Research On Detection Methods Of Abnormal Behavior In Security For IIoT Based On Random Inspection

With the deep integration of industrialization and informatization,Industrial Internet of Things(IIo T)is facing an increasingly severe information security problem while getting rapidly promotion.In particular,the characteristic of IIo T nodes which is vulnerable makes its security highly concerned.Most of the common sensor nodes of the IIo T are deployed in open environments,facing a number of threats from all aspects of the network.Since the common sensor nodes have extremely limited resources,it has become a long-term problem for the safety of IIo T that the common sensor nodes cannot be equipped with high energy consumption safety protection system.The thesis proposes an algorithm for detecting abnormal behaviors in security of the IIo T based on random inspections to achieve low-power and high-efficiency security protection for IIo T.The main works of the thesis as follows:Firstly,an abnormal behavior detection model for safety of IIo T is Established.The typical architecture of IIo T system is analyzed in this thesis,and an abnormal behavior detection model based on the data acquisition layer of IIo T is established which is built on the cluster of the IIo T network.The hierarchical mechanism for detection and management of abnormal behavior is constructed,and packet integrity,packet transmission rate and transmission delay ratio are used to detect abnormal behavior.This thesis introduces a double threshold reputation measurement mechanism,completes the abnormal behavior monitoring of the whole network nodes and ensures the safety and reliability of the communication data.This thesis selects three node communication parameter attributes of packet integrity,packet transmission rate and transmission delay ratio,and uses these three attributes as the detection indicators of abnormal behavior,and introduces a double threshold reputation evaluation mechanism to complete the abnormal behavior of nodes across the network.Ensure the safety and reliability of node communication data.Secondly,this thesis proposed an abnormal behavior detection algorithm based on random inspection.For different types of malicious nodes,the new algorithm manages the number of nodes in the network,so that the sink nodes can more easily inspect the entire network,and then set different inspection video rates according to the reputation value interval.The sink node inspection is used to detect and analyze the existence of nodes in the network.The abnormal behavior improves the randomness and efficiency of the detection algorithm,and set up blacklists of malicious nodes and whitelists of trusted nodes,further reducing network energy consumption.The thesis simulates and analyze on the Network Simulator 2 platform,the abnormal behavior detection algorithms were tested and compared from aspects such as network performance analysis,energy consumption,node reputation value analysis and malicious node identification analysis.Simulation results show that the new abnormal behavior detection algorithm improves the detection accuracy without affecting the life of the network.The proposed random inspection mechanism helps the algorithm find malicious nodes with low energy consumption and high efficiency,and improves the security of the IIoT.