Internet of Things (IT-703 (B)) - Unit 5 Short Notes

UNIT 5: Internet of Things - Comprehensive Study Notes

Based on RGPV past paper analysis (May 2022 - Jun 2025). Focus on definitions, comparisons, protocols, and architectures.

I. IoT Fundamentals & Architectural Framework

IoT Ecosystem & Conceptual Framework

• Definition: IoT is a network of physical objects ("things") embedded with sensors, software, and connectivity to exchange data with other devices/systems over the internet.

• Core Characteristics (8 key traits):

  1. Connectivity: Seamless communication.

  2. Things: Physical/virtual objects with unique IDs.

  3. Data: Raw data from sensors becomes actionable information.

  4. Communication: Various protocols/wireless standards.

  5. Intelligence: Data processing & decision-making (often via analytics/cloud).

  6. Actionability: Results trigger physical/digital actions.

  7. Services: IoT as a service platform (e.g., monitoring, automation).

  8. Semantics: Meaningful interpretation of data (context-awareness).

IoT Ecosystem Components:

• Things/Devices: Sensors, actuators, embedded systems.

• Communication: Networks & protocols (LAN, WAN, LPWAN).

• Data Processing & Storage: Edge computing, cloud platforms.

• Applications & Analytics: User interfaces, business logic, data analysis.

• Management & Security: Device management, security models, privacy policies.

IoT Reference Architecture & Information Model

A common 3-layer or 5-layer model:

1. Perception Layer (Device Layer): Sensors/actuators for data collection/action.

2. Network Layer (Gateway/Transport Layer): Data transport via networks (Wi-Fi, ZigBee, cellular). IoT Gateways aggregate data, protocol translation, security.

3. Application Layer: User-facing apps, analytics, specific IoT services.

4. Middleware/Service Layer (in 5-layer): Service discovery, data management, SOA.

5. Business Layer (in 5-layer): Overall management, business models.

Information Model: Defines how data is structured, represented, and exchanged (e.g., using standards like oneM2M, LwM2M). It ensures interoperability.

IoT Service-Oriented Architecture (SOA) & Challenges

• SOA in IoT: Treats device functions as reusable "services" (e.g., "read temperature", "turn on light"). Enables loose coupling, interoperability, and composition of complex services from simple ones.

• Challenges:

  • Resource Constraints: Limited power, memory, CPU on devices.

  • Heterogeneity: Diverse devices, protocols, data formats.

  • Scalability: Millions of devices.

  • Real-time Requirements: Latency-sensitive applications.

  • Security & Privacy: Increased attack surface.

Design Perspectives

• Logical vs. Physical Design:

  • Logical Design: What the system does. Defines functions, services, data flows, and interactions (abstract). e.g., "System shall monitor temperature."

  • Physical Design: How it's built. Specifies hardware (sensors, microcontrollers), communication modules, physical layout, and power sources. e.g., "Use DHT22 sensor with ESP32 via Wi-Fi."

• IoT Level-Based Systems:

  | Feature | Level 3 System | Level 4 System | | :--- | :--- | :--- | | Connectivity | Single device → single app (point-to-point). | Device → Gateway → Cloud → Multiple Apps (networked). | | Example | Smart bulb controlled by one phone app. | Smart home system (bulb, thermostat, camera) all managed via cloud platform. | | Complexity | Low. | High. | | Scalability | Poor. | Excellent. | | Data Flow | Direct. | Via intermediate layers (gateway, cloud). |

Evolution & Context: M2M vs IoT vs WoT

Reasons for shift from M2M to IoT: Need for scalability, integration with enterprise systems, big data analytics, cloud computing, and standardized IP-based communication.

II. Device Layer: Sensors, Actuators & Hardware

Sensors

• Sensor Node: A complete unit containing sensor, microcontroller, communication module, and power source.

• Key Features: Sensitivity, accuracy, precision, range, resolution, response time, stability, power consumption.

• Types:

  • By Signal: Scalar (measures magnitude, e.g., temperature sensor), Vector (measures magnitude & direction, e.g., accelerometer, magnetometer).

  • By Output: Analog (continuous voltage/current, needs ADC), Digital (discrete binary output, e.g., I2C/SPI sensors).

