Real time hybrid intrusion detection system using signature matching algorithm and fuzzy-GA

In Internet-based communication, different types of networks are used to provide services to users. Due to exploration of different types of vulnerabilities, day by day threat of attacks in network is increasing. There is possibility of getting system infected by internal or external intruders. Single computer system is shared by multiple users. Users use multiuser system by creating their own account, which is protected by unique id and password. Sometimes, authorized users attack system for some malicious purposes. Internal attacks include deleting some important files or data, altering some important data. For any intrusion detection system, it is very difficult to identify authorized internal intruders. SQL injection is one such attack, which can be launched by internal attacker. The purpose of external intruder is to perform malicious activities in remote system. To avoid these two types of attacks, the robust intrusion detection system is needed. In this paper, we have implemented hybrid intrusion detection system, which includes identification of both internal and external attacks. Signature matching algorithm is implemented to identify internal attacks. Fuzzy genetic algorithm is applied for implementation of external attacks detection. The system is hybrid and compatible in offline as well as online environment. Experimental results show that the accuracy of system is better than some of existing systems.