Title :
Multiple Markov Models for Detecting Internet Anomalies from BGP Data
Author :
Gardiner, Judith D.
Author_Institution :
Ohio Supercomput. Center, Columbus, OH, USA
Abstract :
This project explores a new mechanism for early detection of Internet disturbances, including both natural and malicious events. We used multiple hidden Markov models to analyze a type of global routing data called Border Gateway Protocol (BGP). Reasonably good discrimination was achieved between quiet periods and disturbances, and some discrimination was achieved between natural and malicious events. This project was exploratory in nature; no validation has been done on the results.
Keywords :
Internet; Markov processes; computer network security; routing protocols; Internet anomaly detection; Markov models; border gateway protocol; Grippers; Hidden Markov models; IEEE news; Internet; Markov processes; Routing protocols; Training;
Conference_Titel :
DoD High Performance Computing Modernization Program Users Group Conference (HPCMP-UGC), 2009
Conference_Location :
San Diego, CA
Print_ISBN :
978-1-4244-5768-7
Electronic_ISBN :
978-1-4244-5769-4
DOI :
10.1109/HPCMP-UGC.2009.61
