Research On Security Technology Of Industrial Internet Protocol Based On Deep Learning

With the deep integration of the new generation of communication technology and the industrial economy,the Industrial Internet has become the cornerstone of the next industrial revolution.With the continuous development of the industrial Internet,security accidents occur frequently,which shows that there are huge security challenges hidden under the vigorous development of the industrial Internet.As the infrastructure of the industrial Internet,the security of the industrial Internet protocol is particularly important.However,there are few researches on the security of industrial Internet protocols at present,and industrial Internet protocols may face potentially huge security threats.Vulnerability mining technology can discover potential security threats in advance.In order to ensure the security of industrial Internet protocols,it is an important means to ensure the security of industrial Internet to carry out vulnerability mining in order to find possible security holes as early as possible and repair them in time to prevent them before they happen.currently Among the mainstream vulnerability mining methods,fuzz testing technology is the most widely used and effective method nowadays.However,the current mainstream fuzz testing methods for network protocols have their limitations,such as requiring prior knowledge of the protocol,low reliability,low degree of automation,and poor effect.Therefore,this thesis designs a fuzzing method for industrial Internet protocols based on deep learning.This method is suitable for fuzz testing of various industrial Internet protocols,and has the advantages of no need for prior knowledge,high reliability,high degree of automation,and good vulnerability mining effect.The work of this thesis mainly includes the following contents:1.Aiming at the problem that mainstream network protocol fuzzing tools require prior knowledge of protocols,low reliability,and low degree of automation,a fuzzing method for industrial Internet protocols based on AFL(American Fuzzy Loop)tool is designed.The method requires no prior knowledge of the protocol,is highly automated,and has high reliability.2.Aiming at the problem that the fuzzing effect of AFL tool greatly depends on the initial seed,a seed selection method based on deep learning is designed.This method uses the convolutional neural network to extract the eigenvalues of the industrial Internet protocols,and then quantitatively calculates the similarity between the industrial Internet protocols based on the eigenvalues,and finally calculates the difference between each industrial Internet protocol and other industrial Internet protocols based on a large number of samples.The similarity threshold of the Internet protocol is used to guide the initial seed selection of the AFL tool.This thesis has carried out experimental verification on the method proposed above,and the verification results show that the fuzz testing method proposed in this thesis has a high degree of automation,high reliability,and good vulnerability mining effect,which can greatly improve the efficiency of vulnerability mining for industrial Internet protocols,so as to better guarantee the security of the industrial Internet.