Make the most of the technical possibilities available
There are countless ways of implementing the technical possibilities offered by the IoT and Industry 4.0. Different protocols, different hardware options, frameworks, and platforms – usually, the environment has the biggest influence on the final choice.
A couple of examples
- If Microsoft is already used heavily at your company, the Azure IoT might be an attractive option.
- If open source plays a major part, open source projects will, accordingly, tend to be preferred over and above commercial products.
- If you already use TeamViewer and want to transport data safely into the cloud with little effort, you’ll take a look at solutions that allow you to do just this.
Often, it’s not the technology itself or its manufacturer that is decisive but what you and we make of the technology. Generally, only a fraction of the offered functions are used. But these few functions must then work well!
Successful projects with the help of reference architectures
A look at reference architectures is a necessary part of the concept phase. That’s because the reference architectures provide guiding principles that form the framework for an implementation. This is particularly important if your company still doesn’t really have any guidelines of this nature. Frequently, the question arises as to where a function will be hosted.
An example: You want to use machine learning and a camera to detect whether or not a component has been produced correctly. But where should the neural network (the technology used to realize machine learning) make this check? Directly in the camera? On the camera? On your local computer? Or in the cloud? All of these are valid possibilities – but you need to decide on one.
The clearer your requirements here, the easier it will be later on for you to get an overview of your IT landscape, rectify faults, and make enhancements.
What are the advantages of reference architectures?
Reference architectures are useful for the following reasons:
- They create standards for your IT applications. This makes troubleshooting, general maintenance, and the introduction of enhancements easier. Overall, you save on costs and retain flexibility.
- You gain comparability, and can easily check whether a change (such as switching to another cloud provider) would be a good move. Your IT is structured, which in turn makes it transparent.
- Increasing organization means that you can more quickly survey all of your IT applications and their interactions. New employees can be trained up quickly, and special knowledge – often reserved for experienced employees – is kept to a minimum.
It doesn’t matter whether your company pursues a known reference architecture or develops its own. However, before creating your own, you should take a look at existing reference architectures and evaluate them. Remember that reference architectures usually cover specific functions, tasks, and dimensions. It is rare for them to precisely meet your needs. Instead, they tend to be over-dimensioned, and you’ll require only some of their functions. Knowing this allows you to protect yourself from unnecessary complexity when introducing a new topic. In addition, bear in mind that you might need to extend an existing reference architecture in line with your needs.
Reference Architectural Model Industrie 4.0
The Reference Architectural Model Industrie 4.0 (RAMI 4.0) was created by Zentralverband Elektrotechnik- und Elektronikindustrie e.V. (ZVEI) and Plattform Industrie 4.0. It allows a solution or scenario to be positioned on the basis of its most important coordinates.
To do this, it uses a 3D image with the following axes:
- The “Hierarchy Levels” axis depicts the various functions in a factory.
- The “Life Cycle Value Stream” axis describes the life cycle of plants and products.
- The “Layers” axis depicts the software or machine using a layered architecture.
Several things are of note here:
- The concept of interlinking these three axes and arranging Industry 4.0 applications accordingly is extremely attractive, and enables the visualization of potential that was previously unknown.
- A certain amount of initial training is required in order to understand the complexity of the model and to interpret it. Only then do the benefits become apparent. In our experience, getting to grips with the model and method are worthwhile. Even an examination of the thought processes of the creators of the model is extremely helpful.
Basically, RAMI 4.0 is a reference architecture model that aims to describe Industry 4.0 solutions in general. The flexibility that the model requires is, naturally, reflected in its complexity. When looking at it, find the topics that are particularly relevant for you. Set priorities. Alternatively, the Industrial Internet Reference Architecture (IIRA) is worth a perusal, despite its different focus (Industrial IoT).
The Azure IoT reference architecture, which we describe below, is much more tangible.
Microsoft Azure IoT reference architecture
The Microsoft Azure IoT reference architecture is primarily tailored to Microsoft requirements, but can naturally be transferred to other providers of IoT platforms, too (whereby these also describe reference architectures). We recommend that you examine at least one reference architecture in detail.
