A Porsche Digital Twin: Driven by Data Streaming & NoSQL
There are few inventions that have shaped the modern world quite like the automobile. But what is next for the car and the automotive industry? Sridhar Mamella, Data Engineer at Porsche AG, and Haris Causegic, Software Engineer at Porsche AG, write about the development of a digital twin for Porsche.
Over the last decade, automobiles have become ever more sophisticated and turned into computers on wheels. They are increasingly equipped with on-board sensors, cameras, and computers. When talking about modern vehicles, we are no longer talking solely about electro-mechanical machines. In fact, today, it takes much more than just powerful engines/motors, responsive brakes, or lots of torque to deliver a premium sports car. With the advent of digitalization, everything changed. The future of the automobile and the automotive industry is digital!
In fact, cars have become increasingly software-driven in recent years, as more and more digital technology and features are added to the mix. This is true not only for the vehicles, but also for the automotive industry as a whole. During the entire lifecycle of a car, massive amounts of synchronous and asynchronous data are generated.
To leverage this process, we are embracing the concept of digital twins at Porsche.
What is a digital twin?
Digital twin describes data aggregation along a product lifecycle. This data can be classified, analyzed, and used to predict events within the product lifecycle and recommend appropriate actions. As virtual replicas of physical objects or systems, digital twins outlive the physical lifecycle, so they can be used to sustainably generate value for decisions and features. Digital twins can be used to reduce costs, optimize processes, and develop new features.
Why do we need them? Vehicle data is generated by various sources, and it is used globally by dozens of systems, covering the complete lifecycle of our cars: R&D, production plants, over-the-air (OTA) updates, connected services and many more.
For example, we get Predictive Maintenance Protocols (PMP) from the vehicles or other systems and merge them with the rest of the vehicle data. This allows us to give other systems an overview of the vehicle and the current wear and tear. Thus, errors and possible problems in components can be detected at an early stage before a defect occurs.
Another use case are synthetic vehicles. To test vehicle development, IT systems need to have access to their data. Instead of using real vehicles to the projects for tests, we can create synthetic vehicles (as a digital vehicle), with which the systems can test various scenarios.
Let’s have a look how it all started and how we developed the vision of a digital twin for Porsche.
Stage 0 — On-Premise
At first, we deployed everything on-premises, with a relational database at the core of the system. However, with our first approach, we faced several problems: continuously growing data complexity, there were several data owners that had to be involved, it was difficult to scale infrastructure (no scale-out/horizontal scaling). To resolve this, we knew we had to completely rethink our approach.
We realized that cloud first is the way to go. The whole team is capable of doing DevOps in the cloud. In the long term, OnPremise services are to be eliminated.
Stage 1 — Technical Enablement
With relational databases, we encountered multiple problems:
- Constant schema changes resulted in constraint violations
- Small data model changes required lots of work
- with NoSQL a thing of the past
2. Performance issues
- Problems with scalability & SQL queries being slow
- A high-performance, full-featured text search engine library
3. Increased requirements for rollback
- Changes need to be reversable fast
- Not possible with relational DB, as changes require schema changes
With the transition to NoSQL, complicated schema changes and performance issues are now things of the past. For technical enablement, the following is needed:
1. NoSQL Database: A global cloud Database as a Service
2. Microservices: System/Architectural redesign
3. Cloud Enablement: Development team must have know-how for cloud related development
DBaaS & Microservices are already implemented, but system runs mostly on-premises.
Our vision: Replacing on-premise
With a cloud migration, the foundation for future digitalization is laid. This enables us to add exciting new features and make even better sports cars. When everything is running in the cloud, we can make use of their global infrastructure. With this, we can gain value from vehicle data from all over the world and connect Porsche workshops worldwide.
The digitization of the automotive industry is in full swing. With the development of a digital twin, we at Porsche are taking the next step in our digital transformation journey. It helps us to turn data into profitable Insight and establish a link between virtual and real production.
About this publication: Where innovation meets tradition. There’s more to Porsche than sports cars — we’re tackling new challenges, develop digital products and think digital with a focus on the customer. On our Medium blog, we tell these stories. It’s about our #nextvisions, smart technologies and the people that drive our digital journey. Please follow us on Twitter (Porsche Digital, Next Visions), Instagram (Porsche Digital, Next Visions, Porsche Newsroom) and LinkedIn (Porsche AG, Porsche Digital) for more.