Research On Program Sequence Monitoring Method Of Safety Computer Platform Based On Cloud Computing

With the development of railway construction in China,the computing demand of train control system is increasing day by day.The application of cloud computing technology to rail transit signal system has become a research hotspot in recent years.The safety computer is the core of the train operation control system,which can directly affect its performance and has a very important impact on the safe and stable operation of the railway signal system.While the use of cloud computing technology for the safety computer platform can reduce the number of equipment,computing stress and storage stress,it also brings new security challenges.As the cloud platform is in the open network environment,how to improve the safety and availability of the safety computer platform in cloud computing environment is the key problem to apply cloud computing to the train operation control system.This thesis proposes innovative improvements to the architecture of the safety computer platform and to the program sequence monitoring method,so as to improve the safety of the safety computer platform based on cloud computing.The main work completed in this thesis includes:Firstly,considering that the safety computer platform faces many information security problems such as security vulnerabilities and malicious network attacks,a solution of real-time monitoring the operation of computer programs through edge safety node is proposed.At the same time,several homomorphic encryption algorithms are studied and an improved program sequence logic monitoring method based on Paillier homomorphic encryption is proposed.Secondly,in view of the static,homogeneous defects of the existing safety computer platform structure and the defense requirements brought by the cloud environment,a new security defense architecture design scheme with dynamic,heterogeneous and random features is proposed.In the existing architecture,negative feedback control link、dynamic scheduling of hot standby executors and heterogeneous design among executors are introduced,so as to realize the safety computer defense model based on "dynamic heterogeneous redundancy" structure.The logical monitoring flow design is completed by setting reasonable software flow feature code and combining with the addition homomorphism of Paillier algorithm,so as to monitor the execution of the cloud deployment and running program.The key data of the software flow feature code is encrypted to improve the privacy and credibility of data,prevent the flow code information used for monitoring from being forged or maliciously tampered.Then,generalized stochastic Petri net is used to establish the attack fault model of the safety computer platform,and the available probability and escape probability of the platform are solved quantitatively by continuous time Markov chain.By comparing and analyzing the probability indexes of the platform under different attack intensities,heterogeneity and reconstruction recovery rates,it is concluded that the dynamic heterogeneous redundant structure with encryption security management has higher availability and lower escape probability,so as to achieve better anti-attack effect.It provides a way of how to improve network security protection ability of train control safety computer system in the dangerous cloud computing environment.Finally,the experimental platform which is based on the dynamic heterogeneous redundancy architecture is built using Open Stack cloud computing management platform.The logical monitoring function based on the software process code in ciphertext and the dynamic scheduling of hot standby executors are implemented.Finally,the edge safety node can effectively detect program anomalies by the experiment of forgery and tampering software process feature code,which proves that the program sequence monitoring scheme designed in this thesis has certain defensive effect against threats such as data tampering.In this thesis,there are 50 figures,11 tables and 90 references.