Whether in production environments, smart trailers or even at the point of sale: Internet of Things (IoT) infrastructures are highly developed technical elements embedded in complex architectures. From individual sensors in garage doors or truck trailers to entire sensor chains in smart industrial plants - the use cases and underlying processes are wide-ranging. It is not uncommon for sophisticated applications to require regular updates at one end of the chain as soon as a new generation of even more sophisticated devices has been installed at the other end as the IoT infrastructure grows.

As fast as the development in the field of networked devices is currently progressing, the characteristics of the individual devices are already diverse and wide-ranging in practical areas of application - especially if these have been growing for years. It is not uncommon for an environment to use different hardware architectures with different processors or network capabilities, different operating systems, versions of the same operating system or versions of their software and even different transmission paths for the collected data. The complexity is compounded by the - by definition - distributed nature of IoT across different locations. Different locations that may even require different connection and transmission paths. Chaos sends its regards. They either communicate with each other or with services that run in the cloud or in a private data center.

Regardless of what happens to the data in the first step, many modern industrial companies that build IoT products and services are now pursuing a cloud-first approach. This means that the devices must always communicate with a service running in the cloud or centrally. This processes and stores data and provides insights, which in turn flow back to the devices.

The bottom line is a heterogeneous collection of transmission paths, hardware and software, which often presents companies with new challenges if they want to maintain an overview. Without robust monitoring, even the most sophisticated IoT devices are sometimes flying blind thanks to increasingly complex applications. Only those who are able to build up an understanding of communication, data flows and the entire end-to-end system can look beyond the individual devices and services to assess the health and performance of the overall structure.

Performance at a glance

Although the white spots in the Federal Ministry of Transport and Digital Infrastructure's broadband atlas are becoming fewer and fewer, IoT installations in remote locations in particular cannot simply ramp up instances when their applications are in greater demand and performance decreases. Also, not every IoT installation provides for such scalability at all, but is a self-contained system limited by factory settings. Where "classic" cloud applications simply scale horizontally at other locations and in other environments, the physical hardware at the deployment site limits the possibilities. This makes it all the more important to seamlessly monitor the performance of IoT devices and their decentralized data processing, for example to ensure seamless predictive maintenance or to avoid slowing down smart supply chains.