Requirements

To exploit the proposed contents of this course, it is recommend to know or have experience in the following topics:

• Use of SSH
• SLURM Basic knowledge
• Use of containers/dockers

Schedule

Total hours: 9 hours.

• Class 1: Monday 13/3, 14:00hs to 17:00hs
• Class 2: Tuesday 14/3, 14:00hs to 17:00hs
• Class 3: Friday 17/3, 14:00hs to 17:00hs

Short Program

Objetives

The use of high-performance computing resources by the scientific, educational, and industrial communities is growing day by day. This promotes the inclusion of a wide range of users, and the opportunity for the creation of complex workflows in HPC. At the same time, it requires a different approach regarding the traditional access methods to these resources.

In this course, we will present FirecREST, an alternative method in which computational resources can be accessed through a programmable web interface (RESTful API) in a way that HPC resources can be exposed in a more friendly fashion to the different communities, and thus allows managing complex workflows through secure and standard interfaces

The program of this course covers the installation of the API, its configuration, and connectivity to an HPC cluster with working examples of use for automation and graphic user interface.

At the end of this course, students should be able to deploy this service in their HPC Center.

Topics:

• FirecREST, RESTful HPC (theory)

  • Motivation, why using RESTful APIs in HPC
  • Use case: procedure automatization in HPC
  • Use case: web application to handle HPC workloads

• FirecREST Infrastructure (theory)

  • Authentication layers and authorization
  • SSH for command execution
  • Storage
  • API Gateway
  • Firewall

• FirecREST installation in an HPC center (practical)

  • Hands-on: VM configuration configuración for FirecREST installation

• Development of clients for FirecREST (practical)

  • Hands-on: application development form managing HPC jobs and data transfer