The objective of the "FabOS" project is to develop an open, distributed, real-time-capable and secure operating system for production that will be the IT backbone for the adaptable automation of the factory of the future and the foundation of an ecosystem for data-driven services and AI applications. Hybrid cloud platforms and IIoT applications are core elements of cyber-physical architectures and will form the basis of future production solutions. FabOS forms a platform that provides components modelled on an operating system that link machines, infrastructure and services, just as an operating system does with user programs and hardware in the form of an abstraction layer. A reduction of IT costs and encapsulation of IT complexity to increase usability leads to a reduction of automation costs. FabOS supports the changeability of systems and infrastructure in production from the sensor to the machine to the complete factory without system boundaries. This includes a uniform lifecycle management of all IT resources, production equipment, technical building equipment and infrastructure as well as the creation of a consistent infrastructure for real-time domain-spanning value creation networks for the AI-supported autonomous production of the future.

For this, a flexible and extensible architecture for future requirements of cognitive services, real-time applications and data marketplaces will be created, which enables end-to-end solutions from the cloud via the edge to the shop floor into the machine for real-time applications. This is the prerequisite for the deployment of real-time capable, intelligent control loops at any level: field, factory, company, value-added network.

In order to provide a foundation for IT solutions for SMEs without a vendor lock-in effect, open source components are developed for easy adaptation by solution providers and approaches from community and crowdsourcing are used for innovative solutions. Thus, FabOS can act as an enabler for production digitalization and new business models for German manufacturing companies and solution providers of AI-based services. Last but not least, security by design, taking into account the indivisibility of security, protection and privacy of the data, is an aspect that is considered from the outset and permeates the entire system.

The motivation to create a FabOS is multi-layered. IT architectures for production technology are still far behind "standard" IT. The very heterogeneous IT landscape in production is currently difficult to master with an integrated IT system. This makes it difficult for companies to efficiently access the data required for ML and AI applications. Today, because of this lack of access to data and information from production, there is usually a hard separation between the knowledge domain of production and data analysts. This limits the combination of IT flexibility with the predictability and reliability of industrial facilities and future AI services. In addition, there is a risk of dependence on individual providers of production IT, IT infrastructure, control software, analytics and AI services, which must be prevented in the future just as much as the migration of engineering and production know-how to platforms dominated by oligopolies.

Furthermore, it will be possible to develop new concepts for security precautions through a real-time-capable distributed operating system. Today, a robot or a machine is analyzed on its own and secured with the help of various sensors that communicate directly with the robot. In the future, it would be possible with FabOS to quickly and easily connect safety sensors that make their information available to all robots in the vicinity. By sharing sensor resources in this way, robots can be secured together and costs for sensors can be saved. A typical example is a factory building with several robots through which a person moves. Until now, each robot has detected the movement of a person with its own sensors in order to react accordingly. With the help of FabOS, the position of the person in the building can be determined by fewer sensors overall and communicated to all robots in a timely manner by means of a suitable communication infrastructure, so that a navigation AI reschedules the routes.

Current solutions are usually too specific, difficult to migrate and only cover partial areas. FabOS creates a comprehensive solution for production that generalizes partial solutions and offers added value over the sum of the individual solutions. The adaptability, resilience and usability of such a system have been identified as the primary success factors.

Adaptability is created through the use of virtualization technology modelled on SDx approaches (software-definedverything - cf. VLAN, VM/Container) to make the functionality of the controller and machine modules and the provision of services more flexible. The adaptation of real-time edge-cloud technologies enables a distributed infrastructure for real-time value networks and data-driven services for cognitive applications and AI applications. Consistent solutions from the integration of brownfield systems using connector adapters to the connection of new intelligent systems, services and infrastructure with native interfaces are considered.

Resilience is created through the consideration of security in all its forms (functional security, security of use, data/information security, operational security, privacy), from the outset and through consideration "by design", not just afterwards. In particular approaches for AI-supported resilience, both for FabOS basic services and for application services are regarded here, by combining data and domain knowledge.

Usability is created by increasing applicability and usefulness in order to also allow companies that have only a small IT department or no IT department at all. Here, target group-specific configuration profiles based on suitable deployment scenarios and the maturity level of the user company should support adaptation and application. Deployment and reconfiguration is supported and guided by an intuitive toolchain, which must be designed and implemented together with partners.