Route leak identification: a step toward making inter-domain routing more reliable

Abstract

Route leaks are one of the anomalies of inter-domain routing that have the capacity to produce large Internet service disruptions. Route leaks are caused because of violation of routing policies among Autonomous Systems. Unfortunately, there are not many studies that formally and thoroughly analyze the route leak problem. There exist few conventional solutions that can be used as a first line of defense, such as route filters. However, these palliatives become unfeasible in terms of scalability, mainly due to the administrative overhead and cost of maintaining the filters updated. As a result, a significant part of the Internet is defenseless against route leak attacks. In this paper, we define, describe, and examine the different types of route leaks that threaten the security and reliability of the routing system. Our main contributions can be summarized as follows. We develop a rather basic theoretical framework, which, under realistic assumptions, enables a domain to autonomously determine if a particular route advertisement received corresponds to a route leak. We reason the possible occurrence of route leaks in different scenarios, with the aim of formulating requirements for their identification, and hence thereof prevention to improve routing reliability.