Date of Award
Master of Science in Information Technology (MSIT)
Distributed Denial of Service (DDoS) attacks on web applications has been a persistent threat. Successful attacks can lead to inaccessible service to legitimate users in time and loss of business reputation. Most research effort on DDoS focused on network layer attacks. Existing approaches on application layer DDoS attack mitigation have limitations such as the lack of detection ability for low rate DDoS and not being able to detect attacks targeting resource files. In this work, we propose DDoS attack detection using concepts from information retrieval and machine learning. We include two popular concepts from information retrieval: Term Frequency (TF)-Inverse Document Frequency (IDF) and Latent Semantic Indexing (LSI). We analyzed web server log data generated in a distributed environment. Our evaluation results indicate that while all the approaches can detect various ranges of attacks, information retrieval approaches can identify attacks ongoing in a given session. All the approaches can detect three well known application level DDoS attacks (trivial, intermediate, advanced). Further, these approaches can enable an administrator identifying new pattern of DDoS attacks.