Publication Date

Spring 2018

Degree Type

Master's Project

Degree Name

Master of Science (MS)


Computer Science


In this paper, we present the results of using Hidden Markov Models for learning the behavior of Docker containers. This is for use in anomaly-detection based intrusion detection system. Containers provide isolation between the host system and the containerized environment by efficiently packaging applications along with their dependencies. This way, containers become a portable software environment for applications to run and scale. Unlike virtual machines, containers share the same kernel as the host operating system. This is leveraged to monitor the system calls of the container from the host system for anomaly detection. Thus, the monitoring system is not required to have any knowledge about the container nature, neither does the host system or the container being monitored need to be modified.