Fundamentals of statistical signal processing: estimation theory
Fundamentals of statistical signal processing: estimation theory
The anatomy of a large-scale hypertextual Web search engine
WWW7 Proceedings of the seventh international conference on World Wide Web 7
Collaborating against common enemies
IMC '05 Proceedings of the 5th ACM SIGCOMM conference on Internet Measurement
Optimal worm-scanning method using vulnerable-host distributions
International Journal of Security and Networks
Highly predictive blacklisting
SS'08 Proceedings of the 17th conference on Security symposium
Hi-index | 0.00 |
We recently established a new experimental Internet service for creating customized source address blacklists for DShield.org contributors. This new service utilizes a radically different approach to blacklist formulation that we refer to as Highly Predictive Blacklists (HPB). A highly predictive blacklist is a list of malicious Internet addresses that is formulated through an analysis of the roughly 30 million firewall log entries that are contributed to the DShield repository each day from across the Internet. The HPB service employs a link analysis algorithm similar to the Google PageRank scheme to cross-compare contributors in search of overlaps among the attackers they report. The attacker addresses included within an HPB are selected by favoring the inclusion of those addresses that have been encountered by contributors who share degrees of overlap with the target HPB owner. Our experiments show that highly predictive blacklist entries consistently yield filters that are exercised at higher rates than those from conventional blacklist methods. In addition, this increase in blacklist filter "hit rates" can last multiple days into the future. In this paper, we provide an overview of our algorithm and present our usage experiences. We discuss the envisioned benefits that we believe HPBs can provide toward reducing unwanted communications for those networks that utilize this service.