Detection of Insider Attacks Against Interior Routing Protocols

Abstract

Internal routing protocols are critical to the operation of private networks worldwide. Of these internal routing protocols Open Shortest Path First routing protocol is the most widely used. Unfortunately, OSPF is vulnerable to insider attacks. Such attacks are reasonably straightforward to carry out by insiders but, because they are based on protocol weaknesses, very difficult to defend against. Solutions in the past have usually relied on monitoring routing tables in all routers across the network and looking for inconsistencies. This approach is cumbersome, computationally expensive and likely to result in many false positives. In this paper we present a novel technique for detecting such attacks using Recurrence Quantification Analysis (RQA). RQA is a non linear data analysis technique suitable for monitoring the behaviour of dynamic systems such as network routing. We implemented on our network testbed some of the well known but difficult t o defend against attacks. We then captured traffic t hat was generated as a result of the attacks and calculated RQA metrics based on it. Using these metrics we were able to demonstrate that RQA can be used to detect attacks quickly and with minimal computation.