BMW uses Ansible to manage HPC cluster

BMW Group has migrated its high-performance computing (HPC) cluster comprising 15,000 Windows and Linux servers to the Red Hat Ansible Automation Platform to improve manageability and IT automation. The Red Hat platform also enables self-service, reducing the IT admin time needed to deploy applications for users.

The HPC infrastructure is used in performance-intensive applications such as 3D modelling, crash test simulations and new car design. 

For the past 30 years, the HPC environment has been managed by a custom application known as scVenus, which needed replacing as it neared end-of-life.

The German carmaker, which owns brands including Mini, BMW and Rolls-Royce, assessed the Red Hat Ansible Automation Platform and worked with the Red Hat technical account management team, along with partners including IBM and Puzzle ITC AG, on the migration of applications, servers and client software.

Replacing a 30-year-old legacy system across 15,000 servers while maintaining continuous vehicle production is a monumental challenge

Discussing the migration, Martina Pretzl, project leader for HPC at BMW Group, said: “Replacing a 30-year-old legacy system across 15,000 servers while maintaining continuous vehicle production is a monumental challenge. By collaborating closely with Red Hat and our partners, we operated as a single, unified team to deliver this highly complex migration on time and within budget, establishing a modern, agile automation foundation.”

According to Red Hat, moving to the Ansible Automation Platform has enabled BMW Group to use Ansible Playbooks. This, it said, has enabled complex infrastructure tasks that previously required up to 60 hours of processing time on thousands of nodes to be reduced to just 20 minutes. Other routine management processes have been cut by 75%, from 12 hours to three hours.

BMW Group has also introduced “shift left” self-service capabilities through a unified self-service portal to enable users to manage their own software licence restarts and support requests.

The car manufacturer has also deployed infrastructure as code for resilience. This enables redundancy and automated disaster recovery across thousands of endpoints and multiple availability zones.

Looking ahead, the IT team at BMW Group plans to introduce event-driven automation to resolve routine IT service issues automatically. BMW Group is also transitioning towards a more cloud-native, containerised architecture, and plans to use AI coding assistants alongside the Red Hat Ansible Automation Platform’s automated policy as code, enabling faster development while maintaining strict compliance.

Johan Barnard, team lead for Ansible Core at BMW Group, said: “Red Hat Ansible Automation will be with us for the foreseeable future. AI can’t be trusted directly to run your production systems – you need automation as that trust layer.”

Earlier this year, BMW Group progressed its cloud-first strategy with NTT Data’s managed services for its global network, across 40,000 network devices, 260,000 ports and 30,000 servers.

In 2022, it began working with HPE to use the company’s GreenLake edge-to-cloud IT infrastructure to provide cloud services for big data, backup, recovery and compliant archiving for the development of electric vehicles. The HPE technology has enabled BMW Group to manage distributed data via a single platform, with a globally consistent cloud experience.

It is also a reference customer for Amazon Web Services, and has been collaborating with the public cloud service provider to develop commercial off-the-shelf cloud technology for securely managing vehicle data.