Skip to main content

Software & Hardware

The DSRI specifications

Software

We use OKD 4.14, the Origin Community Distribution of Kubernetes that powers RedHat OpenShift, a distribution of the Kubernetes container orchestration tool. Kubernetes takes care of deploying the Docker containers on the cluster of servers, the OKD distribution extends it to improve security, and provide a user-friendly web UI to manage your applications.

DSRI provides a graphical user interface on top of the Kubernetes containers orchestration to easily deploy and manage workspaces and services.

We use RedHat Ceph storage for the distributed storage.

DSRI works best when you work with code, scripts to run, and web applications. Especially if they require an important amount of computing resources. If you work on desktop softwares with graphical user interface, such as Matlab or Spyder, the installation will be much more complex, and usually using your laptop will be more comfortable, stable and reactive than accessing a desktop interface on a remote server through the UM VPN.

Here is a non-exhaustive list of some of the applications that can easily be deployed on the DSRI:

  • Multiple flavors of JupyterLab (scipy, tensorflow, all-spark, and more)

  • JupyterHub with GitHub authentication

  • RStudio, with a complementary Shiny server

  • VisualStudio Code server

  • Tensorflow or PyTorch on Nvidia GPU (with JupyterLab or VisualStudio Code)

  • SQL databases (MariaDB, MySQL, PostgreSQL)

  • NoSQL databases (MongoDB, Redis)

  • Graph databases (GraphDB, Blazegraph, Virtuoso)

  • Apache Flink cluster for streaming applications

  • Apache Spark cluster for distributed computing

  • Or any program installed in a Docker image!

Hardware

  • 16 CPU nodes
RAM (GB)CPU (cores)Storage (TB)
Node capacity512 GB64 cores (128 threads)120 TB
Total capacity8 192 GB1 024 cores1 920 TB
  • 1 GPU node: Nvidia DGX1 8x Tesla V100 - 32GB GPU
GPUsRAM (GB)CPU (cores)
GPU node capacity8512 GB40 cores
DSRI infrastructure

Learn more about DSRI

See the following presentation about the Data Science Research Infrastructure

DSRI April 2021 Community Event Presentation