Reproducible & Scalable Research Computing with Containers

Scientific software is often dependent on other packages and libraries (and specific versions of these). The dependencies are hard to reproduce upon redeployment by collaborators and users who may not share the same system setup. This creates an unnecessary barrier to sharing many excellent software packages making them less shareable and reproducible.

Containers are an effective solution to this problem. The course teaches the fundamentals of creating, deploying and managing containers. You will learn about the essential commands and scripts, about Docker Hub and Singularity (HPC-friendly solution). The workshop will be delivered through a combination of slides, demonstrations and hands-on practice. 

Part 1: Introducing containers, Docker command lines 

In this part, you will learn why containers are used, what containers are, and you will get hands-on practice with Docker (the most popular container tool) command lines.

Part 2: Create containers with examples 

In this part, you will get familiar with creating container images in Docker environment. You will learn to compose Dockerfile and build container images through examples including bash, Python and R. You will get hands-on practice as well. 

Part 3: Share container images and scale up with singularity. 

In this part, you will practice how to share your own container images with others using either Docker Hub distribution or other ways. You will also learn how to run images using singularity on high performance computing cluster or cloud.