• Common IoT Sensors: Temperature/Humidity (DHT11/22), PIR (motion), Gas (MQ series), Pressure (BMP280), Proximity (Ultrasonic), Image (Camera), GPS.

Sensor Characteristics & Errors (Frequently Tested)

Actuators

• Role: Convert electrical/control signals into physical action (opposite of sensor).

• Four Common Selection Characteristics:

  1. Force/Torque Output: Required mechanical strength.

  2. Stroke/Displacement: Distance/movement needed.

  3. Speed/Response Time: How fast it must act.

  4. Power Source & Efficiency: Voltage, current, energy consumption.

• Types:

  • Mechanical: Electric motors (DC, stepper), solenoids, relays.

  • Soft: Made of flexible materials (silicone) for gentle handling.

  • Shape Memory Polymer (SMP): Change shape with temperature/light stimulus.

Hardware Platforms & Interfacing

• Microcontrollers (MCU): e.g., Arduino (ATmega328), ESP32/8266. Low-cost, low-power, good for dedicated sensor/control tasks. Use GPIO, SPI, I2C, UART for interfacing.

• Single-Board Computers (SBC): e.g., Raspberry Pi (Broadcom SoC). Runs full OS (Linux), more powerful, for complex processing/gateway roles.

  • Raspberry Pi vs Desktop: Pi is ARM-based, lower power, no internal storage (uses SD), less RAM/CPU, designed for embedded/IoT, not general computing.

• Interfacing:

  • GPIO (General Purpose Input/Output): Digital pins for simple on/off or communication.

  • I2C (Inter-Integrated Circuit): 2-wire (SDA, SCL) serial bus for multiple low-speed devices (sensors, EEPROM).

  • SPI (Serial Peripheral Interface): 4-wire (MOSI, MISO, SCLK, CS) faster synchronous serial for high-speed devices (displays, SD cards).

• Pneumatic: Use of compressed air/gas to generate motion/force (e.g., pneumatic actuators/cylinders). An example of an actuator type.

III. Communication & Networking Protocols

Wireless Communication Technologies

Network Types by Topology/Connection:

• Physical Topology: Star (all to hub/AP), Mesh (multi-hop), Tree (hierarchical), Bus (single backbone).

• Connection Type: PAN (Personal), LAN (Local), WAN (Wide), LPWAN (Low-Power WAN, e.g., LoRaWAN, NB-IoT).

IP & Adaptation Layer: 6LoWPAN

• Role: Enables IPv6 packets to be carried efficiently over IEEE 802.15.4 networks (which have small MTU ~127 bytes).

• How: Header Compression (compresses 40-byte IPv6 header to ~1-2 bytes), Fragmentation & Reassembly (splits IPv6 packet to fit 802.15.4 frame).

• Difference from IPv4/IPv6: It's not a new IP version. It's an adaptation layer between IPv6 and 802.15.4 MAC. IPv4 has no such adaptation for constrained networks; native IPv6 is too heavy for 802.15.4.

Application Layer Protocols (Very High Frequency)

Differentiation: AMQP vs MQTT

• AMQP: More complex, enterprise-oriented, guaranteed delivery, rich routing (exchanges), standard for financial/business systems.

• MQTT: Extremely lightweight, simple, minimal overhead, ideal for constrained devices, simple pub/sub.

Other Enablers & Technologies

• RFID (Radio Frequency Identification):

  • Principle: Uses radio waves to automatically identify and track tags attached to objects.

  • Components: Tag (passive/active, with ID), Reader, Antenna, Backend System.

  • Link to IoT: Provides unique digital identity to physical objects, enabling them to be "things" in IoT. Used in inventory, access control, supply chain.

• NFC (Near Field Communication):

  • Principle: Short-range (≤10 cm) wireless tech. based on RFID standards. Enables two-way communication.

  • Use in IoT: Device pairing (e.g., phone to IoT device), contactless payments, data exchange (touch to configure), access control.

• IoT Enablers: Broad category including Cloud Computing, Big Data Analytics, AI/ML, Mobile Computing, IPv6, M2M Platforms, Semantic Technologies that make IoT systems viable and scalable.

