As the Internet grows, the complexity of the routing system increases and it is need to analyze the routing behaviors of the Internet. Routing technology among inter-domains is one of the crucial technologists that can affect network performance. Therefore it is important to understand inter-domain routing behavior and network topology.
The Internet routing instability, of the rapid fluctuation of network reachability information, is an important problem currently facing Internet engineering community. High levels of network instability can lead to packet loss, increased network latency and time to convergence. At the extreme, this instability could result in the loss of internal connectivity in wide-area networks.
In this thesis, in order to analyze inter-domain topology variation, we measure the day-to-day variation of domain degree and its court over time by using linear regression model, and discuss the variation of inter-domain topology. We also propose fast BGP convergence policy model to reduce the effects of BGP routing instability. The proposed model is based in state automaton and we verify it though the analysis of BGP routing table gathered from major IXPs in the Internet during two years. The proposed model can be used to identify potential convergence problems, and to determine which kind of routing policy is adequate to reduce transition period when BGP routing instability or network configuration change happens.
Our proposed model can be used to set up routing policy in domains for the purpose of minimizing the effects and the propagation of BGP routing instability.