Authors : Asfiya Shireen Shaikh Mukhtar and Ghousiya Farheen Shaikh Mukhtar2
Page Nos : 64-66
Description :
A firewall is a technology that connects a network to one or more external networks by acting as the network's interface. It is responsible for implementing the network's security policy by deciding which packets should be permitted to travel across the network based on criteria set by the network administrator. Any error in the formulation of the rules may result in the security of the system being compromised, as unwanted traffic may be allowed to pass through while appropriate traffic is prevented from passing through. An anomaly in policy may result from manual rule formation because it produces a collection of regulations that conflicts with itself, redundant with itself, or overshadowed with itself, which is a result of the manual defining of rules. Manual identification and resolution of these anomalies is necessary, but it is a time-consuming and error-prone task that must be done by hand. Previous research on abnormalities in firewall policy has mostly focused on the analysis and identification of these anomalies, with little attention paid to the causes of these anomalies. Previous works describe the potential relationships between rules, as well as the anomalies that may occur as a result of the relationships, and they provide methods for identifying the anomalies through the analysis of the rules in question.
In this research, we present a method for identifying the anomalies through the analysis of the rules in
question separately by Convolution Neural Network and Recurrent Neural Network.
