Independence from individual technologies and platforms through microservices

A microservice is a software architecture that is based on many small, self-contained services. It represents a sub-function of a larger system. In general, they do not fulfil several functions at the same time, but specialise in one task: for example, a microservice can be used for Predictive maintenanceThe system can predict failures based on existing machine data, regardless of the interface or technology. Another microservice then takes care of the Visualisation of the datawhile another application reads data from a machine and writes it to a data memory.

A microservice is also always self-contained and can be created on the basis of any technology. Only the interfaces for communication with other systems need to be defined, but they can also be implemented independently of the programming language. This makes it possible to adapt, expand or replace them independently of the entire system. This creates independence from individual technologies or platforms.

In contrast to this is monolithic software, which combines and centralises all functions. Here, everything is developed on one platform or with one technology, which creates a dependency on the specific platform, programming language, protocols or company.

Independence is particularly important in the IIoT environment, as few standards or platforms have been established to date and companies need to react flexibly and independently to developments.

As microservices are a detached software architecture, they can be used in all sectors and areas. They offer major advantages over monolithic software.

For example in the beverage industry: A beverage bottling plant is operated with condition monitoring software that has been implemented monolithically in Java, for example by Monolith AG. However, problems arise: If a tablet app is to be developed for this system, the monolith must first be adapted, as access to the user assignment cannot take place without an authentication service. In order to analyse the filling speed, an additional algorithm needs to be written, which must also be connected directly to the machine. This requires an interface to the monolith. However, this is not available, as there is usually no single service for data storage. How can this challenge be solved? The company must either commission Monolith AG (lock-in effect) or turn to an agency that can develop in Java. In the latter case, the agency's programmers must first familiarise themselves with the monolithic system, which means a great deal of time and increased costs.

Using microservices results in a completely different scenario: if defined microservices have been used for the bottling plant, any company can be commissioned. The only requirement is that the interface must be known. This means that the company remains largely independent and flexible. The modular structure allows the software to be flexibly adapted to current requirements. The use of microservices is therefore cost-efficient and time-saving for a company.

Only a few prerequisites are needed to be able to use microservices. To implement a microservice, the interface for communication with other systems must be defined beforehand. A microservice is always self-contained and can therefore be created on the basis of any technology. Microservices guarantee further independence from other systems through their own data storage.

Microservices only cover a very manageable area. They can therefore be easily replaced or rebuilt. For a SCRUM team, for example, it is possible to develop a service within a month without any problems. Microservices are also horizontally scalable. This means they can be scaled infinitely with additional computing capacity without the need for software customisation. Logging, monitoring and security methods should be implemented to make the overall status observable and secure.

The benefit for the user is always the focus of a microservice. During implementation, care must therefore be taken to deliver a usable version as early as possible that can be iteratively improved. After implementation, microservices can be subsequently adapted, expanded or even replaced - independently of the entire system.

It is advisable to work with a small team for the implementation. According to SCRUM, this is a maximum number of seven members. The duration of the implementation should also be manageable and take a maximum of one to two months.