Extending Case-Based Reasoning (CBR) Approaches to Semi-automated Network Alert Reporting.

摘要

A substantial amount of cyber security analyst time is spent handling well-known and na ve threats and policy violations on the local network. This includes both the time spent actually identifying and analyzing the activity as well as generating and filing reports associated with the activity. With increasing concern over advanced persistent threats, there is an interest in the development of techniques to automatically handle well-known threats and policy violations. We propose extensions to existing case-based reasoning approaches to support the unique requirements of cybersecurity report generation. Specifically, we consider the fact that we are reporting on hostile actors that will attempt to game the system or manipulate the system to actually aid the actors in obfuscating their activity. In this report, we describe the need for automated reporting, the applicability of case-based reasoning, our proposed extension to the standard case-based reasoning system model, and provide examples of the modified case-based reasoning system as applied to example cybersecurity scenarios.