Detection of Denial of Service (DoS) Attacks in Local Area Networks Based on Outgoing Packets

• Main Author: Manaa, Mehdi Ebady
• Format: Thesis
• Language: English; English
• Published: 2012
• Subjects: TK5101-6720 Telecommunication
• Online Access: https://etd.uum.edu.my/2922/1/Mehdi_Ebady_Manaa.pdf; https://etd.uum.edu.my/2922/3/Mehdi_Ebady_Manaa.pdf; https://etd.uum.edu.my/2922/

Denial of Service (DoS) is a security threat which compromises the confidentiality of information stored in Local Area Networks (LANs) due to unauthorized access by spoofed IP addresses. DoS is harmful to LANs as the flooding of packets may delay other users from accessing the server and in severe cases, the server may need to be shut down, wasting valuable resources, especially in critical real-time services such as in e-commerce and the medical field. The objective of this project is to propose a new DoS detection system to protect organizations from unauthenticated access to important information which may jeopardize the confidentiality, privacy and integrity of information in Local Area Networks. The new DoS detection system monitors the traffic flow of packets and filters the packets based on their IP addresses to determine whether they are genuine requests for network services or DoS attacks. Results obtained demonstrate that the detection accuracy of the new DoS detection system was in good agreement with the detection accuracy from the network protocol analyzer, Wireshark. For high-rate DoS attacks, the accuracy was 100% whereas for low-rate DoS attacks, the accuracy was 67%.