Towards Coordinated, Network-Wide Traffic Monitoring for Early Detection of DDoS Flooding Attacks

摘要

DDoS flooding attacks are one of the biggest concerns for security professionals and they are typically explicit attempts to disrupt legitimate users' access to services. Developing a comprehensive defense mechanism against such attacks requires a comprehensive understanding of the problem and the techniques that have been used thus far in preventing, detecting, and responding to various such attacks.ududIn this thesis, we dig into the problem of DDoS flooding attacks from four directions: (1) We study the origin of these attacks, their variations, and various existing defense mechanisms against them. Our literature review gives insight into a list of key required features for the next generation of DDoS flooding defense mechanisms. The most important requirement on this list is to see more distributed DDoS flooding defense mechanisms in near future, (2) In such systems, the success in detecting DDoS flooding attacks earlier and in a distributed fashion is highly dependent on the quality and quantity of the traffic flows that are covered by the employed traffic monitoring mechanisms. This motivates us to study and understand the challenges of existing traffic monitoring mechanisms, (3) We propose a novel distributed, coordinated, network-wide traffic monitoring (DiCoTraM) approach that addresses the key challenges of current traffic monitoring mechanisms. DiCoTraM enhances flow coverage to enable effective, early detection of DDoS flooding attacks. We compare and evaluate the performance of DiCoTraM with various other traffic monitoring mechanisms in terms of their total flow coverage and DDoS flooding attack flow coverage, and (4) We evaluate the effectiveness of DiCoTraM with cSamp, an existing traffic monitoring mechanism that outperforms most of other traffic monitoring mechanisms, with regards to supporting early detection of DDoS flooding attacks (i.e., at the intermediate network) by employing two existing DDoS flooding detection mechanisms over them. We then compare the effectiveness of DiCoTraM with that of cSamp by comparing the detection rates and false positive rates achieved when the selected detection mechanisms are employed over DiCoTraM and cSamp. The results show that DiCoTraM outperforms other traffic monitoring mechanisms in terms of DDoS flooding attack flow coverage.