Internet of Things (IoT): Concept and architecture
1. Definition
The Internet of Things (IoT) is a network of physical objects — "things" — embedded with sensors, software, actuators and network connectivity so that they can collect, exchange and act on data without human intervention. The term was coined by Kevin Ashton (1999). Its core idea is to bridge the physical world and the digital world: real-world parameters (temperature, motion, location, level) become data, and data becomes physical action (switch on, move, alert).
Examples: a smart watch counting steps, an air conditioner cooling a room because a sensor says it is hot, a factory machine raising a maintenance ticket before it breaks, a soil-moisture sensor ordering irrigation.
2. Enabling Technologies
| Technology | Role in IoT |
|---|---|
| Sensors | Convert physical quantities into electrical signals (temperature, humidity, pressure, motion/PIR, gas, light, current, GPS, camera, accelerometer) |
| Actuators | Convert electrical commands into physical action (relay, motor, valve, servo, buzzer, display) |
| Embedded controllers / MCU | Low-power processors (ESP32, STM32, Arduino, Raspberry Pi) that run the logic |
| Connectivity | Wi-Fi, Bluetooth/BLE, ZigBee, Z-Wave, LoRaWAN, NB-IoT, 4G/5G, Ethernet — long range vs low power trade-offs |
| Protocols | MQTT (lightweight publish/subscribe), CoAP, HTTP/REST, AMQP, ZigBee/Thread |
| Cloud & edge computing | Storage, device management, analytics, rules — increasingly pushed to the edge for latency and bandwidth |
| Data analytics and AI/ML | Turn raw telemetry into predictions and automation |
| IPv6 | Practically unlimited addresses, so every device can be addressed directly |
| Security infrastructure | Device identity, encryption, secure boot, certificates |
3. Characteristics of IoT Systems
Connectivity (any-time, any-place) · sensing and actuation · heterogeneity (many device types, vendors, protocols) · dynamic and self-configuring (devices join/leave, mesh routing) · enormous scale (billions of devices, thousands of data points per second) · intelligence (local or cloud analytics, autonomy) · resource constraints (battery, memory, CPU) · data-centric behaviour (the value is in the data stream) · security and privacy sensitivity (networked microphones, cameras, locks).
4. IoT Architecture — Four Layers
┌──────────────────────────────────────────────────────────────┐
│ 4. APPLICATION LAYER │
│ Smart home app, health monitoring, smart traffic, │
│ industrial dashboards, agriculture, logistics │
├──────────────────────────────────────────────────────────────┤
│ 3. MIDDLEWARE / PROCESSING LAYER │
│ Cloud/edge servers, databases, data analytics, AI models, │
│ device management, APIs, dashboards (the "brain") │
├──────────────────────────────────────────────────────────────┤
│ 2. NETWORK LAYER │
│ Transmission: Wi-Fi, BLE, ZigBee, LoRaWAN, NB-IoT, 4G/5G │
│ Protocols: MQTT, CoAP, HTTP — plus gateways that translate│
├──────────────────────────────────────────────────────────────┤
│ 1. PERCEPTION / SENSING LAYER │
│ Physical devices: sensors, actuators, cameras, meters, │
│ microcontrollers, embedded firmware │
└──────────────────────────────────────────────────────────────┘
A fifth (business) layer is sometimes added on top, covering business models, billing, decision-making and governance.
Sensors → network → cloud/edge → application → action is the whole loop; a gateway typically sits between the sensing and network layers, translating ZigBee/BLE into IP and buffering data when the link is down.
5. The IoT Ecosystem in Practice
Device (sensor + MCU + radio, often battery powered) → Gateway/hub (protocol translation, local rules, buffering) → Network (internet/cellular/LPWAN) → Cloud platform (device registry, ingestion, storage, analytics, ML, APIs, OTA firmware updates) → Application (mobile app, web dashboard, alerts) → User/actuator (the physical action). Edge computing moves some processing to the gateway or device to cut latency and cloud cost; a digital twin is a virtual model of the physical system kept in sync with real telemetry.
6. Working Example — a smart air conditioner
- A temperature and humidity sensor samples the room every minute.
- The MCU averages readings and publishes them over MQTT to the cloud through Wi-Fi.
- The cloud stores them, compares with the user's comfort setting and applies a rule or an ML model.
- A command is sent back (or executed locally at the gateway) to switch the compressor on/off, or to open a vent.
- The user sees the status in the app and can override; usage data feeds energy-saving recommendations.
7. Benefits and Challenges
| Benefits | Challenges |
|---|---|
| Automation and convenience; remote monitoring and control | Security — weak passwords, unpatched firmware, botnets (Mirai), data leakage |
| Efficiency — energy, water, fuel and maintenance savings | Privacy — always-on cameras/microphones, profiling, data ownership |
| Better decisions from real-time data and analytics | Interoperability — competing standards (Matter aims to fix this) |
| New services and business models; predictive maintenance | Power — battery life vs transmission range and data rate |
| Improved safety (gas leaks, machine faults, patient alerts) | Scale and data volume — storage, bandwidth, processing cost |
| Accessibility for elderly and disabled users | Connectivity gaps, cost of deployment, e-waste, regulations |
8. Exam Points
- Define IoT and explain its concept with two real examples — 5 marks, almost certain.
- Draw and explain the IoT architecture layers (perception, network, middleware, application) — 7 marks, very frequent.
- List the enabling technologies/characteristics of IoT — 5 marks.
- Explain the role of sensors, actuators, gateways and cloud platforms in IoT — 5 marks.
- What are the main challenges of IoT (security, privacy, power, interoperability)? — 5 marks.
- Name any four IoT communication technologies with a use case each — 3-5 marks.