Open this lesson in your favourite AI. It'll walk you through the why, explain the demo, and quiz you on the try-it list.
The phrase 'Internet of Things' is misused to mean everything from a Wi-Fi lightbulb to industrial SCADA. Before you build anything, you need a precise mental model: a device that senses or acts on the physical world, a gateway/network path that aggregates and routes its data, and a cloud (or backend) layer that stores, analyzes, and exposes that data to humans and other systems. Without this three-tier picture you'll over-engineer trivial cases and under-architect the hard ones — sticking a cellular modem on a $5 sensor that only ever talks to a hub three metres away, or shipping a 'cloud-native' security camera that drops every clip the moment Wi-Fi blips.
The canonical three-tier IoT architecture.
Use these three in order. Each builds on the one before.
In one paragraph, explain the three tiers of an IoT system and why most products fit into this model.
Walk me through what happens, step by step, from the moment a temperature sensor reads 22.3 °C to the moment that value appears on a dashboard 600 km away.
Given a smart agriculture deployment with 10,000 soil-moisture sensors across 50 km², how would you decide between a star topology (each sensor talks cellular directly) versus a mesh with LoRa gateways every 2 km, and what cost/reliability/maintenance trade-offs would drive the call?
[ DEVICE ] [ GATEWAY ] [ CLOUD ]
───────── ────────── ─────────
Sensors ──┐ ┌── Aggregates messages from ┌── Stores telemetry
Actuators ─┤ Reads / writes │ many devices │ (time-series DB)
MCU/SoC ─┤ the physical world │ │
Radio ─┤ (1 Hz to 10 kHz) │── Translates protocol │── Exposes APIs
Battery ─┘ │ (BLE/Zigbee → MQTT/HTTPS) │ (REST / GraphQL)
│ │
│ │── Buffers when uplink down │── Runs rules,
│ (BLE / Zigbee / │ │ ML, dashboards
│ LoRa / Wi-Fi) │── Enforces local policy │
▼ │ (LAN-only commands, │── Pushes commands
│ privacy filters) │ back to devices
│ │
└── Optional: edge inference │── User app / web
└── (dashboards,
mobile apps)
COLLAPSED ARCHITECTURES (common in practice):
• A Wi-Fi smart plug talks MQTT directly to the cloud — the home router is the
gateway in the network sense, but the plug is its own "gateway" in the protocol
sense.
• An industrial PLC reads 200 modbus sensors and uplinks one summary record — the
device IS the gateway.
• A BLE-only fitness tracker uses your phone as a transient gateway whenever you're
near it; offline storage on the device must bridge the gaps.