Evaluating Docker for Lightweight Virtualization of Distributed and Time-Sensitive Applications in Industrial Automation
Michael Sollfrank, Frieder Loch, Steef Denteneer, Birgit Vogel‐Heuser · IEEE Transactions on Industrial Informatics · 2020
A trend, accompanying the change of automation systems and their architectures, is the virtualization of software components. Virtualization strengthens platform-independent development and the provision of secure and isolated applications. Virtualization introduces well-defined interfaces to strengthen modularity, which facilitates the scalability of applications. However, virtualization includes additional software components and layers and, thus, additional computing costs. This additional effort can conflict with the real-time requirements of automation processes. Current research lacks the investigation of the time behavior of container-based virtualizations concerning their use in real-time systems. An assessment concerning real-time applications is required to prepare it for use in industrial automation. This article examines the effects of virtualization on the time delays of a software component based on Docker containers by providing measurements on a hardware testbed in a realistic use case. The experiments indicate that Docker virtualization can meet soft real-time requirements and can be used in industrial automation.