Efficient Techniques for Network Attack Identification

摘要

The ubiquity of the Internet has set off an explosion in the number of networked applications developed for the consumption of users. Unfortunately, many are maliciously designed to harm network infrastructure and systems. Considerable effort has been focused on detecting attacks by such applications that are known as malware, and predicting the breaches in security rendered by them. However, large networks pose a formidable challenge to representing and analyzing such attacks using scalable models. Furthermore, the availability of only partial information on the vulnerabilities admitted by a large fraction of the networked systems makes forecasting the future targets of such applications equally hard. In this paper, we present innovative methods to identify and characterize network attacks, and to forecast their propagation over time.