This article is part of a 3-part series. Be sure to read them all!
Part 1
Part 2

IoT sensors are powerful devices that can fundamentally change the way systems and entire supply chains operate. But putting sensors to work is no simple task. In the last installment of this three-part series about IoT sensors, we examined the role of sensors in developing more efficient solutions. Yet understanding what sensors do and how they work is only a starting point. A sound IoT sensor strategy ultimately revolves around several key questions, according to Benson Chan, senior partner at Strategy of Things, an IoT consulting firm in Hayward, Calif.:

• What is the IoT application? It’s critical to thoroughly understand the use case, Chan says. This is the starting point for determining how to design an application, what type of sensors to use, and where the sensors must be located, among other questions.

• What data is required? Identifying the specific data that’s needed determines how to design and build an IoT framework. For example, tagging events in real time or by location requires specific kinds of sensor and communication protocol. Precision data collection involves sensors capable of measuring conditions at a higher level of accuracy.

• What is the framework surrounding the data? Does the IoT application require data in real time or at certain intervals? Do events or circumstances trigger the need for data? It’s crucial to match the IoT framework and data feed to project requirements. This also determines the IoT network topology, including whether to use centralized processing and storage, edge processing, or fog networks.

• What conditions will the sensors operate in? Pressure, temperature, vibration, radiation, and other factors take a toll on sensors, and their performance. It’s important to consider the environmental conditions that sensors will operate in and use that information to determine how they are housed, how often they are inspected and when they should be replaced, Chan says.

• How will the sensors be powered? Depending on how and where IoT sensors are used, it may be necessary to rely on electrical power, battery, solar, or other power methods.

• What are the security requirements? The sensitivity and value of data must guide careful thinking about security methods, including authentication, authorization, encryption, and data storage. It’s also crucial to examine vendor firmware and patching policies for specific sensors and devices.

• What data formats are required? IoT solutions are a collection of sensors, software, and systems. Ensuring that data flows freely across various hardware and software components, and increasingly across company boundaries as well, is vital. This translates into a need to use APIs, middleware, and open source components effectively.

• What is the total cost? It’s crucial to determine the upfront expense for an IoT application, as well as the ongoing operational, maintenance, and expansion costs. For example, depleted batteries can boost the total cost of ownership. “While the batteries themselves may not be very expensive, the labor to access and replace them can be very high, negatively affecting operating expenses,” notes Roman Staszewski, president of IoT firm Zenseio LLC, of McKinney, Texas, in a blog post. 

• Do the IoT sensors require specialized skillsets? Some IoT frameworks are simple to set up and operate while others are incredibly complex. Staszewski has pointed out that IoT sensors that are difficult to configure and use often introduce major challenges. A similarly significant problem is that “embedded engineers are too expensive to hire” and training costs can undermine ROI.

A sensor is the starting point for all IoT projects. Choosing the right sensor, and the right supporting framework, will greatly improve your IoT success rate.

Samuel Greengard is a business and technology writer based in West Linn, Ore. He is the author of The Internet of Things (MIT Press, 2015).