IV. IoT Platforms, Cloud & Data

Cloud Computing for IoT

• Usefulness:

  • Storage: Massive, scalable storage for time-series sensor data.

  • Processing: On-demand compute for analytics, ML, batch/stream processing.

  • Device Management: Provisioning, monitoring, updating fleets of devices.

  • Scalability & Cost: Pay-as-you-go, no upfront infrastructure cost.

• Cloud Service Models:

  • IaaS (Infrastructure as a Service): VMs, storage, networks (e.g., AWS EC2). User manages OS/apps.

  • PaaS (Platform as a Service): Runtime environment, DBs, dev tools (e.g., AWS IoT Core, Azure IoT Hub). User focuses on app/data.

  • SaaS (Software as a Service): Ready-to-use applications (e.g., Salesforce). User just uses software.

• Cloud Storage Models:

  • Object Storage (e.g., AWS S3): Unstructured data (images, logs).

  • Block Storage (e.g., EBS): Raw storage volumes for VMs.

  • File Storage (e.g., EFS): Shared file system.

• Cloud Communication APIs: RESTful APIs provided by cloud IoT platforms (AWS IoT, Azure IoT) for:

  • Device registration & authentication.

  • Sending telemetry data (publish).

  • Receiving commands (subscribe).

  • Device shadow/state management.

IoT Platforms

• What: Integrated software/hardware suites that provide tools for device management, data ingestion, analytics, and application development.

• How they facilitate:

  • Device Connectivity & Management: SDKs, protocols (MQTT/CoAP), lifecycle management.

  • Data Ingestion & Storage: Scalable pipelines.

  • Analytics & Visualization: Built-in dashboards, ML tools.

  • Application Enablement: APIs, rules engines, integration with other services.

  • Examples: AWS IoT, Microsoft Azure IoT, Google Cloud IoT Core, IBM Watson IoT, open-source (ThingsBoard, Kaa).

Data Analytics in IoT

• Role: Transform raw sensor data into actionable insights, predictions, and automated decisions.

• Process: Data Ingestion → Storage → Processing (batch/stream) → Analysis (descriptive, predictive, prescriptive) → Visualization/Actuation.

• How differs in M2M vs IoT:

  • M2M: Analytics is often local, simple, and reactive (e.g., threshold alert). Data volume low, siloed.

  • IoT: Analytics is centralized, complex, predictive/prescriptive. Uses big data tools, ML/AI on massive, diverse, real-time streams for optimization, forecasting, anomaly detection.

• Extracting Insights: Through statistical analysis, machine learning (regression, classification, clustering), stream processing (Apache Flink, Spark Streaming), and visualization (dashboards, alerts).

V. Security & Privacy in IoT

Need for Security & Models

• Why Required?

  • Physical Vulnerability: Devices in public/unattended locations.

  • Resource Constraints: Hard to implement strong crypto.

  • Large Attack Surface: Millions of devices.

  • Critical Impact: Attacks on medical devices, infrastructure.

  • Privacy: Sensitive user data (location, habits) collection.

• Security Models (Frameworks/Approaches):

  • CIA Triad Adaptation: Confidentiality (encryption), Integrity (hashing, signatures), Availability (DDoS mitigation).

  • Zero Trust Model: "Never trust, always verify." Continuous authentication/authorization.

  • End-to-End Security: Security at device, network, cloud, and application layers.

  • Identity & Access Management (IAM): Device identity (certificates, tokens), least privilege access.

  • Security by Design: Integrating security from initial design phase.

Vulnerabilities & Attacks

• Kinds of Vulnerabilities:

  • Hardware: Physical tampering, side-channel attacks.

  • Software/Firmware: Unpatched bugs, weak/default passwords, insecure APIs.

  • Network: Unencrypted traffic, weak protocols, open ports.

  • Privacy: Excessive data collection, lack of anonymization.

• Attacks Exploiting Application/Service Layer:

  • Injection Attacks: SQL, command injection via device APIs.

  • Broken Authentication: Weak credentials, session hijacking.

  • Sensitive Data Exposure: Lack of encryption in transit/at rest.

