Detecting BGP Routing Anomalies Using Machine Learning: A Review

The Border Gateway Protocol (BGP) is a protocol for exchanging IP prefixes online. It allows for incremental path-vector routing protocol and reachability between Autonomous Systems (ASes), enabling efficient global internet activity. BGP anomalies have been known to cause ASes to malfunction in various ways; therefore, detecting them is critical. Machine Learning (ML) solutions currently guarantee improved BGP irregularity recognition based on BGP update messages volume and manner features, which are frequently boisterous and bursty. This work organizes these anomalies and provides the most recent approaches for detecting discrepancies. We also look into a few crucial requirements for the up-and-coming detection of internet routing abnormalities methods.