BGP Route Anomaly Detection using Fisher-Markov Feature Selection and LSTM Networks

Abstract

According to Statistica.com 4.53 billion about 58.8% of total population of the world are connected to the internet and user of internet. And it is increasing at the rate of 1157%. The Internet is a decentralized global network comprised of tens of thousands of Autonomous System. The Border Gateway Protocol is the internet’s default interdomain routing protocol that manages connectivity among Autonomous Systems .Border Gateway Protocol has been the defacto inter-domain routing protocol since it was introduced and it is the destination based routing nature. Anomaly Detection of BGP traffic is significantly important in improving both security and robustness of the Internet. The objective of this paper is to work on two factor detection and prevention of the anomaly of the Border Gateway Protocol Route using the Deep learning Long Short Term Memory Technique. We developed the model that would classify the anomaly prefix hijack from the Dataset of Border Gateway Protocol Route.