Industrial control network protocol vulnerability mining method based on inverse sample

Abstract

The invention discloses an industrial control network protocol vulnerability mining method based on an inverse sample, and the method comprises the steps: taking a Modbus TCP protocol message as a test case, and constructing a vulnerability mining test case generation model of an industrial control system; learning semantics of a protocol data unit through a recurrent neural network, and representing probability distribution of data values by using a Softmax layer with temperature parameters; and then comparing the relationship between the random variable threshold and the maximum probabilityto determine whether to replace the current data value with the data value with the minimum probability so as to increase the possibility of industrial control abnormality, and finally supplementing aprotocol MBAP message header part with a random value according to protocol specifications to form a complete test case. According to the method, the problems of high redundancy and low efficiency ofan industrial control network protocol can be solved, and the case vulnerability mining capability is improved.

Description

technical field [0001] The invention relates to the field of industrial control network security, in particular to a method for semantic learning and vulnerability mining of industrial control network protocols based on counter samples. Background technique [0002] With the introduction of Internet technologies such as Industry 4.0, industrial control systems continue to closely integrate automation and intelligence, and network security risks also follow. Industrial control systems are widely used in energy, water conservancy, transportation, manufacturing and other fields, and the number of exploiting vulnerabilities to attack industrial control equipment continues unabated. [0003] In recent years, there have been multiple serious security vulnerabilities in the MicroLogix 1400 series PLC of Rockwell Automation, causing denial of service attacks, tampering with device configuration and other attacks; the website of the Ministry of Energy and Coal Industry of Ukraine was...

AI Technical Summary

Problems solved by technology

The research on vulnerability mining technology in the security field at home and abroad has achieved some success, but the traditional industrial control network protocol test case generation method applied to the industrial control vulnerability mining process has problems of high redundancy and low efficiency.

Images