All posts by Lars Hohmuth

21 Aug 2017

Edge computing: taking digitisation to the edge

Edge Computing beschreibt eine Datenverarbeitung am Rand des IT-Netzwerks

While the whole world talks about the benefits of Cloud Computing, a new trend – edge/fog computing – is reshaping the IT landscape. In contrast to the oft-cited data cloud, fog or edge computing involves decentralised data processing on-site. The term „fog“ emphasises the conceptual closeness to the „clould“ but things happen much closer to the user.  Edge computing basically refers to the same thing; the term describes data processing that takes place at the edge of the IT network. Especially for Industry 4.0 scenarios, the previous limits of IT (the office or the server room) are being extended into production areas. Locally and decentralised at the source of the data, not in air-conditioned data centres.

Decentralized data processing at the edge of the network

There are clear advantages to edge computing: as more and more end devices are networked in the course of the internet of things, huge mountains of data are created which require extremely fast and therefore expensive data connections for their proper transmission. Fog computing, on the other hand, takes care of basic data processing locally and forwards only the results to the central cloud providers. As a result, the amount of data being transferred decreases considerably. This reduced data volume also results in a reduced vulnerability to malicious attacks. It can also lead to significantly lower latency for industrial control systems because the data is processed locally and does not leave the local networks. There are also compliance reasons for using edge computing – storing raw data locally provides improved security and protection for your data.

Suitable devices for edge computing applications


Small versatile computers are required for edge computing; these must be able to work reliably in a harsh environment, handle various protocols and be capable of processing data.


In the traditional production environment, data processing has tended to focus on operating individual machines or facilities. In contrast, the digital transformation underway in industrial enterprises now aims at an all-encompassing networking of all “things” – Industrial Internet of Things (IIoT) is the magic word. Simple IoT gateways usually only communicate with cloud providers without saving or analysing the data. For such applications, traditional industrial PCs are unsuitable because they have excessive performance, size, and costs. Small versatile computers are therefore required; these must be able to work reliably in a harsh environment, handle various protocols and process data. On the MICA, for example, database programs and protocols can be installed and connected like smartphone apps. So data can be collected and processed on-site with speed and flexibility. Since the MICA is designed for protection class IP67, it can transition seamlessly from prototyping to permanent installation on a machine.


The best of both worlds

The synergy results from the good scalability and availability of the cloud, which can be used for capacity bottlenecks in local IT resources or as a fallback solution. It also provides a central storage location for collecting and presenting the results of the processed raw data from multiple locations.

Cloud providers also offer post-processing services, such as machine learning or big data analytics, which are difficult to implement locally. The collected data can then be accessed centrally on the cloud storage. The provides MICA-based solutions for such applications. These help bridge the gap between fog/edge computing and the cloud.

IT decision-makers are increasingly facing the question of whether to process their data in the cloud or in the fog. In the long term, it will be hard to avoid a combination of both.



18 Aug 2017

Cloud computing for small to mid-sized companies

Small and mid-sized companies have long resisted the cloud computing trend. However, this phalanx of the resistance is slowly crumbling. In fact, more and more decision-makers are realising that the use of cloud computing can provide major cost savings. Many requirements from customers and employers cannot be delivered without the support of cloud-based services. Data security remains a continual issue – as it should.

Data in the cloud – gone with the wind?

Cloud computing refers to non-local data processing. The hardware and software are not located locally at the company (referred to as on-premise or private cloud) but rather located in a remote data centre and then made available by a cloud provider via internet. Perhaps it is the term itself which causes some of the decision makers discomfort. Is company data safe in the cloud, or is it “gone with the wind”?

Cloud Computing or On Premise?

The storage, performance and security offered by cloud computing providers is generally more cost-effective than providing the hardware and personnel locally. And this is also a great advantage for SMEs, since they don’t need expensive server rooms and IT specialists. The ability to analyse large amounts of data at high speeds, in particular, often cannot be implemented by the local IT department at a reasonable cost.  Software is also increasingly being used as a service from the cloud (software-as-a-service, or “SaaS”). Thus, expensive programs do not have to be purchased and installed. Cloud computing is also advantageous compared to on-premise installations when several sites are to be networked and supplied with identical software. The same applies when providing operational information on mobile devices. Many such applications are already available from our partners.

Big data for machine learning 

Das Bild zeigt in einem Dashboard zwei Graphen von Klimadaten (Temperatur, Luftfeuchte) über die Zeit.
Combining environmental data (climatic data shown in photo) with order data and machine data


A typical cloud computing scenario in manufacturing companies is context-based information presentation. Employees from the manufacturing or service department automatically receive the information that they currently require – typically on their mobile device (tablet, smartphone or smartwatch). The necessary information must therefore be compiled from real-time and historical data and then analysed. The real-time data is collected using existing or newly installed sensors. Temperature sensors, for example, detect the ambient conditions of machines. Cloud services start learning the machine’s behaviour by analysing long track data, and draw conclusions about possible future events, such as disturbances or risks, and put them in relation to the machine temperatures. As a result, machine wear can be detected early based on temperature deviations. This additional information is communicated clearly and at an early stage via notifications or dashboards so that the necessary repairs or replacements can be scheduled. As a result, production losses or delays are avoided. Significant costs and risks do not spring up unexpectedly.

IoT gateways: data from the sensor to the cloud

Condition Monitoring System
Condition Monitoring System

Devices in the automation pyramid, starting with autonomous sensors and controllers, are not normally designed to transfer data to the cloud. There are so-called IoT gateways which are used for this purpose. The best tools for this job are small industrial PCs, such as HARTING’s MICA, because their performance, connectivity and return on investment (ROI), have been optimized for these tasks. As an IoT gateway, the MICA collects data from sensors or the PLC. It then converts this into an interpretable data format (such as OPC-UA) and transmits the data to the cloud using an installed Cloud Connector. The container architecture of the MICA enables connectors for various cloud providers to be easily installed as apps. Apps for well-known cloud providers (Dimension Data, IBM Bluemix, Microsoft Azure and SAP Hana) have already been implemented or certified.

Data security for cloud computing

Many companies are, understandably, very concerned about data security. This includes both protection against third-party access (hackers) and against the unauthorised use of data by the provider of the cloud infrastructure. This protection usually is provided by VPN (virtual private network) connections between the IoT gateway and the cloud or even the end application. However, these must be easy to operate and compatible with standardised hardware and conventional communication channels.