Choosing the right IoT platform

The Internet of Things (IoT) is one of the hottest talking points in tech right now. Almost every day, new IoT products and services are unveiled and strategic alliances announced. What’s more, the range of what are known as ‘IoT platforms’ is practically exploding – even if, on closer inspection, some of these offerings do not (yet) live up to this billing. There are now more 350 commercial IoT platforms, and they are growing in number all the time. The range of platforms on offer varies greatly, while selecting a platform has far-reaching consequences. CIOs and other technological decision-makers thus face a formidable challenge when choosing the IoT platform that will be right for them in the long term.

What are IoT platforms?

IoT solutions make simple ‘things’ – such as machinery, cars or articles of clothing – ‘smart’ by connecting them to the internet. In simple terms, this requires three basic elements:

  • physical ‘things’ equipped with sensors, actuators, appropriate firmware and connectivity

  • web applications or mobile apps enabling users to access and, where necessary, control the data from these ‘things’ on different end devices (such as smartphones, apps or PCs)

  • an IoT cloud platform that receives and stores the data from the ‘things’ together with control signals from users, processes this information in accordance with prescribed rules and provides various administrative capabilities, thus connecting users with these ‘things’.

Many companies use so-called IoT platforms to create an individual IoT cloud platform as quickly and efficiently as possible. These software services already include many of the standard technological functions offered by IoT cloud platforms, ensuring that these functions do not have to be recreated at great expense for each IoT project. While the main IoT platform providers include tech heavyweights such as Amazon, IBM and Microsoft, they are also joined by traditional industrial companies such as General Electric and Hitachi.

Although there are clear differences depending on the focus (consumer vs. Industry 4.0 solutions) and specific sector or application of the IoT platforms, most ‘end-to-end’ offerings generally include the following basic features:

  • connectivity
  • data standardisation/data management
  • data visualisation
  • device and service management
  • external interfaces
  • development support
  • security features.

However, many services advertised as ‘IoT platforms’ do not offer all of the aforementioned features, for example pure connectivity/M2M offerings or IaaS back-end solutions. Despite what their marketing literature often suggests, these offerings do not strictly constitute IoT platforms due to the absence of an ‘end-to-end’ service.

Spoilt for choice

In our IoT projects, we have been seeing for years that companies are very uncertain about selecting the ‘right’ IoT platform for their individual project. There are four primary reasons for this:

1. Large, varied offering

There are over 350 IoT platforms to choose from, and they vary widely. Interestingly, many of the confident assurances offered by providers do not stand up to a thorough code review, making evaluating the actual quality of a platform an extremely time-consuming process. This evaluation is also specifically designed to shed light on aspects that are crucial to the future viability of the platform (such as scalability and the integration of additional device types).

2. Long-term lock-in

As the core of any IoT solution, IoT platforms are deeply integrated at both device and application level. For example, when switching to a different IoT platform, firmware updates usually have to be rolled out to all devices in the field, which requires significant effort and increases the potential for errors. Of course, this is only possible if the devices can be updated in the first place. In the worst-case scenario, companies must either implement a complete vendor lock-in or replace every device.

3. High costs

In addition to significant one-off implementation work, IoT platforms can also incur high running costs, such as for scalable delivery of device connection end points and the number of messages sent via this method.

4. Strategic factors

The realisation from Silicon Valley that ‘data is the new oil’ is now well established, including in many traditional companies. With software making an ever-greater contribution to value creation and competitive positioning, this also means, for example, that a German engineering firm would not necessarily be well advised to share all the operating data for its products with a major US tech company. This is because the latter may sell the resulting information to a direct competitor of its customer or may even itself become one of its client’s (indirect) competitors in the future as industries increasingly converge.

How do you find the right IoT platform?

Selecting the wrong IoT platform may eventually necessitate an expensive platform change, triggering the threat of huge additional costs, considerable ‘time to market’ delays and significant strategic drawbacks. That is why a thorough analysis of requirements and an extremely systematic approach are worth the effort when evaluating the ideal platform for your individual IoT project.

It is important to start by gaining an overview of the products available on the market and the primary scope of their services in order to draw up a rough shortlist. For example, we keep a constantly updated database of all of the IoT platforms known to us and their main characteristics to enable us to quickly set the initial perimeters for our search.

On this basis, it is worth carrying out a detailed analysis of your own project as well as the functionalities and special features of each IoT platform. To help you do this, we have compiled our key findings from more than 30 IoT projects into a checklist with 75 questions and have divided this into ten sections:

  1. Embedded software development kit or SDK (e.g. what device and gateway implementations does the SDK support?)

  2. Cloud connectivity (e.g. how can certificates be updated on devices in the field?)

  3. Scaling (e.g. does the solution scale automatically?)

  4. Server interaction with IoT devices (e.g. are online and offline events actively reported by IoT devices?)

  5. IT security and data protection (e.g. is it possible to place suspicious devices into a quarantine area?)

  6. Vendor lock-in (e.g. what changes are required when switching providers?)

  7. IoT device production (e.g. how are device identification codes generated and how can these be integrated into the device production process?)

  8. Administration (e.g. are applications or interfaces accessible to administrators?)

  9. Hosting (e.g. does the IoT provider support global rollouts with communication end points in key countries?)

  10. Support and SLAs (e.g. how can support for the IoT platform be incorporated into the company’s customer care?)

This checklist helps us to ask the right questions of both the IoT platform providers and your own IoT project and the project team. Of course, this also means that you need to analyse the exact requirements of your project very carefully in advance in order to understand what you need to demand (as a minimum) from providers.

But doesn’t this contradict the very idea of an agile approach? Absolutely not. Of course there are going to be plenty of unforeseeable changes during something as innovative as an IoT project – particularly to the front end of applications in the short term, but also to the connected devices (or device types) in the long term. However, whether you have 10,000 devices within your ideal framework, each of them sending one binary signal every day, or several million devices constantly transmitting dozens of data points that need to be processed in real time makes a huge difference when it comes to choosing a platform.

We would be glad to discuss the different options within an individual workshop: Workshop IoT platform.

Contact: Jan Rodig (jr@tresmo.de)/ +49 176 105 191 50