Research On Trust Mechanism For Inter-Domain Routing Control

The aim of Trustworthy and Controllable Network (TCN) is to make the network system behavior to be monitorable, evaluable and predictable, on the basis of which an independent control plane is constructed to realize the trustworthiness and controllability of the network. It is known that the inter-domain routing is the cornerstone of the Internet, which also plays an important role of TCN. In this system, in order to achieve the inter-domain routing convergence, the routing advertisements are exchanged between the Autonomous Systems (ASes) based on the Border Gateway Protocol (BGP). However, BGP works with the assumption that the routing behavior of ASes is credible. Such an assumption leads to the situation that the inter-domain systems are interfered by false routing information, resulting in many network outage incidents. Therefore, how to effectively suppress the generation and dissemination of the false routing information is a challenging and promising issue.Previous solutions to improve the security of BGP or the diagnosis methods of routing information are not sufficient to achieve the objective. Addressing this issue, the trust mechanism has been applied to evaluate on the routing behaviors of ASes, and the evaluation results provide an effective criteria for routing decisions. However, existing trust mechanisms for the inter-domain routing system still have some drawbacks, leading to the situation that the evaluation results cannot accurately reflect the actual routing behaviors of ASes:(1) the prediction methods of routing behaviors are not adopted, so that the evaluation results cannot reflect the future behavior of ASes correctly; (2) there is lack of incentive mechanisms to encourage ASes to share trust information, so that the completeness and accuracy of the evaluation results cannot be guaranteed; (3) there is lack of effective methods to filter the unreliable trust information.In this paper, in order to achieve the accurate trust evaluation on the routing behaviors of ASes, the theoretical researches are conducted in three aspects:the trust evaluation model, the prediction method and the incentive mechanism. First, a trust model for inter-domain routing system is presented, based on which the trust evaluation method for routing behavior of ASes is proposed. Second, in order to guarantee the accuracy of the evaluation results, a prediction algorithm is presented. Third, an incentive mechanism for trust recommendation is presented to effectively encourage ASes to share trust information. Finally, in order to conduct a comprehensive evaluation on the proposed algorithms and verify the effectiveness and feasibility of this work, the inter-domain routing trust control prototype system is designed and implemented on the Southeast University experimental platform.In this paper, the trust mechanism for inter-domain routing control is deeply studied, and an accurate evaluation method is presented, the evaluation results can effectively support the routing decisions so as to suppress the generation and dissemination of the false routing information. Therefore, this work can contribute to guaranteeing the stable operation of inter-domain routing system and enhancing the Internet security.