This free curriculum covers every domain assessed in the Oxford Certified IoT Specialist examination: architecture, devices, connectivity, protocols, edge and cloud, device management, security, industrial IoT, data, privacy and business value. Read each chapter, note the key points, and then sit the 70-question examination.
An Internet of Things solution connects physical objects to digital systems. A typical architecture has layers: devices with sensors and actuators; connectivity, often through gateways that aggregate local devices and translate protocols; an edge or cloud platform that ingests, stores and processes data; and applications that present insights and send commands.
Telemetry is the measurement and status data devices send automatically; commands travel in the other direction. Designing an IoT solution means balancing power, connectivity, cost, latency, security and the value of the data.
Key takeaway: IoT value comes from closing the loop between the physical and digital worlds.
Chapter 2
Devices: sensors, actuators and power
Sensors measure the physical world — temperature, vibration, occupancy (for example passive infrared sensors), humidity, location. Actuators act on it — opening valves, switching motors, locking doors. Battery-powered nodes use low-power microcontrollers rather than general-purpose processors.
Battery life depends mainly on how often the device wakes, measures and transmits. Deep sleep between short active periods, lower sampling and transmission frequency, and efficient radios extend life to years. Devices should fail safely when they lose connectivity, keeping physical processes in a safe default state.
Sensors measure; actuators act
Low-power MCUs for battery devices
Deep sleep and fewer transmissions save energy
Calibrate sensors and design fail-safe behaviour
Key takeaway: Power and safety decisions made at the device level shape the whole solution.
Chapter 3
Connectivity options
Bluetooth Low Energy suits short-range, low-power links with phones and wearables. Wi-Fi offers high data rates but higher power use. Zigbee and Thread create low-power mesh networks in homes and buildings. Low-power wide-area networks (LPWAN) such as LoRaWAN and cellular NB-IoT and LTE-M provide long range and years of battery life for small, infrequent messages — ideal for meters and agriculture. 5G offers very low latency and supports massive numbers of devices for advanced use cases.
Choosing connectivity means weighing range, coverage at the site, power, data rate, latency, cost and the radio environment — metal buildings and basements can block signals, so pilots should test coverage and antenna placement.
BLE: short range, low power
Wi-Fi: high throughput, higher power
Zigbee/Thread: home and building mesh
LoRaWAN, NB-IoT, LTE-M: long range, low power, low data
Test coverage in the real environment
Key takeaway: There is no single best network — only the best fit for the use case.
Chapter 4
Messaging protocols: MQTT and CoAP
MQTT is a lightweight publish/subscribe protocol. Devices publish messages to topics on a broker; subscribers receive them. Quality of Service levels trade reliability for overhead: QoS 0 is at most once, QoS 1 at least once and QoS 2 exactly once. Retained messages give new subscribers the last known value immediately, and the Last Will and Testament notifies others when a device disconnects unexpectedly.
CoAP is designed for constrained devices, using a lightweight request/response model over UDP, similar in spirit to HTTP. Choose protocols according to device constraints, network reliability and the interaction pattern.
MQTT: publish/subscribe through a broker
QoS 0, 1, 2 — more reliability, more overhead
Retained messages and Last Will for state and presence
CoAP: request/response for constrained devices
Key takeaway: Pick the lightest protocol that still meets your reliability needs.
Chapter 5
Edge computing
Edge computing processes data close to where it is generated. It reduces latency for time-critical decisions — such as stopping a dangerous machine within milliseconds — cuts bandwidth and cloud costs by filtering, aggregating or downsampling data, and keeps local operations running during network outages. Safety-critical decisions must never depend on a distant cloud connection.
Edge devices and gateways are still part of the attack surface: they need secure configuration, monitoring and remote updates just like cloud services. A common pattern sends summary features (for example RMS and peak vibration) and anomalies to the cloud while keeping high-frequency raw data local.
Decide locally when latency or safety requires it
Filter and aggregate before sending
Keep operating offline
The edge still needs security and updates
Key takeaway: Process where it makes sense: the edge for speed and resilience, the cloud for scale and insight.
Chapter 6
Cloud platforms, data and digital twins
IoT platforms ingest telemetry, manage devices and trigger actions. Event-driven rules or serverless functions react to incoming data — for example alerting when a freezer temperature exceeds a threshold. Time-series databases store high-volume timestamped measurements efficiently, and accurate time synchronisation lets events from many devices be ordered and correlated.
