With the steady advancements in the technology, the network security is really important these days to protect information from attackers. In this research, the main focus is on designing strong firewall filtering rules so that detection of malicious code is achieved to an optimal level. A proposed framework is introduced to improve the performance parameters such as Server response time, Web content analysis, Bandwidth, and the performance of the Network traffic load. This research work defines a new set of IPtable rules achieved by modifying the kernel source code. This is done using OpenBSD kernel source code, which results in the formation of a mini-firewall. Therefore, a new hybrid approach is proposed by adding packet filtering rules and SNORT technology in mini-firewall for malicious activity detection. It is an efficient and practical technique which will be helpful to mitigate the malware attacks and secure LAMP server. Experimental analysis has been done to conclude that around 70-75% malicious activity can be reduced by using the proposed technique.
An Empirical Comparison on Malicious Activity Detection Using Different Neural Network-Based Models