Memory efficacious pattern matching intrusion detection system

In today´s web-enabled world, the communications happening over the network is intensifying at a capacious rate. Not all communications are credible and malpractice can arise anywhere, anytime. If the normal traffic is slightly modified to delude the intrusion detection system, then the traditional systems might not be able to discern the same effectively. Thus, a system that could detect and ferret out the novel attacks has been proposed. Since any number of users can use a web page, maintaining the availability of the resources and allocating them to the active users as per their need is very essential. The multi-thread concept is used to share the resources that each client can use. Attribute Selection Algorithm is used as the feature extraction algorithm in weka, to yield those relevant features pertaining to the user´s request and helps in achieving a more accurate result. Memory efficiency is brought in with the cascading binary search tree. The patterns are efficiently stored and hence the search for the presence of an attack is accomplished effectively. An Intrusion Detection System which is memory efficient and effective enough in detecting attacks and reducing the false positives is thus proposed.