Digital platforms have the potential to completely change entire industries as well as products. Despite this, automation providers are still very cautious when it comes to IoT platform integration. At the moment, users only have the option of IoT retrofitting. However, connecting devices to an IoT platform is generally no longer rocket science. The functional scope of leading cloud and IoT platforms has been developing at a rapid pace for several years. Device and machine manufacturers who continue to refrain from using these functions for their products will lose competitiveness in inverse proportion to this trend.

In this respect, anyone in the automation industry - who markets assemblies, devices or machines in which a piece of software is used - is well advised to take a closer look at the technical details and application possibilities of the IoT platforms from Amazon, Microsoft, Siemens and others and to bring their products up to date with the latest digital technology.

Building blocks of an IoT platform

IoT platforms consist of different layers, which vary depending on the provider. Despite sometimes having completely different names, at least two functional layers can be identified in most platforms:

IoT platforms have a multi-layered structure. The devices are connected as IoT devices in a star configuration via a hub. The application logic, which consists of standard services and individually created applications, is used to extract information from the device data. In addition, there is a rules engine that is run through according to the data flow principle of incoming and outgoing data and can trigger programmable actions.

© SSV

To connect the individual devices, sensors and actuators (IoT devices), there is usually an IoT hub or device gateway as a central message hub, which handles all communication between the platform and IoT devices and in which the IT security of the transmission channel is also implemented.

- All other components of a platform are combined in the application logic. This includes the numerous standard services of the respective platform (various databases, various runtime environments for apps, data analysis services with AI algorithms, notification functions) as well as a rules engine (data flow engine) as the data processing logic for incoming and outgoing IoT device data. Software modules (applications) are also being developed for the services and the rules engine in order to link the individual functions of the standard services with each other on a task-related basis.

Requirements for the IoT platform

In addition to the application logic, AWS IoT consists of a modular IoT hub that enables MQTT, HTTP 1.1 and WebSockets connections on the protocol side, with a clear focus on MQTT. In terms of security, the devices are expected to support TLS 1.2 with mutual authentication and are equipped with an X.509 certificate plus a private key. All security functions can be set very finely via a management system.

© SSV

Thanks to the IoT hub, the entire application has a star-shaped topology with the individual devices as end points. Such a hub forms the distribution function between all devices of an application and the services of the IoT platform. In this respect, the most important requirements can be divided into three topics:

1. security

As the central communication endpoint for all IoT devices in an application, the IoT hub also includes IT security to secure all connections. This includes

- Device authentication: every device that is allowed to communicate via the IoT hub must be securely identified beforehand;

- Ensuring confidential - i.e. encrypted - data transmission;

- Ensuring data integrity.

In addition, a security management function is required that enables the management of device-based authentication on the one hand and ensures complete control over all device access to the IoT platform on the other.

2. communication diversity

An IoT hub must support multiple communication patterns. In practice, not only the classic 'device-to-platform-application' communication direction is required. 'Platform-application-to-device' and 'device-to-device' are also required in many applications. Of course, the different security requirements must be taken into account. Data transfers from the platform to a device (i.e. IoT device write access) are definitely much riskier than communication in the opposite direction.

3. scalability

In extreme cases, an IoT hub must also function securely and reliably with millions of simultaneously connected devices that generate tens of millions of events per second. The challenge here is to achieve the most constant possible time behavior for the data throughput, regardless of how many active devices are currently connected to a hub and the number of messages this generates.

AWS IoT - a practical example

AWS IoT (AWS IoT Core) is offered as a cloud-based IoT platform. It enables any devices to communicate or interact with AWS cloud applications via an internet connection secured by TLS 1.2. Strictly speaking, AWS IoT is an additional functional layer that has been added to Amazon WebServices, which has been available since 2006, in order to adapt AWS to the special requirements of the Internet of Things.

From the perspective of the AWS IoT architects, IoT applications collect data in the cloud in order to process it with special applications in the environment of the AWS standard functions. Application scenarios are also conceivable in which an AWS application communicates directly with IoT devices, for example to control them remotely via Alexa Voice Service (AVS). In this respect, the platform includes a very comprehensive security system with diverse rights management.

A modular solution is used as the IoT hub. The interface to the devices is a so-called device gateway, to which all IoT devices of an application are connected via TLS. Due to the special features of the TLS protocol, MQTT and WebSockets make it possible to establish relatively long-lasting connections so that a new - time-consuming and data-intensive - TLS handshake is not required every time. A publish/subscribe broker function (message broker) is available for receiving and sending messages via MQTT. It also supports message transmission via HTTP POST request by means of a special server function. Such an HTTP request is functionally treated as an MQTT publish.

Brief overview of the most important Amazon Web Services (AWS) that can be used by an IoT application. Each individual service is a relatively complex application, but its practical use sometimes requires extensive specialist knowledge:

Special feature Device Shadow

With regard to the performance of Device Gateway and Message Broker, Amazon points out that these functional units scale fully automatically as managed services and are also suitable for very large projects with more than one billion IoT devices. In this context, Amazon emphasizes that the Message Broker can also handle millions of simultaneous MQTT connections and, for example, distribute an incoming publish message to millions of subscribers.

Klaus-Dieter Walter, CEO at SSV Software Systems

© Market&Technology

A special feature of the AWS IoT Hub is the device shadow. This function block serves as a virtual data image (virtual representation) for each device connected to the device gateway, in which the last device status is stored for up to one year. This virtual representation is formed by a JSON-based document (maximum 8 KBytes per device). The JSON data can be accessed by other AWS applications or external mobile devices using a special REST API, for example. The device shadow is synchronized with the respective device from time to time. New (status) data is written from the device to the JSON document (Reported State), but variables are also read and transferred to the device (Desired State), which can lead to a changed device status.
The AWS IoT application logic essentially consists of a special rules engine and the standard web services of the Amazon Cloud (see table on page 3).

The rules engine can be used to pre-process, store and analyze all data received from the IoT devices and pass it on to other services. To do this, sets of rules are created with a syntax similar to SQL. If a rule applies, the corresponding action is executed.

Amazon provides various software development kits (SDKs) for different programming languages and runtime environments to support developers. One of these is the AWS IoT Device SDKs. They support the connection of different device architectures to the IoT Hub. There are also a large number of application SDKs for developing your own applications for the AWS cloud in all common languages.

The external appearance and internal technical implementation of almost all technical products has changed significantly in recent decades (see table on the next page with the example of a fictitious pillar drill). This trend will continue in ever shorter innovation cycles and will include business models as well as technical features.

System network requires platform integration

For the current state of the art (level 5 in Table 2, the pillar drill as a product-service hybrid), platform integration of the respective products is required in any case in order to realize the system network envisaged by Industry 4.0 concepts. Some IoT platform functions can run directly on site, i.e. as an edge application, while others require an external cloud service.

A condition monitoring solution that, for example, provides early warnings of impending failure of individual components in a 'system of systems' in the event of unusual operating conditions can be implemented both locally at the edge and in the cloud with the help of an IoT platform. Suitable sensors (such as current, vibration and noise, as each machine has specific current, vibration and noise characteristics) are used to continuously record status data and transmit it to the IoT platform. There, machine learning algorithms or other statistical methods are used to determine the respective component stress and calculate the remaining useful life.

The result can be used to optimize maintenance and quantifiable cost savings can be achieved through productivity improvements.

No innovative product without an IoT platform: overview of the individual development stages of technical products in recent decades using the example of an electric pillar drill. A higher-level IoT platform is already required for the overall function in the system network. Although stage 5 represents the current state of the art (the product-service hybrid), platform integration is still usually carried out via IoT retrofitting: