Executive interview: How Audi provides developer self-service

Audi, part of the VW Group, owns the premium brands Audi, Ducati and Lamborghini. It has been focused on creating a secure, stable central environment for innovative development at scale across the VW Group. As part of this project, Audi IT needed to reduce risks and remove dependencies with a flexible, modular architecture that could support iterative work.

The company chose Red Hat OpenShift as the base platform for Kubika-O, its self-service environment for software development. This environment offers the software developers at Audi a way to design, deploy and scale new applications and project environments rapidly. 

Kubika-O currently supports several key applications that are central to the VW Group’s business operations, such as in the areas of data collection, analysis and automation.

Sebastian Kister is product team lead in the Kubernetes Competence Center at Audi. Describing the role of IT within the car company, Kister says: “Audi IT sees the standardisation of platform projects throughout the VW Group, which has 660,000-plus employees, as a central task – with the provision of robust runtime environments in the cloud.”

To support this goal, Kister says Audi IT has built up a professional service organisation that supersedes the times of rigid “plan-build-run” schemes. In the past, organising IT around plan, build and run structured the IT function of an organisation around major areas of infrastructure that worked across technology domains.

Such an approach was previously seen as a way to organise the IT infrastructure to reduce cost, improve performance, and help prepare companies for next-generation IT infrastructure products.

In 2013, McKinsey noted that the plan-build-run organisational structure for IT offered global IT functions a better way to manage IT compared with organising IT infrastructure functions around “technology towers” for mainframes, servers, storage, middleware, databases, networks and end-users.

But Kister says Audi IT needed the ability to support any adaptations and extensions required at short notice.

Previously, projects became backlogged because of time-consuming provisioning processes that could require lead times as long as six months. He says: “Only with a flexible system can Audi meet dynamically evolving business requirements.”

Like many large businesses, Audi set several criteria when selecting a suitable platform and deciding whether to deploy workloads on a public cloud, private cloud or on-premise. Kister says innovative development projects always require a secure, stable and scalable environment – but it is equally important that IT can provide the business with a flexible, modular architecture that reduces risks, eliminates dependencies and supports iterative work.

“The runtime environment also needed to offer multicloud compatibility, as VW and Audi use several cloud platforms,” he says. This multicloud approach recognises the fact that the group has a global market presence. “Audi also wanted to implement a cloud-agnostic platform to avoid lock-in,” he adds.

Asked about how Audi develops software and the way teams are organised, Kister says: “As you’d expect from a global brand like Audi, we have a wide range of demands from across all our various teams. We use an agile management approach, enabling the delivery of customer-centric solutions that we can rapidly get to market, and we release many service version updates on the go.”

Architecturally speaking, Kister says Audi has decided to use containerisation to decouple operating systems from workloads. Why use containers rather than more traditional virtual machines? “The benefits of having lightweight workloads is definitely the most important argument for containerisation,” he says. “Our project teams develop and run our customers’ containers, but we focus on working with them to allow them to transition with us and empower them to work alongside us.”

With automation through Red Hat OpenShift, Audi IT is able to provide tailored Kubernetes clusters and add-ons for each Kubika-O project. As a result, experienced developers and those new to Kubernetes can work more efficiently to create, deliver and migrate innovative solutions across on-premise and cloud environments.

According to Kister, the modular, responsive container infrastructure provided by Red Hat OpenShift contains all the necessary functions and services to operate a container management platform for business-critical applications on different infrastructures in a certified manner. The platform enables Audi IT to manage service-level agreements, multiple security layers, automation and cluster management. “With Red Hat OpenShift as a basis, users can, for example, move containerised applications between clouds while retaining all functionality,” he says.

The company has deployed Red Hat OpenShift on Amazon Web Services (AWS) and Microsoft Azure as runtime environments, supported natively on AWS and Azure. Kister says each version is jointly managed, supported and monitored by Red Hat and the respective cloud provider. “The Audi team takes a cloud-agnostic approach, which leverages complementary solutions to the runtime environment, especially with regard to security,” he adds.

For Kister, these complementary capabilities address IT management requirements such as container image scanning, container runtime scanning, vulnerability checks, monitoring, role-based access controls or continuous integration and continuous delivery and deployment, among others.

Overall, Kister sees the role of IT changing. “You can’t expect a system to run as-is for 10 years any more,” he says. “To survive in an environment of constant change, we needed a multi-tenant system with the ability to make small adjustments and updates whenever necessary.”

This is where developing self-service through Audi’s Kubika-O environment fits in. Such environments are key to digitisation efforts that require continuous innovation through the development of new software-powered products and services.