Details

Autor(en) / Beteiligte

• Chang-Jung Hsieh
• Ting-Yuan Chan

Titel

Detection DDoS attacks based on neural-network using Apache Spark

Ist Teil von

• 2016 International Conference on Applied System Innovation (ICASI), 2016, p.1-4

Ort / Verlag

IEEE

Erscheinungsjahr

2016

Link zum Volltext

Quelle

IEEE Electronic Library (IEL)

Beschreibungen/Notizen

• Network security issues are becoming serious with the growth of Internet, in many types of network attacks, The Distributed Denial of Service (DDoS) has become the vital threat, The Akamai report reveals first quarter of 2015 there are more than doubled attacks in the same quarter last year, In addition, the types of DDoS attacks changing frequency, in the past, attackers used huge volumes of traffic in short time to make victim host unavailable, but now some attackers used low volumes of traffic for a long time making attacks difficult to detect. Nowadays, there already have many methods for the DDoS detection, but no one can fully detect, the difficulty lies in DDoS attacks often have huge volumes of traffic, in first quarter of 2015, there were 8 attacks exceeding 100 Gbps, and their characteristic highly variable, make hard to detect, in order to overcome it, this paper propose DDoS detection method based on Neural Networks, implemented in the Apache Spark cluster, we used 2000 DARPA LLDOS 1.0 dataset to train and perform experiments to our detection system in a real network environment, the results show our detection system able to detect attacks in real time and average detection rates were over 94%.