A digital twin is a virtual representation of a physical asset updated with real-time data; device shadows store desired and reported state so commands can be applied when an offline device reconnects. Reliable analytics require validated, timestamped data: detect outliers and impossible values, handle gaps, use redundant sensors for critical measurements and calibrate regularly. Predictive maintenance uses patterns such as vibration and temperature to anticipate failures.
Event-driven rules for real-time reactions
Time-series storage with accurate timestamps
Digital twins and device shadows
Validate and clean sensor data
Predictive maintenance reduces unplanned downtime
Key takeaway: Data is only valuable if it is trustworthy and acted upon.
Chapter 7
Device management at scale
Provisioning securely registers each device with a unique identity and credentials. Fleet management groups devices, configures them remotely, monitors health and firmware versions and performs bulk operations. Over-the-air updates deliver fixes and security patches: firmware must be signed and verified, updates staged to small groups first, and rollback available. If a pilot group fails after an update, halt, roll back and investigate.
At end of life, decommission securely by wiping credentials and data. Manufacturers remain responsible for security support during the declared support period.
Unique identity per device at provisioning
Fleet management instead of manual access
Signed, staged OTA updates with rollback
Secure decommissioning
Key takeaway: If you cannot update a device safely, you cannot keep it secure.
Chapter 8
IoT security
Many IoT attacks exploit weak basics. The Mirai botnet hijacked devices using default passwords, so unique credentials per device are essential. Each device should have its own cryptographic identity, such as an X.509 certificate, stored in a secure element, so it can be authenticated individually and revoked if compromised. Secure boot ensures only authentic firmware runs. Encrypt communication with TLS even on private networks, segment IoT devices away from critical systems, protect against physical tampering and monitor for unusual behaviour, failed authentication and outdated firmware.
Consumer IoT baselines such as ETSI EN 303 645 and the UK PSTI regime require no universal default passwords, a published vulnerability disclosure route and transparency about the security-update support period. Respond to researcher reports through a coordinated disclosure process.
No default or shared passwords
Unique certificates in secure hardware
Secure boot and signed firmware
TLS everywhere and network segmentation
Vulnerability disclosure and defined support periods
Key takeaway: Security must be designed into the device, the network and the lifecycle.
Chapter 9
Industrial IoT and operational technology
Operational technology monitors and controls physical processes using PLCs, SCADA systems and industrial networks. Safety and availability are often prioritised even above confidentiality. Legacy protocols such as Modbus lack built-in authentication and encryption; OPC UA provides secure, platform-independent data exchange.
The Purdue model organises industrial networks into levels, separating control systems from enterprise IT. Remote access for suppliers should pass through a demilitarised zone with strong authentication and monitoring — PLCs must never be exposed directly to the internet.
OT controls physical processes; safety comes first
Modbus is insecure by design; OPC UA adds security
Separate OT and IT using the Purdue model
Secure, monitored remote access through a DMZ
Key takeaway: In industrial IoT, a cyber incident can become a physical safety incident.
Chapter 10
Privacy in IoT
Smart devices often collect personal data — occupancy, location, voice, health. Data protection laws apply regardless of whether data comes from machines. Collect only what is necessary, explain data use clearly, obtain consent where required and give users control. Detailed occupancy patterns reveal when people are at home and must not be shared with advertisers without a lawful basis and clear consent.
Privacy by design means considering privacy from the first product decision: processing data locally where possible, setting privacy-friendly defaults, protecting stored data and deleting it when no longer needed. Clear, accessible privacy information builds the trust that connected products depend on.
Device data is often personal data
Minimise collection and be transparent
Consent and user control for sensitive uses
Avoid always-on recording without clear notice
Key takeaway: Respecting privacy is essential to customer trust in connected products.
Chapter 11
Interoperability and the business case
Standards such as Matter help devices from different manufacturers work together reliably and securely. A strong business case compares benefits — for example reduced unplanned downtime and maintenance costs — with the total cost of ownership: devices, connectivity, installation, cloud, maintenance and decommissioning.
Start with a pilot to validate connectivity, data quality, operations and value at small scale before full rollout.
Interoperability reduces lock-in and integration effort
Measure value in business outcomes
Include full lifecycle costs
Pilot before scaling
Key takeaway: Successful IoT projects prove value small, then scale with confidence.
Ready? Take the free exam
Take the exam and earn your digital badge free of charge. After passing, you may order the official certificate.