The merging of physical products and software in the manufacturing industry, for example, has progressed exponentially thanks to the development of increasingly powerful microprocessors. In addition, the triumphant advance of the cloud is now also enabling industrial companies to make enormous progress, especially if they want to bring intelligent systems to market that generate large amounts of data; the extremely flexible and highly scalable cloud solutions offered by Microsoft, Amazon and Google often serve as the basis for this.

While the lifecycle of the product itself is relatively long and static, the trend in the software industry is increasingly moving towards dynamic development processes that incorporate lean and kanban concepts familiar from manufacturing. Successful companies such as Netflix or Facebook are now characterized by the fact that they continuously renew and improve their software, sometimes with up to 100 deployments per day. Cloud technologies will develop in a similar way.

It is therefore important to combine traditional, rather slower industrial processes with the dynamic processes of the software industry in the future. Companies must therefore reconcile the different pace of innovation cycles.

Synchronization of two worlds

Cloud technology plays a key role here. A cloud-based solution with Microsoft's Azure, for example, is further developed with a high frequency of several updates and releases per year. The question therefore arises as to how it will be possible to coordinate the life and development cycles of the two areas of product and cloud solution. This challenge should not be underestimated, especially as software for embedded microprocessors used in intelligent systems is generally written in programming languages that are very close to hardware. At the same time, however, the way in which these systems must be able to communicate with a dynamic cloud is constantly changing.

This challenge arises in all phases of the product life cycle: in the design phase, interfaces and data flows must be defined that are generic enough to last throughout the life cycle of the product. In the development phase, changes must be made to the data definition. In addition, solutions must be found for time synchronization with the cloud; one option here is timestamps, for example, which determine when an event has occurred or when the information about this event has reached the cloud. In the maintenance phase, in which costs must be minimized, constant changes are massive cost drivers. They also impair compliance with the contractual obligations that providers of complex systems have to enter into with their end customers.

Cloud providers are aware of the problem and are working on concepts for granting different devices access to their cloud services as neutrally as possible. In particular, this approach should make it possible to offer ready-made services for networked systems from the production and logistics sector. Such services could include the monitoring of devices or the billing of associated services.

Changes on three levels

There is unlikely to be a perfect solution for synchronizing product lifecycles with the innovation cycles of cloud solutions in the near future. However, technical improvements and innovations will be available that can be seen as a major step in the right direction. Decisive changes can be expected on three key levels.

1. management of communication

One of the key concepts for improving the connection of smart devices to the cloud is certification. Devices communicate with the cloud via a so-called 'hub' (called an 'event hub' in Azure, for example); the devices can connect to it and import or query events such as measured values. The challenges here relate to the connection of native protocols or the provision of drivers for controlling specific hardware.

By publishing a certificate by the cloud provider that defines the minimum specification of a device for communication with the cloud, manufacturers of connected devices can prepare their devices for certification right from the design stage. Once a device has been certified, it can then always communicate with the cloud throughout its life cycle and successor products can be integrated more easily. Theoretically at least - the concept alone is of course no guarantee that recertification will not occur within a life cycle of five or ten years. But at least cloud providers are making a public commitment to comply with standards that is transparent for all participants.

A decisive step towards genuine lifecycle management of devices in the cloud can also be expected from the further development of event hubs into 'device hubs'. With event hubs, communication between devices and the cloud is essentially unidirectional, meaning that devices can send messages to the cloud and have them processed there. In the device hub, however, the object itself exists virtually in the cloud. The big advantage is that it can be addressed directly by the cloud. Procedures on the devices can thus be executed by a cloud service.

For example, the cloud service can use a corresponding command to request a device to transmit information about its status, or the cloud service informs the device about defined relevant events. This makes it possible to implement remote maintenance concepts: Software updates can be pushed to the devices, errors are rectified in the device software and their status can be monitored in real time. Another scenario: the transmission of relevant information can be used, for example, to send weather changes to a control unit and thus make the necessary adjustments to increase the efficiency of an energy generation system.

2. management of data streams

As soon as devices are given an identity in the cloud through new hub concepts, this also leads to expanded possibilities for controlling and analyzing the data streams they generate. 'Streamed analytics' is a process in which data is accepted and analyzed by an algorithm in real time. This means that a decision is made immediately after the data is generated as to how it should be used.

Although very complex evaluations cannot be carried out within such a framework, users can achieve a very high degree of parallelization and performance in terms of the volume of data processed. This is highly relevant for the increasingly important areas of machine learning or predictive maintenance, for example.
It can also be specified that the data is stored as values for a later, more complex business intelligence analysis (BI analysis). In addition to the previously known BI analysis based on historical data, it is now also possible to make decisions in real time. This is particularly important for industrial companies that rely on just-in-time optimization of the supply chain, for example. But operators of trading platforms are also no longer competitive without such real-time analyses.

3. data security

Hardware platforms and hardware-related code are not constantly updated and are therefore fundamentally susceptible to security vulnerabilities. Device hubs, on the other hand, enable active security management: the software on the devices can be regularly updated and patched.

Finally, the device hub concept also improves the security of communication in the cloud.

© Fotolia - Maksim Kabakou

Common security concepts, such as password protection, can be applied and passwords can be changed without the need for a field technician to carry out the time-consuming process. In principle, the cloud provider or another IT provider can completely take over security management as part of remote device management. This means that companies can make very good use of the advantages of managed services here.

Of course, this does not mean that the security of one of the most sensitive areas of a company - new innovative products - is placed in the hands of a third party without restriction. Every company will continue to define its own security guidelines in the future. It will also be necessary to implement and manage these in the cloud. However, communication in the IoT certainly opens up completely new aspects of data protection. New economic and regulatory requirements arise from the changed need for protection of 'things', which are now more vulnerable than ever from the outside. These challenges have not yet all been solved, let alone recognized.

However, avoiding the cloud does not per se imply a higher security standard for a smart device. On the contrary: in future, cloud platforms will not only provide new technical possibilities for implementing security concepts; above all, they will also create transparency and assign responsibilities - crucial aspects for regulating security issues in organizational terms via processes in companies and supervisory authorities.

Recognizing the opportunities of the cloud

Production without IT-controlled machines is already considered exotic today, and the use of cloud services is becoming increasingly commonplace. This new digital era is more exciting than ever: technology, visions and concrete scenarios are all available. At Avanade, we are seeing more and more projects from this next stage: companies are increasingly focusing on new IT approaches that enable greater efficiency, agility and freedom in terms of innovation. At first glance, this path may seem challenging and complex. This makes it all the more important for companies to recognize the opportunities and possibilities that the successful integration of intelligent systems into the cloud brings. Because one thing is certain: IoT will very soon be mandatory rather than optional.

Author: Dr. Beatrix Ertsey is Director Managed Services at Avanade.