Bernhard Günthner, Managing Director of Kontron Technologies and Vice President IoT Solutions in the Kontron - S&T Group since October 2019, is well acquainted with application-specific system solutions in the fields of industrial automation and medical. The electrical engineering graduate started out as a software developer for telematics and ticketing systems. He has now been with Kontron for over ten years and has played an active role in shaping the IoT revolution.

The Industrial Internet of Things is generating more and more data that needs to be managed. This data complexity needs to be mastered. What role does edge computing play in this?

Günthner: The integration of control, visualization, security, connectivity and the relocation of complex AI tasks such as machine/deep learning, e.g. for machine vision, from the cloud to the edge means that intelligent edge computing plays a central role. In most application areas, the storage, processing and analysis of data must also meet certain requirements in terms of security and determinism, meaning that this is not usually done in a cloud. This is where the non-time-critical big data evaluations and medium/long-term data storage take place.

At the edge, sensor data is preselected, software applications are processed or even AI calculations are carried out. What should users bear in mind with regard to the edge software landscape? Which technologies should/must be taken into account?

Günthner: Tools for efficient device management enable, among other things, centralized management of all devices in a cloud-based, scaling environment for edge devices, as well as the IoT connectivity of Docker containers, in which encapsulation of customer applications is possible; also management of device fleets/groups and configuration of container groups per device group. Cyber security is also becoming increasingly important: with the help of tools such as 'SecureOS', it is possible to carry out continuous penetration tests and security updates with a hardened operating system for X86 or ARM; in addition, there are secure update mechanisms and container technology for applications.

Kontron is a leading global provider of IoT/Embedded Computer Technology (ECT).

© Kontron

With the EquipmentCloud, we offer a ready-made and scalable solution that supports aspects such as digital transformation and AI, as well as the creation of digital business models. To integrate the solutions into the existing infrastructure, it should be possible to transfer data from production to big data applications via REST, MQTT or AMQP, as well as to connect machines via OPC UA, SOAP, FTP and TCP/IP. Kontron believes that a holistic approach, such as with our SUSiEtec IoT toolset, is the key to success.

As the tasks grow, so do the hardware requirements. What should users look out for when selecting the right edge computing hardware?

Günthner: High-performance computing (HPC) is becoming increasingly important due to growing requirements and ever-increasing data volumes. When making a selection, attention should therefore be paid to scalable platforms and solutions so that performance can be adapted in each case, including AI functionality. In addition, the respective environment with temperature range and cooling options for heat dissipation must be taken into account when making the selection, resulting in a maximum permissible power dissipation (TDP) for safe, continuous operation. Edge servers are also increasingly required for deterministic data exchange between multiple edge clients. Many of these requirements can already be met today with systems based on COM Express Type 7.

However, the development continues: COM-HPC could become one of the leading standards for HPC - both for client and server applications that do not have to be installed in the IT server room, but on site, i.e. at the edge, sometimes in harsh environments. COM-HPC therefore makes it possible to run even compute- and data-intensive AI processes for time-critical applications on site. Time-Sensitive Network (TSN) functionality for manufacturer-independent, deterministic Ethernet communication from the field to the controller to the IT level is now already included in the new Intel Atom and Intel Core i processors, as well as in certain processors from NXP, so that no additional costs are incurred for the connection.