Simplify Your Computations With Containers
Using containers, you can run tools such as FSL and Freesurfer in a container player, avoiding all the headaches of having to install and configure the tools yourself. Containers also make it easier to share and replicate your image processing pipelines and guarantee reproducibility of the results.
Docker and Singularity are two of the most popular container players. Docker has an enormous community of support and provides a wide range of features such as GitHub integration and container orchestration. Singularity targets the use of containers in HPC environments and more native integration with GNU/Linux environments. Both Docker and Singularity have public container image repositories, called hubs, that facilitate the finding and sharing of container images.
NITRC-CE LITE was recently released and is a variation of the NITRC-CE Classic environment. NITRC-CE LITE’s focus is on running analysis using containers and comes with Docker and Singularity pre-installed plus a handful of complementary tools such as DataLad. DataLad provides a way to manage and track your experiment code as well as the data and containers it uses, consumes and produces. NITRC-CE LITE is available for install for free in the Amazon Cloud or by downloading a NITRC-CE LITE virtual machine and running it on your local computer using VirtualBox.
- Matt Travers (NITRC Software Developer), and Yaroslav O. Halchenko (Research Associate Professor, Dartmouth College)