Established clouds can do a lot. However, when it comes to real-time control of machines and real-time evaluations for fast decision-making processes on the machine, cloud computing often reaches its limits. This is because evaluating, analyzing and transferring more and more data to the cloud and returning it to its source takes too long for applications in which fractions of a second are crucial for machine control.

Thanks to the scalability of computing power from the edge to the cloud, the tasks of individual processes can be adapted precisely to the circumstances. In the first step, high-performance servers are used to train tasks that are not yet time-critical. In the real process, the trained model can then be applied immediately on demand, close to the process and in the edge.

The 'Intelligent Edge': Takes over many computing tasks upstream and maintains storage capacity.

© Kontron

Instead of connecting sensors and machines to the cloud and processing the raw data in the cloud, as much computing and storage capacity as possible will be shifted to the intelligent edge in future. This means that data will be filtered, processed and analyzed close to its point of origin, before it is transferred to the network, using trained models in order to keep the control loop fast and efficient directly at the process.

New technologies, such as the 'containerization' of complex functions or the creation of 'digital twins', are essential prerequisites for achieving this scalability.

This is what makes software-defined machines possible in the first place and even allows powerful tools such as edge analytics, artificial intelligence (AI) or machine learning to be outsourced to the edge. Trained models in combination with powerful computer systems then allow complex, time-critical processes to be moved from the cloud to the edge computer without the need for a permanent connection to the cloud.

The Industrial Servers are used where a private cloud is required on-premise. They are then connected to the edge devices - or to so-called fog computers for higher computing power - and store the collected data in the local cloud. They also perform higher-level control tasks or data processing for the public cloud. Less time-critical data can still be outsourced to a public cloud in order to provide management information from distributed systems.

Together with its sister company S&T Technologies, Kontron offers a range of hardware and software products and services for such cases. An important component of the package is the Susietec software framework developed by S&T Technologies. Susietec is a range of software and services that serves as a link between the components in edge, fog and cloud architectures. Susietec also has the ability to integrate machine learning in order to be able to suggest its own decisions based on available data using artificial intelligence methods.

As part of a partnership with Iconics, data can not only be recorded and analyzed in the future in combination with the Supervisory Control and Data Acquisition Software (SCADA) from Iconics, which is optionally integrated into Susietec, but also visualized via the cloud, for example Microsoft Azure.

With the support of open standards such as OPC UA and TSN (Time Sensitive Networking), interface problems can be resolved in the medium term and data can be mapped virtually. Bandwidths are also increased, which accommodates the real-time applications and security aspects of industrial IoT. The conventional boundaries between IT (information technology) and OT (operational technology) are disappearing thanks to the harmonized standards and the use of Susietec.

Susietec in combination with Microsoft Azure

Susietec is integrated with Microsoft's Azure IoT edge services, enabling seamless scalability of computing power between computing resources at the intelligent edge, in the embedded cloud, in the on-premise data center or in the public cloud. In technical terms, these are containers in which software functions are mapped and which can then be moved between the cloud and edge devices. In this way, it is possible to decide flexibly - depending on the desired level of security and performance - where data processing and analysis should take place. The use of Azure IoT Edge therefore significantly speeds up reaction times in processes.

TSN and OPC UA are the future

The IoT software framework Susietec is the link between end devices, edge devices, control computers and clouds.

© Kontron

The current standards TSN (Time Sensitive Networking) and OPC UA (Open Platform Communications Unified Architecture) form the backbone of the embedded cloud. TSN plus OPC UA enables deterministic OT and standardized IT data transmission on the same Ethernet. This is essential for time-critical machine control systems.

On the standards side, Kontron is fully committed to OPC UA and TSN. For example, Kontron is a participant in TSN testbeds as part of the Industrial Internet Consortium (IIC) and the Lab Network Industry 4.0 (LNI) and enables an uncomplicated entry into Time Sensitive Networking with the TSN starter kit.

In combination with the new LS1028 systems-on-chip from NXP, it is even possible to integrate up to five TSN-capable Gigabit Ethernet ports into the chip. With an additional four serial interfaces and a CAN port, not only can costs, space requirements and power loss be reduced, but compact IoT gateways and edge computers can also be realized.

New technologies and business models

The latest technologies for hardware, software and connectivity are constantly expanding the range of tasks for edge computers. In future, machines will be implemented in the software with a full range of functions - similar to cars. However, depending on the model and licensing form, the currently usable range of functions will only be activated via the software. New business models, such as usage-based leasing models like pay-per-use, also require powerful, connected computers, as do new service models in connection with preventive or predictive maintenance.

For many applications, hybrid cloud architectures consisting of their own on-premise architecture with edge and fog computers and the use of public clouds are the ideal combination. Capture data, process it intelligently, optimize processes and efficiency and provide higher-level information platforms: All of this is possible with scalable computer platforms, TSN, OPC UA and IoT edge container software technology, all fully integrated by Susietec.

The edge computing world of Kontron

For edge and fog computing architectures, Kontron offers robust embedded computing solutions based on the SMARC 2.0 (Smart Mobility Architecture) and COM Express standards, among others. One example of a cost-effective solution with guaranteed latency and quality of service for controlling Industry 4.0 architectures is the Kontron SMARC-sAL28 module.

Kontron boards, modules and embedded systems also support the latest Intel Core and Server Class Intel Xeon processors. They offer increased power density as well as flexible, scalable computing power and networking, making them ideal for intelligent edge controllers, edge servers, machine learning and AI applications.

Industrial computer platforms can be used to implement servers that take over on-premise cloud and storage functionalities as well as AI tasks. High modularity and flexible storage options, paired with the latest Intel Core and dual Intel XEON processor platforms and optional GPGPUs (General Purpose Computation on Graphics Processing Unit) allow a wide range of applications. Like the new KISS family, the compact ZINC-CUBE-SKD embedded server is designed for compute-intensive applications such as machine learning or AI.

It is important for edge computers to be able to connect different sensors or PLC devices in order to obtain information from them for modern applications for data analysis and process optimization. Thanks to the latest Intel Atom or Intel Core technology in conjunction with TPM 2.0, the KBox-A and KBox-C series ensure that performance, security and the wide range of applications in the industrial sector are guaranteed. Both computer families are fanless, flexibly scalable and support Microsoft IoT Edge. The KBox-C series also allows the use of PCIe expansion cards, for example to equip systems with TSN Ethernet interfaces.

Security for IoT applications

To ensure robust security in IoT applications, the Kontron product family of boards, gateways, modules and systems features TPM 2.0 components and CodeMeter technology from Wibu-Systems. This means that trusted hardware can be assumed and standard technologies - such as Secure Boot - can be implemented with secure operating systems. They are 'IoT-ready', i.e. secure against tampering, reverse engineering and attempts to copy your own intellectual property.