Cloud computing is considered a central element of the smart factory. Huge amounts of data can be transferred and analyzed, and storage capacities and computing power can be scaled flexibly and practically at will. The cloud reduces complexity for many application scenarios. However, the limitations of the technology quickly became apparent. The cloud has its limits, particularly when processing data from real-time systems, where large volumes of data need to be evaluated very reliably in a short space of time.

This limitation is particularly critical for applications that digitize and actively control production processes. Software that coordinates when material is to be delivered to the machine or analyzes whether a component is defective during the process must function reliably. Centralized control from the cloud entails the risk of the software no longer working due to a slow or interrupted internet connection. The more software is used in the factory, the more important it is that the central server is reliably accessible. In addition, the complexity increases disproportionately the more logic and data run via the central server. This makes extensions and adaptations complex and system maintenance difficult. Central systems generally have this characteristic, whether cloud or on-premise.

Overcoming these boundaries requires a technical paradigm shift: decentralized edge computing. There is no central brain, instead the system is made up of many independent, decentralized entities that can make decisions independently and exchange information and coordinate directly with each other. Although the decentralized entities can continue to exchange information with central brains, for example to determine priorities, they do not require a constant connection in order to act. This significantly reduces the complexity of the central brain, and decisions are made quickly and reliably in a decentralized manner.

Transferred to the IT world, this means that data is stored locally and logic is executed locally on edge computers such as industrial PCs, scanners, gateways or tablets. Communication takes place directly between the edge computers in the local network. Central message brokers or databases are no longer necessary. Edge computing is not yet widespread in the industrial sector, but it offers an attractive solution for factories in particular. The approach promises higher availability and high flexibility. There is no infrastructure overhead from central instances and additional devices can be integrated into the network. The software is installed directly on them.

Inexpensive hardware

Now is a good time to get started with decentralized edge computing, as a number of developments in recent years are now enabling its widespread use. One of these developments is the massive reduction in the cost of powerful hardware. As the software in edge computing is executed locally on end devices and data is also stored there, powerful devices are required. Industrial-grade tablets and gateways have become significantly cheaper and can now be purchased for low three-digit sums.

Another development is a trend of the modern Internet: decentralized computing. Here, computers are connected to form a network, data is exchanged directly "peer to peer" and calculations are carried out exclusively in a decentralized manner. The failure of one computer does not lead to the failure of the entire system. Decentralized systems are highly fail-safe and highly scalable. Blockchains use this architecture to make decentralized decisions. Open source technologies from the blockchain world now make it easier to set up decentralized systems.

New software platform

Actyx has developed the ActyxOS software platform based on this architecture so that as many companies as possible can use Industry 4.0 applications. The aim is to make developers more productive in the digitalization of processes between machines, robots and people. The platform provides the infrastructure for a factory to work digitally and take advantage of the benefits of edge computing. It automatically handles the distribution and storage of data between edge computers and offers integrated developer tools for the rapid implementation of apps on the platform. As the application examples for the system in the production process are diverse, Actyx cooperates with IT experts from a wide range of sectors.

The vision for the future is a marketplace for the manufacturing industry where plant operators download apps from the ActyxOS platform. Thanks to a cooperation with the Italian software manufacturer Alleantia, machines can already be integrated into the process via an app. In the future, further apps for tracking products, worker assistance on mobile devices and cloud integration via app installation will be possible. The apps can also be used to transfer data to cloud systems. The cloud is the better option, especially for complex, less time-critical data analyses or long-term data storage. In order to exploit the full potential of Industry 4.0 applications, both worlds need to be brought together. MaxFischer, co-founder of Actyx