With over 5 million users, 6 million repositories and billions of image pulls per month, Docker Hub has become the to-go image registry for most container users. However, according to a 2019 study [1], 90% of the images have compressed sizes of up to ~0.5GB, and over 7000 images are consisted of a single layer. Surely such numbers can be justified for some very specific use cases, but in general, it would be naive to assume that most images in Docker Hub are adopting good practices when it comes to building efficient container images.

This workshop tackles the best techniques for optimizing Docker images, with live demonstrations and presentations from real use cases.

Attendees can follow the live demonstrations by themselves, as long as they have access to a computer with Docker installed. Everyone is invited to participate and contribute with ideas and thoughts on how to go the extra mile and further improve the processes and practices being shown.

After the hands-on session, a series of speakers will be given the chance to talk about their own experience with image building. The presentations’ topics can range from success stories to innovative optimization techniques, or even showstopper challenges to get the audience engaged.

[1] N. Zhao et al., “Large-Scale Analysis of the Docker Hub Dataset,” 2019 IEEE International Conference on Cluster Computing (CLUSTER), 2019, pp. 1-10, DOI: 10.1109/CLUSTER.2019.8891000.