Home // ICIMP 2018, The Thirteenth International Conference on Internet Monitoring and Protection // View article
Authors:
Andreas Schäfer
Prof. Dr. Michael Massoth
Keywords: anomaly detection; internet of things; unsupervised machine learning; intrusion detection and prevention
Abstract:
Internet of Things (IoT) are globally connected devices which are able to collect and exchange information. The increasing usage of IoT-devices in industrial and private environments result in the need for higher security and constant surveillance of such devices. Since 2016 novel botnets, consisting only of IoT-devices, where observed to execute major Distributed Denial of Service (DDoS) attacks. Due to the autonomous nature of these IoT devices, a compromised device might never be detected by system administrators. This creates the need for continuous monitoring of IoT network traffic. A possible solution for this problem is the permanent monitoring of anomalies within the network traffic of the IoT devices. Anomaly Detection Systems (ADS) monitor the behavior of a system and flag significant deviations from the normal activity as anomalies. This paper presents a new three step approach for anomaly detection in unsupervised communication meta data by cascading X-means clustering, decision tree, and statistical analysis, in order to monitor and protect IoT networks.
Pages: 16 to 21
Copyright: Copyright (c) IARIA, 2018
Publication date: July 22, 2018
Published in: conference
ISSN: 2308-3980
ISBN: 978-1-61208-652-1
Location: Barcelona, Spain
Dates: from July 22, 2018 to July 26, 2018