Research Of Negotiation Mechanisms Of ATN In P2P Environments

Automated trust negotiation (ATN) is an effective way to solve the problem of trust establishment across security domains. Negotiation mechanisms are an important research content in ATN, including policy language, negotiation strategy and negotiation protocol. The main problem faced by a researcher in the existing negotiation mechanisms are, policy language which isn't able to meet requirements of ATN systems, and negotiation strategies which have low negotiation efficiency and negotiation security. Therefore, designing a kind of policy language which is capable of eliminating the problem exists in the present policy languages, making an improved negotiation strategy and negotiation protocol are the primary focus of this research.A flexible policy language FPL (Flexible Policy Language), negotiation strategy based on the stack structure and the corresponding negotiation protocol algorithms by the basis of analyzing the existing policy languages and negotiation strategies were designed. For problems which existing policy languages can not meet ATN system requirements and negotiation strategies'efficiency and security are not high, an improved strategy FPL and the negotiation strategy based on the stack structure were designed respectively. In order to enhance the success rate of interaction between the negotiating parties, a Reputation-Based ATN Model (RBATN) was designed, which improves the success rate of interaction between the negotiating parties from quantitative point of view.A kind of flexible policy language FPL was designed, which supports a variety credentials and access control policies. In FPL, negotiating parties'credentials and access control policies were described in XML, which broadly increases the flexibility and applicability.The negotiation strategy based on the stack structure, which was improved on the basis of DFANS(Deterministic Finite Automation Negotiation Strategy) was designed. This negotiation process was divided into two stages:pre-negotiation, formal negotiation and the ideas of credential sensitivity were added. After making an improvement in the strategy when compared with DFANS negotiation strategy, it possesses the advantage of higher average efficiency and the security, which make more realistic value.A Reputation-Based ATN model (RBATN) was designed, in which the idea of reputation system was added to the traditional model. To improve the success rate of interaction between both parties, it designed the interaction process of both negotiation parties from the first quantitative point of view, in which FPL was used as policy language and negotiation strategy that is based on stack structure. Policy language FPL designed on this paper possesses the advantage of meeting the requirements of ATN systems compared with other policy languages. The simulation experiment result when compares with both negotiation strategies based on the stack structure and other negotiation strategies in the aspect of negotiation efficiency and negotiation security shows that new strategy possesses higher efficiency and minimum disclosure of sensitive credential between the negotiating parties. It also has further improvements in the aspect of securities when compared with DFANS negotiation policies. The simulation experiment with both RBATN model and the traditional model in terms of interaction success rate shows that the interaction success rate of RBATN is higher than the traditional model as number of malicious nodes in P2P environment is increased.