Research On The Security Mechanism Of P2P Networking Monitoring And Trust Model

With the rapid development in the past decades, the P2P networks have been one of the most popular Internet application systems. In the public networks, P2P users and resources obtain the heterogeneous, dynamics and distributed natures. Due to these natures, the P2P systems are very popular and meanwhile bring in novel security threats. However, the traditional security defensing mechanism which is based on centralized management can not effectively constrain the propagation of malicious behaviors. Therefore, considering novel security threats in the P2P environment, in order to restrict malicious behaviors and improve the security, researching on network monitoring and trust model have become hot topics, which have been widely concerned by academe and industry.The establishment of the trust model can identify the malicious users through the collection and analysis of the interaction experiences among the whole network users. However, the malicious users, who have analyzed the working mechanism of trust model, also develop corresponding malicious schemes, such as whitewashing, team cooperation and joint fraud. Therefore, it is a challenge to design a convenient and effective trust model which could solve problems such as the construction and deployment of the trust model, the collection of user experience, the calculation of the value of trust, the initialization of the trust value, the user’s incentive and so on. Meanwhile, in the case of network monitoring, P2P traffic identification is needed, so as to analyze and restrict the network. However, to resist the monitoring and restrictions, P2P software producer and malicious users are also armed with corresponding hacking technologies, such as dynamic ports and symbol substitution. This makes it harder to design monitoring system. In addition, the spreading of rumors has become the pain point of P2P social networks. The propagation and suppression of rumors becomes a hotspot of research. Therefore, in order to limit the malicious users and improve the security of P2P network, this dissertation deeply analyzes the defects and threats in P2P network systems, and designs effective monitoring systems and trust models and models the security environment for multi P2P social networks based interdependent networks.In this dissertation, we address security mechanism and propose security monitoring, trust model and rumor restriction in various P2P scenarios:Firstly, in P2P e-commerce scenario, a promise incentive trust model based on hierarchical fuzzy system is proposed. Based on fuzzy logic theory, by considering eight factors for security, capabilities and promises, the hierarchical fuzzy system is taken to calculate the direct trust value. In this way, the initial reputation and complexity is improved. Also, the evaluation of trust is more objective and reasonable to satisfy the user’s requirement. Taking advantage of eliminating the unreliable recommendation, the constantly updated fuzzy close degree weighted recommendation trust values are adopted to aggregate the global trust value. Simulation results show that the proposed trust model is available and robust.In P2P file sharing scenario, a topological potential based recommendation trust model is proposed in a P2P community architecture. Having analyzed the P2P users’ distribution model, we propose a topological potential based recommendation trust model to integrate the influences, transactions and reputations of nodes. In this model, the trust metrics are divided into intra- and inter- community computing mechanism. Moreover, the algorithm of selecting super node is presented. Simulation results show that our trust model is effective and robust.In addition, a cross layer P2P traffic identification scheme based on IP filter is designed to make up the defects of the existing identification technology. From this view, we design and implement a security monitoring system based on mobile Internet platform. The data acquisition module, data analysis module, P2P traffic identification and monitoring module, database module and network analysis module are designed and analyzed.Moreover in P2P instant messaging scenario, the spam message filtering system is designed. Having analyzed the spam message propagation mechanism, the spam message filtering system is designed based on social network analysis. And the deployment of monitors is proposed, which shows good performance in the system experiment.In the last P2P social networks scenario, the security of interdependent networks, which is consisted of multi network cooperation and dependency, is analyzed by complex network theory. The processes of cascading failure and rumor spreading are modeled. And then the models of cascading failure under target attack and SIDR rumor propagation are proposed. Finally the security measures are raised according to different security policies.In terms of the security threats in P2P network, we analyze and research the traffic identification and monitoring, spam filtering, trust model, incentive mechanism, cascading failures protection and rumor spreading restriction, based on real-time communication, electronic commerce, file sharing and social network scenarios respectively. Note that our proposed security schemes not only recover the P2P system vulnerabilities, but also enhance the P2P network security.