LoRaWAN, NB-IoT or BLE: Choosing a Radio for a Sensor Fleet
Three radios that look interchangeable on a datasheet and are not. How range, power budget, message size and who owns the network decide which one a deployment can actually live with.
A sensor fleet's radio is the one decision that is genuinely hard to reverse. Firmware can be rewritten and a backend can be replaced, but the radio is soldered to a board that is already installed on a wall, a pole, or a pipe. Choosing it on familiarity rather than on constraints is the most expensive habit in the field.
LoRaWAN trades data rate for reach. A device can be kilometres from its gateway, run for years on a primary cell, and send a few dozen bytes at a time. What it cannot do is send much, or often: duty cycle limits in most regions cap how long a device may transmit per hour, and a payload is measured in tens of bytes rather than kilobytes. This suits anything that reports a reading periodically and never needs to be commanded urgently. It suits nothing that streams.
NB-IoT inverts the ownership question. There is no gateway to install because the network already exists, which removes an entire class of deployment work and an entire class of things to maintain. The cost is a subscription per device and a dependency on an operator's coverage and lifecycle, which is a commercial relationship rather than a technical parameter. In a basement or a metal enclosure, coverage is also a question best answered by a site survey than by a coverage map.
BLE is the odd one out, because its range assumption is different in kind rather than in degree. It is a personal-area radio, and a BLE sensor fleet is really a fleet of devices near something else that has connectivity — a phone, a gateway in the same room, a vehicle. Where that assumption holds, BLE is cheap, low-power and ubiquitous. Where it does not, no amount of tuning will make it reach.
The question that resolves most of these arguments is not about the radios at all: who owns the network. LoRaWAN means you run gateways, which means you own coverage, backhaul and the cost of both, and you can put a gateway exactly where the deployment needs one. NB-IoT means an operator owns it, which is less work and less control. BLE means whatever the device is near owns it. Deployments fail far more often on that ownership question than on the radio's specifications.
The practical procedure is to write down the message size, the reporting interval, the battery life target and the worst-case installation location before comparing any datasheets. Those four numbers eliminate at least one option immediately and usually two, and they are specific to the deployment in a way that no comparison table can be.