Jienyan Fan:
"Pattern Recognition Method on Detecting DDoS Attacks for
Large-Scale Internet."

Abstract:
In recent years, distributed denial-of-service (DDoS) attacks have
become a major security threat to Internet services. How to detect and
defend against DDoS attacks is currently a hot topic in both industry
and academia. In this talk, we propose a novel method to detect
low-traffic DDoS attacks at edge routers of autonomous systems. The key
idea is to exploit spatial correlation of DDoS attack traffic.
Specifically, we regard edge routers as an image, whose pixels have two
states, normal and attack. So the DDoS attack detection problem is
modeled as a pattern recognition problem. That is, given observations of
all pixels, we recognize which pixels are in attack state. Our
simulation results show that the proposed framework can detect DDoS
attacks even if the volume of attack traffic on each link is
extremely small. Especially, for the same false alarm probability, our
scheme has a detection probability of 0.97, while the scheme using
threshold has a detection probability of only 0.17, which demonstrates
the superior performance of our scheme. Our scheme is also very time
efficient. Our simulation results show that 16 network traces, each of
which has approximately 78M packets, are processed in 2.7 seconds,
excluding the time to extract features from packets.