Intrusion Confinement by Isolation in Information Systems

System protection mechanisms such as access controls can be fooled by authorized but malicious users, masqueraders, and misfeasors. Intrusion detection techniques are therefore used to supplement them. However, damage could have occurred before an intrusion is detected. In many computing systems the requirement for a high degree of soundness of intrusion reporting can yield poor performance in detecting intrusions, and can cause long detection latency. As a result, serious damage can be caused either because many intrusions are never detected or because the average detection latency is too long. The process of bounding the damage caused by intrusions during the process of intrusion detection is referred to as intrusion confinement. We justify the necessity for intrusion confinement during detection by a probabilistic analysis model, and propose a general solution to achieve intrusion confinement. The crux of the solution is to isolate likely suspicious actions before a definite determination of intrusion is reported. We also present a concrete isolation protocol in the file system context to evaluate the feasibility of the general solution, which can be applied in many types of information systems.