The Research Of Collaborative Intrusion Detection Systems Based On Blockchain

With the diversification and complication of network attacks,the security of computer networks has attracted more and more attention.Our networks are constantly attacked in our daily life and information is leaked.In order to solve these security problems,both the industry and academic are attaching increasing attention on network security.Because the network is interdependent,and the security information of the external global network has been found to be very useful for the assessment of the security risk of the internal network and also useful for finding intrusion activities of the internal network as early as possible and more accurately,the collaborative intrusion detection technology is proposed in a distributed network to improve the accuracy of intrusion detection and system scalability.Although researchers have made a lot of breakthroughs in collaborative intrusion detection technology,data sharing and trust computing still remain as the two major challenges of collaborative intrusion detection technology given that not all participants want to share their information.In addition,mutual trust means that participants must trust each other and not disclose data to the third party when sharing data,Moreover,participants prefer to transmit private data anonymously.Once participants begin to share data anonymously,the system will be untrustworthy.Therefore,this paper presents an in-depth analysis of these problems and corresponding improvement plans.The specific research work includes the following three aspects:Firstly,a general architecture that incorporates block chains into the field of collaborative intrusion detection system(CIDS)is implemented,which enables the collaborative intrusion detection system to quickly detect network attacks.Secondly,we compare and analyze all data sets of current intrusion detection,aiming to figure out and summarize the shortcomings of these data sets.Based on the NSL-KDD data set,we generate a new generated data set by using the generation countermeasure network technology.It overcomes the three main problems of data redundancy,data imbalance and not including the latest attacks in intrusion detection data sets.Finally,we analyze and compare the single-point detection performance of various classifiers,improve the real-time detection rate of the classifier of cooperative intrusion detection system by using deep learning method,and simulate the cooperative intrusion detection system in distributed network,so that the performance of the classifier can be improved and hence the new attack types can be detected.