The Internet of Things (IoT) is just getting started. Honestly though, if you think about it, IoT has always been here. In the beginning, we had the internet. Traditionally, the internet’s purpose was to allow devices to serve content from the internet. Now with IoT, we simply alter the paradigm and consider that the device is sending sensor data to the internet. In other words, we have always had the ability to do IoT, but what has changed is the notion around how devices and device data can be used. One of the reasons for this is the resurgence of big data and new innovative tools which make big data more effective. As new approaches to big data create more compelling business cases, the demand for more device data accelerates.
One of the reservations companies have against starting an IoT project is a misconception that because they don’t manufacture the end-product they cannot deliver an IoT solution. When you think about the weather, you can view weather forecasts on the internet. However, you cannot change the weather as you monitor it. In other words, you monitor the weather passively. In many cases, you can monitor third party devices without changing the device by monitoring characteristics about the environment in which the device is operating. For example, a hygienist will tell you that they know when a dentist drill will fail based on the sound the drill makes. Therefore, you don’t need to change the drill you simply monitor the ambient acoustics with machine learning to do your predictive analytics.
A caution I have for companies selecting IoT platforms is to not roll your own architecture. It’s true some organizations have a decade old architecture because they were doing “IoT” before it was IoT, but for new initiatives rolling your own IoT architecture is fraught with security concerns. The reason is because the “S” in IoT stands for Security. Joking aside, the cost of failure in an IoT environment is high, especially in industrial or healthcare applications, which is why you need to make sure your bases are covered. Using a mature IoT platform like Microsoft Azure is one way to mitigate this concern. This concern also has regional sensitivities. Through our work with industrial IoT, we have noticed European manufacturing workloads to have a heightened sense of concern. This is indeed warranted as a famous incident of hackers compromising a German mill’s blast furnace a few years ago points out. In this specific workload, one of our approaches in the past was to use a military spec one-way optical diode so the field gateway we created can talk unicast to the internet. Besides security considerations, another reason to use an IoT platform like Azure relates to the supportability of the solution. Using an obscure IoT platform means there are fewer partners in the market able to assist you, especially if that IoT platform vendor is looking to be acquired and has an unknown roadmap.
When you look at IoT through an innovation lens, you can better plan around trends in the market. For example, by using a theory like the Theory of Interdependence and Modularity, we can plan for modularity entering the IoT space. According to Gartner, within a couple years about half of companies will reject products from new vendors which contractually block access to the underlying product’s sensor data. In other words, unless a company is leasing the asset, a company assumes that the data generated by that asset is in fact theirs. This mindset is important as it will force vendors to ultimately add open APIs to their products to stay relevant. Granted, this type of modularity in the ecosystem will still be years out, knowing how this modularity will enter and exist in the market will help you plan which parts of your IoT stack should be considered disposable, when future solutions will accommodate natively. This understanding helps you determine which parts of your solution will require you to integrate forward until future solutions exist. Simply waiting until future solutions exist is still not an option because the value of IoT to either new revenue or increased margins requires action now, given the current arms race with technology, fueled by increases in corporate discretionary funding of futures and innovation projects.
The Internet of Things is just getting started. At the same time, it’s been around for a while. What has changed is the way we think about device data and the cloud. Your introduction into the IoT space should be well thought out with respect to what your customer really needs. Customers do not pay for IoT because no one buys IoT. Customers pay for increased capabilities and value in IoT enabled services and products. Besides delivering value, you will need to ensure that your platform selection is maintainable, safe, and can adapt to future changes in the IoT ecosystem.