The word Smart is often misused in marketing. Just slap that word in front of any device and suddenly you have a device capable of doing the most amazing things. We have smart fridges, smart thermostats and even smart underwear with sensors woven in the undies and bras that will measure vitals like heart rate, temperature, pressure, motion and so on. Really, I am not kidding. I believe that, unless you are a top athlete or have a chronic condition that needs monitoring, you need to be a bit of a data fetishist to wear such garments. What I mean with smart is not only connected but also capable of adding value by means of its ‘smartness’.
A smart fridge and smart thermostat at this moment are also not the most sensible products either (in my honest opinion). It would be smart if the fridge would say: you cannot eat those meatballs from two days ago since they have gone off. Or the thermostat actually manages the climate indoors in a sense that it will take into account who is home, who will be home soon (and who will leave) and taking into account the weather and traffic conditions and manage on the preferences of all individuals and economic principles (since heating costs money). I have no doubt that it would be technically possible but the timeframe and cost are to be determined.
However, I see one category of smart devices or things emerging that I find very interesting and that is the category of smart camera’s. These camera’s do not only register for instance what is happening but also interpret the information. It could be that it does that directly in the device or by sending information to the cloud. Qualcomm and the University of Amsterdam, combined QuVa lab, researches this deep learning technology for applications in Qualcomm’s products.
Cameras will actually not only record what is happening, but also do something with that information. A very simple example and something that you already see is the possibility to do detection of motion. If something is moving, an alert will be generated. If you have pets, a camera with such capabilities will get a lot of triggers since pets can roam through the house. So you want to exclude pets from triggers. Other use cases are cleaners who you know will clean your office but never at midnight. So the cleaners (with perhaps facial recognition to ID them) are allowed to be there, but only between cleaning hours.
Use it for security
You can imagine a lot of detection that could be done with a camera and combined with other information like date and time creates a smart and connected camera that does more than only record. But how are these cameras going to communicate with you? Via their app? Or by sending messages to WhatsApp or Messenger? It might even be directly to the police if they would have access to the video feed. You grant them access to view the video. Regardless who is receiving the messages, it needs to be secure. You need to secure the camera and the transport of messages so that someone with bad intentions cannot hack your feed and intercept messages. Although I believe that the average burglar does not possess the skills to do such hacking there is the issue of newspaper articles for vendors who do not want their camera’s being a newspaper headline because they were hacked. I believe these camera’s will become more and more prevalent. We might even see event driven architectures where camera data is actually used to recognize patterns. Input streams can come from many types of devices and combined. A door opening and closing and a rise of temperature can indicate a person locked in a cargo hold. Even though people might not be visible, other sensors can actually detect their presence.
Smart IT infrastructure
It’s another good example world of IoT becoming more and more part of your IT landscape. Especially when you combine machine learning together with more basic sensors like motion sensing, camera streams or temperature rising or dropping (i.e. with perishables). We will end up with a whole new world of technology that will be part of your IT landscape. And of course you need to prepare an environment where you are adding devices with such capabilities into your IT landscape. They will need to communicate or talk to other systems, being part of the larger IT infrastructure. Integration will be part of the perspective that we know now, like soap messages and APIs. Soap seems to be an odd candidate since it is falling out of favor but there might be backend systems they need to connect to with the messages that are soap based. This model is trigger based. If something is out of the ordinary, we want a message. It might be that we simply just listen to a stream of data coming in and let machine learning detect what is happening using models that we created. It means that there will be modern ways of integration as well as older ways that we need to support. For such an environment you need a top-notch Enterprise Integrator to be ready for the future.
If you are not sure your current solution is up to the task, I would suggest that you take a look at our ESB selection guide. It will help you determine what characteristics such a solution should have to actually enable those kinds of future technology.