A Detection and Offense Mechanism to Defend Against Application Layer DDoS Attacks

Application layer DDoS attacks, which are legitimate in packets and protocols, gradually become a pressing problem for commerce, politics and military. We build an attack model and characterize layer-7 attacks into three classes: session flooding attacks, request flooding attacks and asymmetric attacks. We proposed a mechanism named as DOW (defense and offense wall), which defends against layer-7 attacks using combination of detection technology and currency technology. An anomaly dete-ction method based on K-means clustering is introduced to detect and filter request flooding attacks and asymmetric attacks. To defend against session-flooding attacks, we propose an encouragement model that uses client´s session rate as currency. Detection model drops suspicious sessions, while currency model encourages more legitimate sessions. By collaboration of these two models, normal clients could gain higher service rate and lower delay of response time.