy through distributed processing across multiple network nodes.

Inventive Principle:

Principle #1Segmentation

Solution Approach 2:

The patent introduces local fingerprinting agents as intermediaries between devices and the cloud service. These agents collect device characteristics locally and can perform preliminary fingerprinting analysis, reducing the need to transmit raw data to the cloud and thereby reducing both latency and bandwidth usage while maintaining identification accuracy.

Inventive Principle:

Principle #24Intermediary (Mediator)

2Measurement precision

If cloud-based device fingerprinting services are used, then device identification capability is improved, but bandwidth usage increases

Engineering Contradiction:

Improvedevice identification accuracyVSAvoidnetwork bandwidth consumption

Core Design Contradiction:

Measurement precisionVSLoss of energy

Solution Approach 1:

The patent extracts the fingerprinting analysis function from the cloud service and implements it locally on network devices through fingerprinting agents. This extraction allows devices to perform fingerprinting independently using local computational resources, eliminating the need to transmit device characteristic data to the cloud and thereby reducing bandwidth consumption while maintaining identification accuracy.

Inventive Principle:

Principle #2Taking out (Extraction)

Solution Approach 2:

The patent enables network devices to perform device fingerprinting autonomously through local agents that collect, process, and analyze device characteristics independently. This self-service capability eliminates dependency on cloud-based processing for routine fingerprinting operations, reducing bandwidth usage while maintaining identification accuracy through local analysis.

Inventive Principle:

Principle #25Self-service

3Object-affected harmful factors

If MAC randomization is implemented, then device privacy is improved, but device tracking capability deteriorates

Engineering Contradiction:

Improvedevice privacy protectionVSAvoiddevice tracking accuracy

Core Design Contradiction:

Object-affected harmful factorsVSMeasurement precision

Solution Approach 1:

The patent implements fingerprinting agents on multiple types of network devices (routers, switches, access points, controllers) that collectively observe and track devices across different network segments and time periods. This multi-functional distributed approach enables continuous tracking of devices even when MAC addresses change, as the same physical device will exhibit consistent behavioral patterns across different access points and time intervals, thereby maintaining tracking accuracy while preserving privacy through MAC randomization.

Inventive Principle:

Principle #6Universality (Multi-functionality)

Solution Approach 2:

The patent ensures continuous device observation and fingerprinting collection across multiple network devices and time periods. By maintaining persistent monitoring through distributed agents that track device behaviors, connection patterns, and usage characteristics over time, the system can accurately identify and track devices even when MAC addresses are randomized, as the continuous collection of behavioral data reveals consistent device identities.

Inventive Principle:

Principle #20Continuity of useful action

Data Source

PatentUS12627564B2Device classification at the edge

Publication Date: 2026.05.12 CAMBIUM NETWORKS

AI summary

A method, device, and storage medium are provided and configured to control one or more network functions for user devices in a local network. A fingerprint determination model is provided and configured to determine one or more characteristics of user devices based on messages generated by respective user devices. A first message from a user device is received and processed to determine a set of one or more user device message characteristics. Device fingerprint data, representing at least one determined characteristic of the user device, is generated by processing the set of one or more user device message characteristics using the fingerprint determination model. One or more network functions for the user device are controlled based on the at least one determined characteristics of the user device represented in the device fingerprint data.