Balancing reliability, cost and context in IoT networks
This article was guest written by Iain White, IoT Architect at Intelligent Growth Solutions (IGS). Iain is responsible for integrating sensors, actuators, data, and analytics across our platform to optimise both crop growth and yield.
In vertical farming, reliable data is central to creating a stable, predictable and repeatable growing environment. This is where a thorough Internet of Things (IoT) network comes into its own: collating a bank of actionable datapocints and enabling growers to tweak the environment, directly impacting plant quality.
But a high‑quality IoT network doesn’t come from treating every datapoint as equal or duplicating sensors unnecessarily. It comes from knowing which signals matter most, where redundancy (the strategic duplication of critical hardware or software to ensure continuous, reliable data transmission and prevent downtime if a component fails) makes sense, and how the context of an environment can shape decision‑making.
We use IoT in a pragmatic way to help our customers utilise data to optimise crop yield and quality. Our systems are built to give growers visibility of farm performance, but also the insight to understand what can have the biggest impact on growing crops. My role is to ensure our IoT networks are not only reliable but also aligned with a customer’s wider operations.
How a reliable IoT design can add customer value
It’s important not to adopt a one‑size‑fits‑all approach to data streams or the value placed on their reliability. A common mistake when constructing IoT networks is assuming that all streams need the same level of protection, when this isn’t the case. Some signals influence real‑time decisions that affect crop quality, whereas others provide visibility rather than immediate control.
Generally, these datapoints can be split into two categories:
- data that feeds a control loop – this needs redundancy or a mechanism that can be triggered for safety purposes
- data that is purely telemetry – if data is analysed retrospectively, duplicating the hardware used to collect it usually adds cost and complexity without meaningful benefit
By adopting this approach, we know which areas to focus on to ensure continuous crop quality versus those which enable us to monitor trends from a wider system perspective, but don't necessarily benefit from real-time feedback. Both contribute to reliability, ensuring a high-quality end product (the crop) through well‑functioning vertical farming technology.
Redundancy through clever distribution
When designing a vertical farm, particularly when considering large vertical units such as our Growth Towers, you need to consider multiple sensors within a single location. This doesn’t mean placing two sensors side by side, however, something that would add cost without delivering additional visibility.
A better approach is spatial redundancy: distributing sensors across different points in the growing environment. This gives growers two advantages:
- balancing reliability and cost – having the right sensors in the right places ensures efficient use of spend, placing data collection points exactly where they are most useful
- outlier detection – in the unlikely event of a sensor failure, it would be easily identified against neighbouring readings
This approach builds reliability through the layout of datapoints, providing thorough context on plant growth through strategic placement.
Using historical telemetry data to inform current design
When you collect large amounts of data, it often proves useful in unexpected ways. After more than a decade of designing and deploying vertical farming technology worldwide, we’re fortunate to have a library of datasets that inform both current and future design iterations. One example of this is how we repositioned the lift to improve connectivity.
.jpg)
During testing, we found that sensor devices were dropping off the network at specific times. This seemed sporadic, but after reviewing Received Signal Strength Indicator (RSSI), we discovered the dips aligned with lift movements. The lift was briefly blocking radio pathways, so we repositioned lift access points to eliminate the issue.
Historical telemetry data is relatively inexpensive to store, yet it’s a powerful tool that helps optimise performance for growers and supports future design iterations. We can use raw telemetry data internally to help inform our product development and support customers, translating this into meaningful insights to support user operations. This level of visibility gives us and our customers confidence, as well as the oversight needed to maintain well‑structured data.
Clarity creates reliability
Reliable IoT systems are built on clarity. In practice, this means clear design, clear understanding of the role of each network component, and clear visibility into the behaviour of the system it monitors. When you can see what is happening, optimising performance becomes easier, giving growers confidence that their systems will perform and meet their needs day to day. It doesn’t need to be overly complicated – it just needs to be clear. This helps embed repeatable, predictable performance in a system designed to stand the test of time.
Enjoy what you’ve read? Read more about how we optimise vertical farming technology design through automation, modular design and more.
