Detection of Sophisticated Network Enabled Threats via a Novel Micro-Proxy Architecture

Abstract

With the increasing use of novel exploitation techniques in modern malicious software it can be argued that current intrusion detection and intrusion prevention systems are failing to keep pace. While some intrusion prevention systems have the capability to detect evasion techniques they all fail to detect novel unknown exploitation techniques. Traditional proxy approaches have failed to protect the universe of discourse that a network enabled service can be engaged in as they view all information flows of the same type in a uniform manner. In this paper we propose a micro-proxy architecture that utilizes reverse engineering techniques to identify a valid universe of discourse for a network service. This valid universe of discourse is then applied to validate legitimate transactions to a service. Thus in effect, the micro proxy implements a default deny policy via the analysis of the application level discourse.