Container Platform (Singularity/Apptainer)
Container usage in HPC environments: Singularity/Apptainer, Podman and Docker integration. Reproducible computing and software distribution.
In HPC environments where multiple research groups share the same cluster and each has different software dependencies, a Singularity/Apptainer-based container platform is an indispensable solution. By packaging an application, its dependencies and its runtime environment into a single portable `.sif` file, it achieves both reproducibility and portability. Mevasis delivers end-to-end container platform services — from installation and central image repository configuration to SLURM integration — with experienced HPC engineers.
Containers also eliminate the conflict between a stable shared system and evolving user software, so each research group can version its own environment without touching the cluster's base installation. This speeds up onboarding and cuts administrator workload.
Key Features
Secure, Rootless Operation
Apptainer runs with user-level privileges, so users never require elevated system access and cluster security is preserved. Because no privileged daemon runs on compute nodes, the attack surface stays minimal. This meets the strict security requirements of shared multi-tenant HPC environments.
SLURM and MPI Compatibility
Apptainer containers run natively inside SLURM job scripts, so users launch them exactly like any other application. Multi-node MPI workloads are supported through the standard MPI implementations already on the cluster. This keeps the container platform transparent to existing workflows.
Docker Image Import
Thanks to OCI compatibility, existing Docker images are converted to Apptainer format with a single command. This lets your team reuse the images they already maintain instead of rebuilding containers from scratch. Converting is fast and preserves the original layers and entry points.
Central Image Repository
A shared .sif image library prevents disk waste by storing each image once for all users to reference. Version control and metadata management guarantee that every researcher runs the exact same container build. This improves reproducibility and simplifies software administration.
Frequently Asked Questions
If different research groups or teams share the same HPC cluster and each has different software dependencies, a container platform is indispensable. It should be chosen particularly for multi-tenant setups where it is not possible to change a Python environment or library version system-wide; for scientific workloads where the results produced need to be reproducible in the future; and for scenarios where you want to move an application tested on your own data center to the cloud or to another HPC system.
Mevasis provides end-to-end container platform services — from Apptainer/Singularity installation and configuration, Docker/OCI image conversion, and central image repository setup through to SLURM integration. Experienced HPC engineers analyze your existing infrastructure, determine the most appropriate deployment architecture and make the transition seamless by delivering hands-on training to user teams.
Because the scope of the container platform service varies depending on factors such as existing infrastructure size, number of users, custom image repository requirements and training needs, pricing is project-specific. Share your requirements via the request form; our team will provide a detailed proposal as soon as possible.
Let's Build Future Together.
