IoT Code Injection Detection Using Local and Centralized Ledgers
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Large Language Models (LLMs) pose a security risk by enabling AI-based hacker systems to inject malicious code into IoT devices, compromising their control and operation.
Innovation Solution
Implement a verification system that uses a combination of local and centralized ledgers to monitor and verify the behavior of IoT devices, tracking state changes through blockchain technology to detect and prevent code injections.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Productivity
If LLMs are deployed to improve organizational efficiency, then productivity increases, but security vulnerabilities increase due to code injection risks
Solution Approach 1:
The patent introduces a verification system as an intermediary between the LLM and the IoT device. This system includes a local ledger on the device and a centralized ledger that verify code before execution, acting as a mediator that prevents malicious code from reaching the device while still allowing legitimate code to execute, thus maintaining productivity while blocking security threats
Solution Approach 2:
The verification system performs preliminary verification of code integrity before the code is executed on the IoT device. By checking the code against the ledgers in advance and only allowing verified code to run, the system prevents code injection attacks from taking effect, thereby maintaining both productivity and security
2Reliability
If code verification systems are implemented to prevent injections, then security improves, but device complexity increases
Solution Approach 1:
The verification system is segmented into multiple independent components: a local ledger residing on the IoT device, a centralized ledger on the server, and a verification module that coordinates between them. This segmentation allows each component to be simple and focused, reducing overall system complexity while maintaining security through distributed verification
Solution Approach 2:
The system uses cryptographic hashes (copies) of the code rather than storing and verifying the entire codebase. The local ledger stores hash values that represent the expected code state, and verification is performed by comparing hashes rather than full code comparison, significantly reducing the complexity and resource requirements of the verification system
Data Source
AI summary
Techniques related to the management of communication devices using a combination of local and centralized blockchains to enable the detection of state changes that deviate from the expected behaviors are disclosed. In one example aspect, a method for detecting code injection activity in communication devices by a machine learning based platform includes determining, by a verification system, baseline information about a communication device; detecting a state change of the communication device, responsive to the state change of the communication device; appending, by the verification system, a current-state block representing the state change to a prior-state block representing a previous state of the communication device in a blockchain; and detecting whether a suspicious attack has occurred by comparing the payload of the current-state block to the baseline information.