Uma ferramenta para simulação e monitoramento de comportamentos anômalos em contêineres Docker

Abstract

Container virtualization has become, in recent years, one of the main alternatives in the development and provision of services in the cloud computing environment. The Docker tool became popular in this context and was widely adopted by the whole set of proposed advantages. The ease of configuring, running and maintaining multiple containers on the same machine has added to this technology some challenges in the area of information security. However, any vulnerability present in a container could expose both the container itself and the host machine to unauthorized external access. Thus, analyzing and monitoring Docker containers in order to identify anomalous behavior has become one of the main challenges for companies specializing in cloud computing and their users. This work proposes the development of a tool capable of simulating anomalous behavior in Docker containers based on a previously defined configuration. With this, it is intended that the tool proposed here can help in the simulation of possible anomalous behaviors in a configurable way. In this way, the present work aims to contribute to the academic community as a tool that generates and exposes a set of data that can be used to analysis and creation of machine learning models, capable of learning and identifying anomalous behavior in Docker containers.