Vulnerabilities Mining In Industrial Networks Based On Fuzzing

With the development of industrial network and internet technology, the connec-tion between industrial network and traditional IP network is getting closer and closer.Now, the traditional industrial network and IP network usually construct a hybrid net-work in many conditions. While traditional industrial networks and IP networks are linked more and more closely, the security threats to industrial networks are becoming more and more prominent. In recent years, viruses and Trojans, which are specifically designed to attack industrial networks, are becoming more and more popular. Therefore,to discover vulnerabilities in industrial network by fuzz testing (Fuzzing) and improve the security level of industrial network, has a very important significance.Now,there are some mature frameworks of fuzz testing,which can generate a large number of effective test cases to find vulnerabilities in a variety of software products and network equipment. At the same time, some researchers apply fuzz testing to the vulnerability mining of industrial network. But relatively speaking, many frameworks of fuzz testing have their own drawbacks, such as low efficiency of the test case gener-ation algorithm and inefficient monitoring methods, so they can't exploit vulnerabilities in the service subnets and control subnets of industrial network effectively.Based on the research of industrial network and the technology of vulnerability mining, this paper analyzes the loopholes which may exist in industrial network, make some improvement for current fuzz testing technology schemes, and then design a vul-nerability mining system which can be applied in heterogeneous industrial network.This system can take combined vulnerability mining for the target. At the same time,this paper propose a test case generation algorithm-MFTCG (Multidimensional Fuzz-ing Test Case Generator) for this system. This algorithm can make multidimensional variation of different fields in the network protocol, to discover vulnerabilities which only can be triggered by multidimensional random mutations. Moreover, the system can effectively filter the test cases to improve the efficiency of generating test cases.And after that, it can take combined vulnerability mining for the target. According to the different targets, the software module of the vulnerability mining system can mon-itor the targets by different means. While this module was deployed in the industrial network, several vulnerabilities were found in the file server and the industrial protocol conversion gateway. If these vulnerabilities are exploited by the attacker, it will cause some serious consequences,such as serious data loss and the parafunctions of industrial control network.This project is supported by the Electronic Development Fund of Ministry of In-dustry and Information Technology. At present, the software module of vulnerability mining system based on Fuzzing has passed the test of the national authoritative com-munication security software organization, which was organized by the partner. The author has published a related paper "Vulnerabilities Mining In Industrial Network Based On Fuzzing", while participating in the application of one patent.With the increasing integration of IP networks and industrial networks, the security of heterogeneous networks will become more and more prominent. As an effective se-curity protection technology, a more in-depth application of vulnerability mining will be taken in the area of industrial networks security. This paper made some useful at-tempts, and achieved some initial results.