The Big vision: From shopfloor edge computing to cloud analytics
The industry wide awareness of the potential of data, often driven by the latest trends like artificial intelligence or machine learning, has been growing rapidly over the last years. This raised Brose’s wish to utilize data in new ways to generate more customer value and profit as well as to strengthen a competitive edge. Brose is the fourth largest family-owned automotive supplier and equips every second new car worldwide with at least one of its products.
Brose’s big vision is to connect a multitude of legacy data sources, foremost shopfloor production machinery, to a central platform that can host data uses cases that accelerate the company’s business and generate value. The use cases range from providing analytical dashboarding for monitoring to hosting AI & machine learning applications to cut costs and raise quality. The vision is split into two projects, which Ultra Tendency delivered in separate workstreams.
The first project, Industrial IoT (IIOT), aims to provide a fast and high-throughput connectivity of shopfloor machinery data to a central data platform in the cloud.
The second project, Data Driven Brose (DDB), has the goal of providing the central analytics platform where the data is stored, high performance ETL takes place and use case applications are hosted.
The vision of Brose is an end-to-end connected platform. Both workstreams must work in close collaboration to create a robust and futureproof architecture from shopfloor devices to cloud platforms.
The Drivers – Industrial IoT
Flexibility and enhanced time to market for new applications are the main driver for the introduction of a global IIoT platform for Brose. The platform supports automated rollouts, easy software changes without the need to change machine programming and connectivity towards machines, internal systems, and cloud platforms. An important factor for the success of the IIoT platform is the ability to deploy and operate the platform globally across all Brose factories.
The Solution - Industrial IoT
Ultra Tendency conducted an initial pilot phase that provided the input for an architecture decision. Based on a set of requirements, the Ultra Tendency team implemented three different architecture prototypes in close collaboration with the Brose team. The phase was concluded by an assessment of the different options based on an upfront created decision matrix leveraging Brose-specific functional and non-functional requirements as well as international standards. Applying this decision matrix to the three implemented prototypes resulted in a clear architecture recommendation.
The actual project implementation of the Brose global IIoT platform was done using the Eclipse IoT stack and Confluent, based on Apache Kafka. Confluent acts as the central communication hub to connect the IoT applications with other Brose systems and the Data-Driven-Brose infrastructure deployed in AWS. The actual IoT applications are implemented using Eclipse Kura and the global device management is based on Eclipse Kapua. This approach based on open-source tools allows a short implementation phase, no vendor lock in and a good extensibility for future use cases.
During the implementation, the Ultra Tendency team collaborates closely with the Brose team to integrate the new IoT layer with the manufacturing execution system. A specific requirement for the use cases in the Brose IIoT platform are high availability and scalability of use-cases and a very low latency (<30 msec) for the interaction between the Kura-based IoT devices and machines since the IoT devices are not just capturing data but also controlling the production process on the machine.
With the successful rollout of the IIoT platform to the first factories, Brose and Ultra Tendency created the capability to quickly rollout additional use cases. The current use cases can be extended by simple Java coding and automatically rolled out to the Kura devices across Brose factories worldwide. Ultra Tendency supports the global rollout throughout 2022.
The Drivers - Data-Driven Brose Platform (DDB)
In order for Brose to become a data-driven company, a central Data Lake platform is needed, in which large amounts of data can be stored and that can host different use cases.
Solution Data-Driven Brose
Ultra Tendency developed in cooperation with Brose the concept for the data lake, planned and implemented it. The first step consisted of performing a Proof of Concept (PoC) to determine a suitable cloud provider and to facilitate the decision of an architectural design of the solution. From here the Ultra Tendency DevOps-Engineering team took over to translate and roll-out the architecture into a running system in the cloud. This included supporting Brose in operating the platform from the start to the go live of the first use cases. It took only six months from the first PoC to achieve this. Ultra Tendency used its extensive Big Data & Cloud experience to provide cost efficient solutions that integrate seamlessly with Broses corporate culture, governance and processes by following a DevOps approach.
A CI/CD process was used roll-out and release the platform via Infrastructure as Code (IaC). This approach enables fast and agile development of the platform. Together with using Terraform in an efficient manner this acted as a lighthouse inside Brose’s organization. It radiated towards other departments and projects that subsequently starting using IaC with Terraform successfully. Suitable staging environments for Use Cases and applications as well as a suitable multi account management were specific governance requirements. These were satisfied by Ultra Tendency’s solutions and implemented to Brose satisfaction. Ultra Tendency proposed concepts for general Security concerns like encryption of data at rest and the establishment of a secure communication between on premises and the cloud. The concepts were implemented and resulted in a platform that also satisfies the highest standards of security.
Ultra Tendency also consulted Brose and teamed up closely during all development stages to allow knowledge transfer and to facilitate enablement. Brose wished to build up the expertise necessary to develop and operate parts of the big data platform completely by itself.
By the end of the first year the Data-Driven-Brose platform successfully hosted several use cases. The first use case serves up to 180 users from different Brose factories globally. Near real time images of welding seams are classified by cutting edge computer vision machine learning algorithms. It became a decisive factor in cost reduction, increasing production speed and product quality with homogenous results throughout all factories.
Another use case enables the storage of production line machinery data from shopfloors on the platform. The data is transformed and loaded to produce dashboards that send information back to the shopfloor workers who can optimize their workflow and thereby reduce costs and increase productivity.
For this year Brose and Ultra Tendency plan to expand the platform by building an Machine Learning Operations (MLOps) environment as well as self-service capabilities. These capabilities provide sandbox environments and support “data as product” approaches, working towards a Data Mesh setup. Also planned is a continuous extension of governance and data management implementations, e.g. data zoning and -catalogues to optimally react towards organizational and processual changes inside Brose. Finally, the partners will establish a Kubernetes-based container environment to operate micro service applications.
The new IIoT platform already forms the basis for live diagnostic tools for maintenance, IT, and operators. The next major step is to open the platform to equipment suppliers outside of Brose to partially automate the delivery approval and to use this platform to significantly reduce the software complexity of new equipment.
Brose is the fourth largest family-owned automotive supplier. Every second new car worldwide is equipped with at least one Brose product. The company's intelligent solutions for vehicle access and interior ensure more comfort and flexibility. Innovative concepts for thermal management increase efficiency and contribute to environmental and climate protection. Brose's understanding of systems enables new functions in vehicles of all kinds - whether on four or two wheels. Including the Brose Sitech joint venture, the company employs around 30,000 people at 70 locations in 25 countries. For 2022, the Brose Group expects sales of more than 7 billion euros.