Research On Recognition And Vulnerability Scanning Of IoT Devices

With the rapid development of the Internet of Things technology,a large number of devices are networked and exposed in cyberspace.Due to the uneven security design level of Io T devices,and the neglect of management during use,they are easy to become the preferred targets of network attackers.Not only will they cause losses to device owners,they may also be used as botnet nodes.The security of the entire cyberspace poses a threat,and its harmfulness cannot be ignored.Asset identification is the prerequisite for security management of Io T devices,and vulnerability scanning is an important means to discover hidden dangers of Io T devices.This article analyzes and researches existing asset identification and vulnerability scanning technologies,and proposes several improvements or optimization techniques.The main work is as follows:1.In view of the existing device identification methods based on message content features that rely on message content text features and difficulty in large-scale data labeling,a machine learning-based Io T device classification method is proposed,which is extracted from the device's WEB homepage Features and generates feature vector fingerprint space,combined with random forest algorithm to improve the accuracy of Io T device classification.This method is suitable for asset identification of Io T devices,and can provide support for the precise implementation of vulnerability scanning of Io T devices.2.Aiming at the problem of omissions and redundancy in the existing CPE generation method due to the limited software version number format or relying on the retrieval of the CPE hierarchy tree sequence,an improved CPE generation algorithm is proposed,which combines the flag information of the device service and the NVD The generated CPE dictionary is extracted from the vulnerability database to infer the CPE corresponding to the device,which can be used to retrieve the possible vulnerabilities of the device in the vulnerability database.This algorithm can effectively improve the accuracy of CPE generation,thereby reducing the false negative rate and false positive rate of Io T device vulnerability scanning.3.Aiming at the slow scanning speed and poor scalability of OpenVAS scanner,a vulnerability scanning method based on OpenVAS scanner cluster is proposed.Scanning plug-ins are selected according to the results of device identification and CPE matching vulnerabilities,and through customization The task scheduler deploys scanning tasks to each scanner cluster node according to the strategy,which effectively improves the scanning efficiency of OpenVAS.4.Designed and implemented a prototype of a vulnerability scanning system for Io T devices based on the above technologies.The system adopts the B/S architecture,the web front end is responsible for interacting with users,and the back end is responsible for the function realization of the entire system.The entire system adopts a modular design,which mainly includes 8 modules including asset detection and identification module,vulnerability scanning module based on CPE matching,and vulnerability scanning module based on OpenVAS scanner cluster.Modules use message queues for communication,support task scheduling on distributed processes and nodes,and increase the number of concurrent processes and nodes as needed.5.A system test environment was built to verify that the vulnerability scanning system meets the design requirements in terms of function.Compared with OpenVAS,because the system scans based on the device identification results,plug-in selection is more targeted,unnecessary plug-in operations are reduced,the efficiency of a single IP scan task is at least doubled,and the scan results are exactly the same.For large-scale network scanning,the number of concurrent processes and nodes can be adjusted according to the network scale to further improve scanning efficiency.In addition,some modules of the system use a plug-in design,which is highly extensible.It can quickly write and add plug-ins according to actual needs,and has strong adaptability.