Research On Fuzz Testing Of DVB-RCS Protocol Based On Reinforcement Learning Algorithm

With the continuous development of information technology and human information transmission needs,satellite communication technology has gradually become one of the three major means of information transmission in the field of communication today,and has become one of the important strategic resources for national political and economic development.DVB-RCS protocol is a typical representative of satellite communication technology,which improves the efficiency and reliability of data transmission through satellite network,and has become an important part of civil and military communication.DVB-RCS protocol has the characteristics of complex protocol stack and variable protocol operation state,which can be customized and flexibly configured according to the actual use requirements.The current security research on DVB-RCS protocol is mostly focused on its communication mechanism and encryption mechanism,and not enough attention is paid to the design and implementation of the protocol itself.Fuzzy testing is currently an important means to detect the security of the protocol.The existing methods for protocol fuzzy testing focus on protocol inverse analysis and test sample generation,which are not suitable for the abovementioned security research needs of DVB-RCS.At the same time,there are disadvantages of blindness and randomness in the sample variation aspect of traditional fuzzy testing,which leads to the efficiency of fuzzy testing for protocols being affected.Second,protocols usually have state transfer characteristics,and state traversal fuzzy testing based on predecessor sequence guidance is usually used nowadays,which has strict requirements on the order and structure design of test samples and increases the difficulty of fuzzy testing.Based on the above background,this paper designs and implements DRFUZZ,a fuzzy testing system for DVB-RCS protocol,and proposes a fuzzy testing technique based on reinforcement learning algorithm to improve the code coverage of the full state of DVB-RCS protocol.DRFUZZ uses reinforcement learning algorithm to optimize the sample variation in the fuzzy testing process,and reduces the randomness and blindness of the sample variation by considering the performance of the tested system and the protocol state change after sending the test samples,and achieves the path-guided state-traversal fuzzy testing of DVB-RCS protocol.After the experimental comparison and analysis,DRFUZZ can optimize the test sample generation,effectively improve the code coverage in the testing process,and can detect vulnerabilities more efficiently,which proves the feasibility and effectiveness of the ideas in this paper.At the same time,the research content and system implementation of this paper,which are extensible and general,provide ideas for optimizing the fuzzy testing process of other stateful protocols.