| The movement of data packets through the Internet requires coordination and cooperation between independently managed Autonomous Systems (ASes) controlled by a wide array of organizations with differing goals. The Border Gateway Protocol (BGP) allows routers within these networks to exchange information in order to ensure the proper delivery of packets. However, the basic design of the protocol was formulated when there was an implicit trust relationship between all parts of the network, leading to a situation where false information (either malicious or accidental) could adversely affect the operation of the Internet.;Today, the Internet is not composed solely of trustworthy entities. Malicious and selfish parties knowingly choose to deviate from standardized methods of network operations. Even when an entity intends to follow protocol rules, mistakes and errors can inject incorrect data into the decision-making process.;The core functionality of BGP is to exchange network reachability information between ASes and compute routes to the advertised networks. The core specification of BGP, developed in an era of implicit trust across the Internet, does not contain any type of protection against attacks. Of note, BGP has no way to validate the authority of an AS to announce a network prefix or a mechanism to ensure the authenticity of the path advertised.;The contribution of this thesis is to advance the state-of-the-art in the authentication and authorization of the information carried in BGP messages. The mechanism described herein will provide BGP routers the ability to detect erroneous and malicious changes to the network prefix and path information carried in BGP messages. Experimental results indicate that this approach not only provides a higher level of protection for BGP than other approaches, but also requires far less additional infrastructure to deploy. |
