Theoretical bounds on control-plane self-monitoring in routing protocols
Proceedings of the 2007 ACM SIGMETRICS international conference on Measurement and modeling of computer systems
| Hi-index | 0.00 |
A growing body of knowledge about misconfigurations and their effects on modern network routing protocols has been accumulated by network researchers and practitioners. The focus in identifying these anomalies and their effects has been on the particular and on short-term practicalities. To our knowledge, nobody has asked the broader and harder question: given a protocol P, what misconfigurations can be detected and how? Are there classes of misconfigurations that cannot be detected, and can these somehow be identified.We present preliminary results in our attempt to address this issue. We first classify detection methods into what we call weak detection and strong detection. A weak detection method is one that, given a node's state, checks specific properties that should hold when the protocol is being correctly implemented. It may however fail to identify a detectable misconfiguration because it did not verify the appropriate property. Strong detection methods must, in contrast, detect any misconfiguration that is detectable by any property.Our work investigates such strong detection techniques for different routing protocols. We present preliminary results and demonstrate an O(|V|3) algorithm that implements strong detection on the Distance Vector protocol.