IoT Gateway White List Creation via Segmented ML and Rule Extraction
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Creating white lists for IoT gateways is a time-consuming process, leaving IoT devices vulnerable during the creation period, as machine learning-based approaches can take several days and are not always effective in securing access control.
Innovation Solution
A creation apparatus that collects information on IoT devices and their white lists, extracts lists based on specific conditions such as the number of devices or installed locations, and couples them to create a tentative white list, which is then applied to IoT gateways for quick and secure access control.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Extent of automation
If machine learning is used to create white lists, then the white lists can be created automatically, but it takes several days to complete the creation process
Solution Approach 1:
The patent segments the white list creation process into two distinct phases: (1) a machine learning phase that runs in the background to learn communication patterns, and (2) a rule-based extraction phase that quickly generates actionable white lists from the learned patterns. This segmentation allows the time-consuming ML training to occur separately from the time-critical white list generation, resolving the contradiction between automation and time consumption.
Solution Approach 2:
The system performs preliminary machine learning training in advance to build a model of normal communication patterns. This preliminary action creates a knowledge base that can be quickly queried later to generate white lists without requiring lengthy processing at the moment of creation, thus achieving both automation and speed.
2Extent of automation
If machine learning is used to create white lists, then automation is achieved, but the creation process takes several days which leaves IoT devices vulnerable
Solution Approach 1:
The patent implements a dynamic white list management system where the white list is continuously updated in stages. Instead of waiting for complete ML training, the system generates interim white lists based on extracted rules and gradually refines them as more data becomes available. This dynamic approach maintains security reliability by providing protective coverage throughout the creation process rather than leaving a security gap.
Solution Approach 2:
The system applies preliminary security measures by extracting rules from initially collected communication data and generating a preliminary white list before the full machine learning process completes. This preliminary anti-action provides immediate security protection while the longer-term ML model continues to develop, preventing vulnerable periods without sacrificing automation.
3Productivity
If white lists are created quickly using extracted rules, then security response time is improved, but the comprehensiveness and reliability of the white list may be reduced
Solution Approach 1:
The patent merges two different approaches to white list creation: the rule-based extraction method that provides quick results and the machine learning method that provides comprehensive analysis. By combining both methods, the system achieves both speed and reliability - the rule-based extraction delivers immediate white lists for quick deployment, while the ML component continuously refines and complements these lists to ensure completeness and accuracy.
Solution Approach 2:
The system implements feedback mechanisms where the initially generated white lists are monitored for effectiveness, and this feedback is used to refine both the rule extraction parameters and the machine learning model. This continuous feedback loop ensures that quickly created white lists improve in accuracy and completeness over time, maintaining reliability while preserving the speed advantage.
Data Source
AI summary
A collection unit (15a) collects information on IoT devices connected to IoT gateways and white lists of the IoT devices, retained by the IoT gateways. An extraction unit (15b) extracts white lists of IoT devices that satisfies a prescribed condition related to the number of the IoT devices of each model or the number of installed locations of the IoT devices of each model from the collected white lists of the IoT devices using the collected information on the IoT devices so as to create a tentative white list. A coupling unit (15c) couples the created tentative white list and the white lists retained by the respective IoT gateways together so as to create a white list applied to the respective IoT gateways.