In the first abstraction, the reference architecture is divided as follows:
- Things – these are the sensors (IoT devices), edge devices, and (in some cases) the cloud gateways, which can also be added to the IoT platform through bulk device positioning. Note that software is installed on sensors and on other components, and this software needs to be updated from time to time.
- Insights – these are the components that receive, transform, and save data as well as analyzing the data in the traditional way.
- Action – finally come the actions that we want to perform with the IoT (real actions or making decisions, for example).
The functions of the Azure IoT reference architecture
The functions of the individual components in the Microsoft Azure IoT reference architecture are depicted clearly there. The most important ones are described in brief here in the form of an overview:
- The IoT devices are the sensors that can, if you wish, already carry out data analysis themselves (edge devices). The sensors are assigned to an Azure IoT hub individually or by means of the DPS (Device Provisioning Service).
- The cloud gateway is responsible for managing the sensors. It provides secure communication between the sensors and the cloud.
- Data processing is divided into several stages. There are different paths here: The “warm” path is intended for the immediate interpretation of the data, so when you need to react to an event on an ad-hoc basis, for example. The “cold” path is used for the downstream interpretation of the data in order to generate new information from it by means of machine learning, for example. The “stream” path provides traditional methods for the data processing of large datasets.
- Lastly, services are offered for the visualization and implementation of business logic (UI reporting, business integration services).
In comparison with RAMI 4.0, the Microsoft Azure IoT reference architecture is significantly more tangible with regard to logical components. In addition, the fact that data can be processed ad-hoc (warm) and downstream (cold) is beneficial. The focus is not on Industry 4.0 applications here; instead, the processing of data and the use of the Azure IoT is in the foreground. However, the breaking down of an automated factory into its various layers and/or associated product life cycles is not described and must be solved on a case-by-case basis if required.
Simplified definition and structure of an IoT reference architecture
Let’s take a closer look at the structural composition of an IoT reference architecture:
- Sensors & actuators – this means sensors that record data and pass it on to the next level: Temperature data, telemetry data, humidity, pressure etc. In addition, actuators can be addressed here.
- Data transport – this describes the connection of the sensors and actuators with the next level up. Different protocols play a part here, such as MQTT (Message Queuing Telemetry Transport).
- Data evaluation – this is where the gold hidden in the data is mined. The possibilities are diverse: You’ll hear terms such as data science, data analytics, big data, stream processing, complex event processing, machine learning, artificial intelligence, and more.
- Application, visualization, dashboards – this is the level where the results of the data evaluation are used: The service technician is told if a machine requires a maintenance run in order to avoid production downtime. A single glance at the dashboard is enough to get an overview of the current state of production.
- Security – this aspect is important at all levels and for all functions. Security can have an effect on the design of the hardware (e.g. the use of TPM 2.0 chips) or on the software or its configuration.
So a reference architecture can be extremely simple. However, some explanation is required:
- The individual blocks can be seen as layers, but in fact they depict individual functions.
- The (cloud or IoT) gateways that securely forward the data – but can also evaluate it – should be assigned to the tasks of data evaluation or data transport.
- Data can be saved at all levels.
Get started with your IoT project now!
This breakdown can be made even clearer. But let’s not allow ourselves to get distracted from your IoT project – let’s simply get started! Plenty of things will disclose themselves to you during the course of the project. For specific project ideas, examples, and application scenarios, see our full portfolio. Here you will find a detailed overview of the various IoT technologies:
The most important aspects Data transport
Selecting the right IoT protocol Data evaluation
The important role of data evaluation
Our strengths when it comes to IoT technologies
- We are characterized by the fact that we understand specific technologies (e.g. specific frameworks or programming languages, protocols, and so on) as tools for implementing ideas and specifications.
- We understand the concepts behind these tools and can therefore use them in a targeted manner.
- Taking customer requirements and individual circumstances into account forms part of our daily routine.