Author

Heba Saleous

Date of Award

4-2018

Document Type

Thesis

Degree Name

Master of Science in Information Systems (MSIS)

Department

Information Security

First Advisor

Dr. Zouheir Trabelsi

Second Advisor

Dr. Mohammad Mehedy Masud

Third Advisor

Dr. Huwaida Said

Abstract

The internet has grown to a point where people all over the world have become dependent on this convenient communication medium. However, with this dependency, malicious traffic has become a major concern. Consequently, firewalls have become a mandatory part of any network, due to their ability to filter the traffic based on rules that state which packets should be accepted or denied. However, a network administrator must manually configure filter rules, and packets that do not fit any rule may be subject to wrong judgement by the firewall. Neural networks can learn the filter rules in order to decide if packets that do not fit any specific rules should be accepted or denied. The neural network will be trained with existing packet data and their firewall actions, and then tested to determine the amount of correctly classified packets compared to the firewall.

Share

COinS