The reasons why industrial companies decide to use cloud resources can be manifold: technologies behind terms such as industry-of-things, data analytics or machine learning generate vast amounts of data that need to be collected, transferred and processed. In the rarest of cases, it is even possible to predict with any degree of accuracy in which dimension and at what time the data will actually be generated. In practice, it is more a question of volatile data streams - sometimes with extreme load peaks - combined with very high availability and security requirements. Sometimes companies have simply paid too little attention to the ongoing modernization of their IT, delaying necessary investments for various reasons.

The cloud comes in handy as a flexible basis for these developments. Lengthy, internal procurement processes can be avoided, as can high investments in hardware that may later remain temporarily unused. Prototypes can be tested easily and seasonal load peaks can be covered flexibly - payment is based on consumption!

Just book cloud resources

At least in theory. Because the promise of cloud providers to be able to book IT resources quickly, easily and on demand is not so easy to fulfill in practice. On the one hand, companies that want to use cloud resources are faced with integration and management tasks in order to control the data flow and integrate the new infrastructure into the existing operating model. At the same time, the self-imposed security and compliance rules continue to apply without restriction. On the other hand, the challenge begins much earlier: when setting up the cloud.

The offerings of the major cloud providers, such as Amazon Web Services (AWS), Google Cloud and Microsoft Azure, are extensive and powerful. At the same time, it is not easy even for IT experts to put together the right offer. To take just one example, AWS offers more than 150 different solutions: from the most frequently booked EC2 (Amazon Elastic Compute Cloud) to integration and control tools such as Amazon RDS (Relational Database Service) and Amazon CloudWatch through to the Lambda serverless platform, which can be used to run programs directly. Extensive descriptions, tutorials and manuals are available for all applications in the online resource center. So the first thing that stands in the way of a quick start is the need to familiarize yourself with the offer in detail without knowing whether the selected options are optimal for your own requirements or how they should be dimensioned.

It is almost impossible to estimate in advance what the cloud resources will cost. Almost every service has its own cost structure. In principle, the costs are calculated according to consumption, but the amount per volume varies according to region or the type of interfaces used, for example, and there are certain scales, minimum purchase quantities or discounts. The more services are involved, the more complicated the price structure becomes. Although several tools are available for calculation, it is not possible to find the optimum price for the desired solution.

The cloud scales itself according to demand

There is another way: Cologne-based cloud provider Gridscale, for example, defines the cloud not from the supply side, but from the user side. Each instance has a fixed price that is calculated based on usage. This is transparent and easy to manage. Instances can be switched off quickly via the API if required and then immediately no longer incur any costs.

But cloud usability can mean even more than simple handling: a platform-as-a-service, for example, that adapts flexibly to the business as a fully automated cloud operation, ideally without user intervention. Such an intelligent algorithm enables a range of data-based cloud services. Based on the experience gained, the system can calculate forecasts of how the workloads will develop. Additional capacity is automatically made available when it is needed - known as autoscaling. It is also possible, for example, to migrate workloads live. This allows them to be redistributed because maintenance work needs to be carried out on the server or they need to be moved to a cheaper resource. Attempted intrusions into the IT infrastructure are also undoubtedly anomalies - and these can also be detected at an early stage using certain parameters, such as specific patterns in network traffic.

Artificial intelligence is thus finding its way into the data center. A machine learning algorithm monitors all the metrics of the cloud environment, such as CPU utilization and temperature, the number of I/O accesses, latency times and much more. In a lengthy process, the algorithm learns from practical events what anomalies can be compared to normal operation and what measures need to be taken. The goal behind machine learning in the cloud is clear: the user defines which parameters the cloud should fulfill in terms of performance, costs, availability and security. The intelligent algorithm implements these and constantly adapts the cloud accordingly.

More cloud usability thanks to AI-controlled user interface

To come full circle to the initial criticism of AWS's confusing offerings: Thanks to an intelligent algorithm, setting up a cloud can be quite simple. This is because the way in which users proceed during setup or navigate the management platform provides information about how experienced they are and which options make sense for them. Looking to the future, this offers the opportunity to provide every user with the right user interface: A new customer, for example, is only offered basic options; at each step, the engine recommends suitable, optimizing additional features. A cloud expert, on the other hand, can directly unlock in-depth insights into their resources and decide for themselves how much management they want to take on. The use of cloud resources can also be optimized in this way. Frequently used functions are either easily accessible or automated. For example, many users regularly take snapshots of certain data. If the algorithm recognizes the regularity, it can automatically relieve the user of this work.

With an intuitive user interface and largely automated operation, the cloud is becoming the flexible resource that industrial companies need. The core business revolves around production, the automation of complex processes, the integration of suppliers and ultimately the transformation of the business to digital. If cloud offerings can be seamlessly integrated here and are designed in such a way that they can be configured quickly and flexibly without in-depth cloud expert knowledge, then - and only then - will they make a valuable contribution.