How Does IoT Work? Talking Through the Basics of IoT Architecture

From smart farming to smart ovens, connected devices have the potential to solve million dollar business dilemmas and the everyday problems of the average Joe. What makes smart devices truly transformative is their ability to connect to the internet in a system known as the ‘Internet of Things’, or IoT.

What is IoT?

While smart ovens have yet to make their way into the mainstream, they’re a small example of the power of the Internet of Things to change our daily lives. At its most basic level, IoT is a system that connects every and any object to the internet. Thus, the internet serves as the intermediary through which devices can communicate with each other. This exchange of information between devices allows them to influence each other as part of the IoT system.

There are two primary types of IoT applications: enterprise and consumer. Consumer IoT is generally products marketed directly to, well, consumers. Most ‘smart’ devices on the market would fall under the consumer IoT category. Enterprise IoT, on the other hand, delivers real-time insights to companies about their assets, removing the need for manual tracking, check-ups, and guesswork.

While consumer IoT has dominated mainstream media, enterprise IoT is poised to deliver the most disruptive change. Rather than turning to expensive and time-consuming manual labor, enterprise IoT systems provide companies with real-time insights into the function and location of every process important to their organization at their fingertips. Because IoT is comprehensive, this article will focus on unpacking an enterprise IoT system.

To begin to wrap our minds around the immense power of enterprise IoT, let’s break it down. While enterprise IoT systems span a huge range of industries, all systems are similar in that they include four conceptual components: devices, connectivity, computing, and business integration. To explain each of these four conceptual components and how they work together, let’s look at an example of an enterprise IoT system:

Enterprise IoT in the Real World: Asset Tracking

It’s hard enough to keep track of your own keys. Imagine trying to simultaneously keep track of thousands of keys that are constantly moving locations and being required to find the location of a given key at any point in time. Sound impossible? That’s where an enterprise IoT asset tracking system comes in. 

While asset tracking is by no means the only use of an IoT system, it’s an example of an IoT system that can be applied to almost any industry. Attaching devices to assets allows organizations to track the location and monitor the assets' performance in real-time. Cox2M’s vehicle management system, Lot Vision, allows users to check vehicle location and battery from a web application. Lot Vision is used at automotive auctions and dealership lots. It enables users to locate a specific vehicle within 6 meters of accuracy in a lot that might contain tens of thousands of vehicles. 

Let’s dive into how each component of IoT functions in this vehicle management system. 

Components of IoT

Devices

Devices are the first layer of an enterprise IoT system. They consist of the hardware (wireless radio, battery, microprocessor, sensors, etc.) and firmware, integrated. Devices are linked to an object and capture data about that object (e.g., location, status, etc.). You can think of devices as gathering the ‘inputs’ to the overall system.

The term “sensors” is often used interchangeably with “devices.” An important distinction between sensors and devices is that a device can bundle multiple sensors within it. For example, your smartphone device contains a GPS module, a camera, a touchscreen, and an accelerometer—all of which are sensors. However, the smartphone is not only a sensor, as it also does computation and wireless communication.

Devices use sensors to gather data from the environment and then communicate that data using wireless radios on the device. In a smart farming system, what devices sense can range from the system's temperature to the humidity of the soil to full video. 

Devices are important because they allow us to gather data about events as they occur, thus allowing us to make more intelligent decisions. Gathering live data can be extremely valuable in medical applications, especially. A device may detect a patient falling and trigger the alarm framework in an IoT system, immediately notifying emergency responders. In this instance, live sensor data could be the difference between life or death. After collecting data from their environment, devices then send that data to the decision-makers elsewhere in the IoT system.

In the Lot Vision system, each vehicle has one attached device, installed by lot personnel. The device gathers location and movement data using GPS and an accelerometer and can be plugged into the OBD port to read vehicle battery and trouble codes.

Connectivity

Data collected from devices need to be sent to the computing layer for processing. These devices can be connected through various methods, including Bluetooth, WiFi, cellular, LoRa, satellite, and much more. Thus, it may be helpful to think about connectivity as the ‘transport’ layer.

In some cases, rather than send data to the Cloud for processing, data can be processed on the device itself or on a gateway, a process known as edge computing. Edge computing is especially important in applications where latency, bandwidth, or battery lifetime are important factors.

Because there are many connectivity options for different applications, there is no one-size-fits-all approach. Each option has a tradeoff between power, range, and bandwidth (how much data can be transmitted). Your cellphone is an example of high power, high bandwidth, and high range connectivity. Because it can transmit and receive large amounts of data over a wide distance, it needs to be changed every 1-2 days or connected to a power source.  

For Lot Vision, because the vehicles are constantly moving, long-range wireless connectivity is necessary. High bandwidth isn’t necessary since the amount of data sent is low (mostly location and device status data). Still, since these devices are battery-powered, it’s necessary to minimize battery drain. If the devices had to be changed frequently, the operational overhead to manage the IoT system and the thousands of devices would be immense.

Therefore we need long-range, low bandwidth, and low energy consumption in our connectivity choice. Cellular NB-IoT is a good choice if the vehicles travel out of range of the auctions or dealerships. LoRa is a great option if the movement is primarily local.

Computing

Typically taking place in the cloud, the computing layer represents algorithms, logic, ingestion servers, cloud infrastructure, and data storage that transform the data and make it useful. As soon as the device data reaches the cloud, data processing begins. This includes anything from applying advanced machine learning to a video stream to identify home intruders to the simple logic of checking whether the measured temperature is within an acceptable range.

For Lot Vision, the raw vehicle location data from the devices is ingested in the cloud, processed, and stored to make it easy to show vehicle location on a map or find vehicles quickly when searched by users. Also, alerts rules are applied to trigger alerts when device batteries are low and need replacement or when vehicles unexpectedly leave a geofence.

Business Integration

The Business Integration layer ties the computing layer into the enterprise to make it valuable. Only knowing where a device is doesn’t matter unless you know what asset it is paired with and information about it. The enterprise usually has existing business systems (e.g., ERPs, CMSs, etc.) that contain information about the enterprise’s assets, sales, marketing, and more.

The true value of IoT lies in combining data from devices with other data streams. For example, in Lot Vision, it’s useful to know where all the vehicles are so that users can find them quickly. However, it’s even more valuable when the enterprise can define different zones on the auction or dealership lot and get insight into how vehicles flow through those zones. If a vehicle is sitting in the staging zone for too long, it would be nice for the enterprise to get an alert.

But how long is too long in a zone? What are the zones in the first place, and why do they matter? Both these questions and more are going to be different for every business. Therefore, business integration is key because it represents the configuration and customization of the IoT system for each specific enterprise to make it valuable within their unique business context.

Business Integration comprises APIs that pull data from existing business systems into the IoT system and push data from the IoT system into the existing business systems. Business Integration is also composed of User Interfaces (UIs) configured and customized to the enterprise to enable users to get the insights they need when they need it.

For Lot Vision, APIs pull vehicle sales data from auction and dealership lots into the IoT system to supplement the vehicle location data. Besides, the Lot Vision UI can configure alerts for when vehicles leave certain zones or remain in a zone for too long.

IoT Transcends

Each layer of an IoT system is crucial to ensuring the enormous amount of data is processed to be ultimately useful to the users. Especially in an enterprise IoT system, where raw data is often coming in from thousands of devices at once, optimizing each layer of the system is critical to its success.

With the potential to automate what currently requires huge amounts of manual labor and time, enterprise IoT has massive implications, and the current applications have only unlocked a small portion of the vast power IoT holds. IoT transcends industry and has the ability to transform the way we all live, work, and play.