Analyzing Technical Lag in Docker Images.
Ahmed Zerouali · ORBi UMONS · 2018
Packaging software into containers is becoming a common practice when deploying services in cloud and other environments.Docker is currently one of the most popular container technologies for building and deploying containers.A key part of this technology is the concept of registry, where container images are stored and shared.The largest one of these registries is Docker Hub, with hundreds of thousands of public Docker images ready for deployment.A container image includes usually a collection of software packages, in many cases corresponding to a Linux-based distribution.By analyzing packages in images, we can learn to which extent those images offer packages that have newer versions available.To measure how "outdated" a container image is, we introduce the concept of "technical lag" for the container, as the difference between a given image, and the most updated image that is possible with the same collection of packages.Using this concept as the framework for the analysis, we empirically measure and study technical lag for 2,245 Docker Hub official images based on the Alpine Linux distribution, with a total of 63,581 packages.Our results show a strong presence of technical lag inside of those images, indicating that even well-maintained containers could benefit of better procedures for updating.