Soft computing in network intrusion detection

The explosive growth of computer networking has brought about limitless possibilities and opportunities along with increased risks of malicious intrusions. It is therefore, critical to have the security mechanisms that are able to prevent security violations. However, complete prevention of security breaches does not appear to be practical. Intrusion detection can be regarded as an alternative, or as a compromise to this situation. Several techniques for detecting intrusions have been studied by many researchers. This paper discusses why intrusion detection systems are needed and presents the main techniques that these systems utilize. In any of these techniques, the need for exploiting the tolerance for imprecision and uncertainty to achieve robustness and low solution costs is evident. It can be noted that, this is in fact, the guiding principle of soft computing and more particularly soft computing. This work reports the preliminary steps taken in the study of the implications and advantages of using fuzzy logic for handling intrusion detection through approximate reasoning and approximate matching.