  • XML External Entities (XXE): If using XML (like XMPP).

  • Denial-of-Service (DoS/DDoS): Overwhelming device or cloud service.

  • Man-in-the-Middle (MitM): Intercepting/altering communication between device and cloud/app.

• Various Attacks on IoT Systems:

  • Device Hijacking: Taking control of device (botnet for DDoS).

  • Data Theft: Stealing collected data.

  • Privacy Breach: Tracking user via device data.

  • Physical Damage: Malicious commands to actuators (e.g., industrial machinery).

  • Replay Attacks: Re-sending valid captured messages.

VI. Applications & Case Studies

Smart Home / Home Automation

• Applications: Lighting control, HVAC, security (cameras, locks), entertainment, appliance monitoring, energy management.

• Design with Raspberry Pi & Hardware (Conceptual Sketch Description):

  • Central Hub/Gateway: Raspberry Pi 4 running Home Assistant/OpenHAB.

  • Sensors: DHT22 (temp/humidity), PIR (motion), MQ-2 (gas), door/window magnetic sensors.

  • Actuators: Relay modules (for lights/fans), smart plugs, servo motors (for locks).

  • Communication:

    • Short-range: ZigBee/Z-Wave modules (USB dongle on Pi) for low-power sensors.

    • Wi-Fi: For Wi-Fi devices (smart plugs, cameras) and Pi's internet connection.

    • Bluetooth: For BLE beacons/sensors.

  • Power: 5V/3A adapter for Pi, separate power for high-current actuators.

  • Cloud Integration: Pi connects to cloud (e.g., AWS IoT) for remote access/analytics.

  • User Interface: Mobile app/Web dashboard served by Pi.

Case Study: Smart Agriculture

• Objective: Optimize water usage, monitor crop health, increase yield.

• IoT Implementation:

  • Sensors: Soil moisture, temperature/humidity, pH, light sensors deployed in fields.

  • Actuators: Automated irrigation valves, fertilizer dispensers.

  • Network: LPWAN (LoRaWAN) for long-range, low-power sensor data to gateway. Gateway uses cellular (4G/5G) to cloud.

  • Platform: Cloud IoT platform (e.g., Azure IoT) ingests data.

  • Analytics: ML models predict irrigation needs based on weather forecasts + soil data. Alerts for disease detection (from image sensors).

  • Actuation: Automated irrigation schedules based on analytics.

  • Benefits: 20-30% water saving, early pest detection, reduced labor.

Other Mentioned Topics

• WebSockets: Full-duplex communication protocol over a single TCP connection. Used in IoT for real-time web dashboards (push updates from server to browser without polling). MQTT can be tunneled over WebSockets for browser-based clients.

• Software Defined Networking (SDN): Separates control plane (centralized controller) from data plane (switches). Maturity: Maturing technology. Used in IoT for dynamic network management, traffic optimization, and security policy enforcement across large-scale IoT deployments. Not yet ubiquitous in small IoT but promising for industrial IoT.

• Context Awareness: System's ability to adapt based on situational information (location, time, user activity, environment). Enabled by sensor fusion.

• Reality Mining: Collecting and analyzing data from device usage (phones, wearables) to understand human behavior, social patterns, and mobility.

Exam Tips & Common Pitfalls

[!TIP]

• Protocols: Be very clear on CoAP (request/response, UDP, CON/NON) vs MQTT (pub/sub, TCP, QoS). Know AMQP's exchange/queue model.

• Architecture: Distinguish Logical vs Physical and Level 3 vs Level 4 with examples.

• Sensors: Memorize Bias, Drift, Hysteresis, Quantization Error definitions.

• Security: Link vulnerabilities to specific attack types (e.g., weak password → broken authentication).

• 6LoWPAN: Emphasize it's an adaptation layer, not a new IP. Its job is header compression & fragmentation for 802.15.4.

• Smart Home Sketch: Label Pi (gateway), sensors/actuators, communication tech (Wi-Fi/ZigBee), and cloud link. Even a rough labeled diagram scores marks.

• M2M vs IoT: Focus on connectivity (siloed vs IP), data (limited vs big), and intelligence (none vs analytics).