The development of major paradigm shifts in computer technology occurs in wave cycles, pretty much in 20-year spans. In the 1960s, mainframe computer technology emerged to meet the needs of large corporations and government organizations, providing highly structured machine data processing, especially for scientific and technical applications. In the 1980s, advances in microprocessor technology and graphical user interfaces made client-server solutions popular. More and more computing power migrated to decentralized and local systems. In the 2000s, the explosive growth of the internet led to an awareness of the need to re-center IT services in the cloud. Consumer services such as video streaming, social networks and eCommerce on the one hand and B2B services such as CRM and ERP systems from the cloud on the other became established, as scalable and fail-safe server software and, above all, high bandwidths for data retrieval were now widely available thanks to the expansion of the broadband infrastructure.

In the current 2020s, a new computing paradigm is emerging in the form of edge computing. In particular, the strong growth of mobile devices as well as modern Industry 4.0 use cases and the trend towards autonomous driving are the drivers behind this. The diverse customer needs on both the industrial and end user side increasingly require local, intelligent data processing in order to provide answers in real time. And this is exactly where edge computing comes into play. Key technological drivers for the success of edge computing are faster and more intelligent processors, artificial intelligence algorithms and the new 5G mobile communications standard with its broadband and real-time properties.

The edge computing market

Leading market researchers currently estimate that there are more than 20 billion devices connected to the internet and other networks. But this is just the beginning: by 2025, this figure is expected to rise to more than 34 billion devices; smartphones, tablets and notebooks are not even included in these surveys. Market researcher IDC expects the global digital data stock to grow by 175 zettabytes by 2025. To put this into perspective: one zettabyte is equivalent to one trillion bytes, a number with 21 decimal places. Due to these huge amounts of data, it is obvious that local processing and utilization of the enormous amounts of data must become increasingly important. If the data volumes generated by hundreds of millions of autonomous vehicles were to be pumped into the cloud, the data lines and, above all, the central processing capacities would be hopelessly overwhelmed. Small, 'embedded' computer systems have been controlling devices, machines and systems for many years. Thanks to the continuous development of semiconductor electronics, they are now so powerful that they can also perform complex calculations and data analyses. Such systems are now developing into a key technology of digitalization under the heading of 'edge computing'.

With the Selective Artificial Intelligence Performance Index (ISIN: DE000VL3SJB4) from Vontobel, investors can invest in the 20 most important shares of leading companies in the field of artificial intelligence that are also heavily involved in edge computing projects via an index certificate. The index covers the entire value chain with companies from the hardware and software platforms, big data and application-oriented artificial intelligence sectors.

© Solactive

Analysts at Transparency Market Research estimate that the global market for embedded systems will be worth over USD 233 billion by the end of 2021. The edge computing market in particular is growing exponentially due to digitalization: the analysts at Grand View Research expect annual growth of 54%. This would put the global market volume for edge computing at around USD 28.84 billion in 2025. A report by Strategy Analytics predicts that around 59% of all data in IoT applications will be processed by edge computing in 2025.

The areas of application

Edge computing will be used in practically all industrial and consumer segments in the future. In particular, this includes the areas of Industry 4.0, autonomous driving, financial services, healthcare, agriculture (smart farming), energy suppliers (energy supply & power grids) and retail. In all areas, the aim is to optimize processes, the use of resources and, above all, to improve the user experience. Today's customers expect an ad-hoc response via their smartphone - whether they are placing an order on an eCommerce site or receiving complex online advice on financial services.

One of the pioneers in the area of stationary retail is the eCommerce giant Amazon. In January 2018, Amazon opened its first Amazon Go store to the public. The store allows customers to log in with a QR code on their mobile app, whereupon cameras and sensors in the store identify customers and register what they choose to buy.

The energy transition in particular offers a wide range of opportunities and challenges for edge computing. Not only can malfunctions of decentralized wind and solar power plants be detected in real time, edge computing also saves money: if a wind power plant is maintained using a combination of edge and cloud computing, the associated costs are only a third of those of a pure cloud solution. In intelligent grid control, edge computing systems in conjunction with sensors will be able to monitor everything from the energy stored by electric vehicles to the energy generated by wind turbines in order to make decisions to reduce costs and make energy generation more efficient.

Edge computing in the context of Industry 4.0

Surveys by market researcher Frost&Sullivan assume that 70% of data in the industrial environment is generated at the edge. Covid-19 combined with ad-hoc changes in demand in production and the simultaneous fragility of supply chains require intelligent, agile and breathing production. This does not necessarily have to mean completely new systems and production lines, as the new semiconductor production facilities at Bosch in Dresden and Infineon in Villach have recently shown. Above all, edge computing also means new intelligence for older production facilities.

Thomas Rappold: "Edge computing offers German industry the opportunity to take pole position in the global digital race."

© Thomas Rappold

SmartFactoryKL shows how it's done: the consortium's demonstrator system, for example, has a weighing module that has been retrofitted with an edge device for quality assurance. The scales are used to weigh the items (business card holders) produced on the system in order to check the correct size of the components. As the scales were previously unable to record data and were not OPC UA-capable, the module manufacturer integrated an edge device into the existing system. It reads the various states of the scales, such as temperature, overload, underload and stabilization time, calculates them and summarizes them into a status message that is transmitted to the cloud. For example, unacceptable vibrations caused by a motor near the scale or contamination would affect the stabilization time; this would be detected during the data analysis. Thanks to the edge device, faults in the motor and on the scales can now be identified and rectified promptly.

Artificial intelligence and 5G as key building blocks

With a new generation of AI chips, it is now possible to execute complex deep learning algorithms in embedded systems and deliver the results in real time. In addition to local AI intelligence, the new 5G mobile communications standard provides a connectivity solution for decentralized edge computing. In future, the 5G standard will be able to support up to one million devices within one square kilometer. Other advantages include the flexible allocation of bandwidth and the low latency times of less than 10 ms that are so important for Industry 4.0 and autonomous driving. In some cases, they are even as low as 1 ms.