The impact of FQDN database updates on name-based routing architecture

Abstract

Sophisticated routing protocols, such as content-based routing, have been studied and proposed to replace Internet Protocol (IP) in the future Internet. Routing based on content will make finding, querying, and retrieving the desired information possible within the routers rather than having the routers forward the packets to data servers or domain name servers. However, one of the main reasons that has prevented achieving content-based routing within the routers in the past has been their storage constraints. In our previous work, we showed that name-based routing, which is a good example of content-based routing, is feasible in practice utilizing Ternary Content Addressable Memory (TCAM). Using our proposed algorithms based on longest alphabet match and Distributed Hash Table (DHT), the Fully Qualified Domain Name (FQDN) database was distributed among less than the number of routers that are currently existing in the Internet. In this paper, we evaluate the feasibility of robustly updating the routing information database entries and show that there are no drastic changes in the routers' forwarding tables. We propose an algorithm for mapping a logical network topology to the physical topology to achieve robustness against dynamically changing databases that also considers the characteristic of TCAMs.