The Anatomy of an IoT Solution
Embark on a crash course to learn how you build an IoT platform or IoT device architecture, and which layers work together to make up such a solution.
Join the DZone community and get the full member experience.Join For Free
Artificial intelligence relies on data. One of the leading ways to collect data these days is by using the Internet of Things. IoT connectivity is often used in devices containing sensors of some sort, allowing us to automatically collect information that either was not collectible before or which needed a human to collect it. Let’s go on a crash course on how you build an IoT platform or IoT device architecture, and which layers work together to make up such a solution. For this example, our IoT connectivity will center around cellular and the use of IoT SIM cards.
To begin with, it’s important to realize that there are four individual layers involved in building an IoT solution. Starting from the bottom — with individual frontline devices containing IoT SIM cards on-board — and working upwards to the networks over which your data is transmitted, and the IT applications that let you manage and control your entire fleet of IoT SIM cards. The layer hierarchy is arranged as follows.
Making a Start: The Sensory Layer
This is where data capture takes place. Without the crucial first steps that happen here, no IoT solution could function properly. There’d be no information for the wider solution to base its automated decisions on.
The data points that are generated here are logged by your IoT devices’ onboard sensors. These sensors could be almost anything, but some commonly used ones include temperature sensors, humidity sensors, pressure sensors, and even sound and video sensors. These sensors are the first element of automation in your IoT solution and are the things that prevent you from having to physically go somewhere and measure how hot, cold, wet, loud, or bright (for example) the environment is with your five human senses. On top of this, because everything’s being done by machines, it’s much more precise too.
There is also data that can be created by humans — device ID numbers for example — but whatever the source may be, any data that a device’s sensors record is given to their IoT SIM cards for the next step in the process, as outlined below.
Making a Connection: The Network Layer
Thanks to their onboard IoT SIM cards, each device at the sensory layer will have the ability to send the data it’s collected over a network of some sort, to a central storage location — for example, a server. That data transfer happens in your IoT solution’s network layer.
As it serves as the conduit that connects your IoT SIM cards and the data they collect to the rest of your IoT solution, there isn’t really much that can be said about the "features" of a solution’s network layer. However, it really can’t be overstated how important the network layer’s role is. Factor in that it also allows layers closer to the top of the solution to pass instructions and software updates to the IoT SIM cards in your front-line devices, and that role only becomes more important!
Making Sense of It All: The Device Management Layer
Until this point, the data in your solution has remained largely untouched… The IoT SIM cards and devices at the sensory layer captured it, but they aren’t concerned with anything but the raw values detected by the sensor. Likewise, the network layer is only concerned with getting the data from the source to the storage and processing location. Here at the device management layer is where the raw data that’s sent by your IoT SIM cards back to your central location is first made sense of, and the information it contains is first understood.
Without the device management layer, an IoT solution would do nothing but generate a stream of meaningless numbers and values. But with it, the next layer in the solution has all the information it needs to analyze situations and provide instructions. For example, questions such as, "What device sent this data?", "When did it send it?", "What was being measured (the metric)?", "What was the value that was recorded for that metric?" are just some of those that are answered in the device management layer.
Making Decisions: The Application Layer
The application layer is where your IoT solution makes its decisions. The application layer takes what is learned at the device management layer, analyzes it, and based on its findings, can either provide insight or instructions to be sent back via the network layer to the IoT SIM cards in your devices. What’s more, the application layer can also track and save values, creating an archive of past values that can be used for reporting purposes or for things like trendspotting.
An example of these last two layers in action might be as follows: The application layer receives information from the device management layer that the temperature in a room is two degrees centigrade. The application layer logs the data value then checks to see if the temperature in the room is allowed to be two degrees centigrade. If so, it does nothing and waits for the next chunk of data to come in. If the room is not allowed to be two degrees centigrade, the application layer looks up what should be done in that event (for instance, "send instructions to turn on the heater to device x") and carries out that action, with the instruction being sent to "device x" via IoT SIM cards and the network layer.
This last step we really feel pulls everything together and turns what otherwise might be four separate steps into one cohesive solution. The key parts of that solution are the connectivity, or IoT SIM cards in this case, which your devices need not only to begin the process but also to receive any instructions or updates that they need to apply — a network for your data to be sent across and a device management/application layer solution.
Published at DZone with permission of Nicola Pickard. See the original article here.
Opinions expressed by DZone contributors are their own.