[!NOTE]
How to use these notes. Read each section. Then do the Exam Focus box at the end of it. Small superscript numbers point to the Footnotes section at the end, where every hard word has one short meaning.
Items marked ★ appeared in past NPTEL Week 1 assignments, or are the kind of fact an MCQ tests.

1. What is IoT?

Definition (simple). IoT connects devices, machines and tools to the internet, mostly by wireless links.^{1Wireless technologies — links without cables, such as Wi-Fi, Bluetooth, ZigBee and cellular.}

Definition (Gartner). IoT is the network of physical objects that contain embedded technology^{2Embedded technology / embedded system — a small computer built inside a device to do one job.} to communicate and to sense or interact with their internal state or the external environment.^{3Gartner — a research and advisory company that publishes IT definitions.}

Key facts

• Over 9 billion things were online at the time of the slides. Experts expected more than 20 billion soon.
• IoT is a union of technologies: low-power embedded systems, cloud computing, big data, machine learning and networking. ★
• Full form: Internet of Things. ★

Origin of the term

• The ITU^{4ITU — International Telecommunication Union, the United Nations agency for telecom standards.} used the term in its 2005 report, titled "The Internet of Things".
• The report predicted that people would become the minority of internet traffic generators. Most traffic would flow between devices and "things".
• It discussed M2M^{5M2M (Machine-to-Machine) — machines exchange data directly, usually over telecom networks, with little human input.} connectivity that reaches common household devices.
• IoT enablers named in the report ★: RFID,^{6RFID (Radio-Frequency Identification) — a tag that a reader identifies by radio waves.} Nanotechnology,^{7Nanotechnology — engineering at the scale of billionths of a metre; makes very small sensors and devices.} Sensors, Smart Networks.

Practice — this year's assignment

Characteristics of IoT ★

1. Efficient, scalable and associated architecture.
2. Unambiguous naming and addressing. (Not "ambiguous". This is a common trick option.)
3. An abundance of sleeping nodes,^{8Sleeping node — a device that stays in low-power sleep most of the time to save battery.} mobile devices and non-IP devices.^{9Non-IP device — a device that does not speak the Internet Protocol directly (for example, a simple ZigBee or RFID device).}
4. Intermittent connectivity.^{10Intermittent connectivity — the link goes on and off, so devices are not always reachable.}

Practice — this year's assignment

Application areas (from the slides)

Modern applications (list to recognise): smart parking, structural health, noise maps, traffic congestion, smart lighting, waste management, smart roads, smart grid, river floods, tank level, photovoltaic installations, water flow, forest-fire detection, air pollution, snow-level monitoring, landslide and avalanche prevention, earthquake early detection, water leakage, radiation levels, explosive and hazardous gases, supply chain control, NFC^{11NFC (Near-Field Communication) — very short-range wireless used for tap-to-pay.} payment, intelligent shopping, perimeter access control.

IoT is used in agriculture for automation and data monitoring. (True. This appeared in a past assignment.) ★

Evolution of connected devices (dates and firsts)

Practice — this year's assignment

Growth model on the slides ("Expected"): Trillions of sensors → billions of smart systems → millions of applications → IoT.

IoT enablers (image slide)

• Implementation: transport, agriculture, vehicles, cities, energy, water, finance, homes, industry, healthcare.
• Connectivity technologies: ZigBee, RFID, Bluetooth, cellular masts, Wi-Fi, LoRa, LoWPAN, and similar.
• Enabling technologies: big data, networks, AI, deep learning, cloud.

2. Baseline technologies: IoT vs M2M vs WoT vs CPS

The slides name three technologies closely related to IoT: M2M, CPS^{14CPS (Cyber-Physical System) — a system in which computing and physical processes control each other closely.} and WoT.^{15WoT (Web of Things) — IoT devices exposed through standard web tools and web addresses.}

IoT vs M2M ★

Practice — this year's assignment

IoT vs WoT

• WoT gives access to and control of IoT resources through mainstream web technology: HTML 5.0, JavaScript, Ajax,^{16Ajax — a web method that updates part of a page without reloading it.} PHP, Ruby on Rails.
• WoT rests on RESTful principles and REST APIs.^{17REST / RESTful API — a simple web style in which each resource has an address and is used with HTTP calls (GET, POST, and so on).}
• WoT faces scalability and security challenges on the way to a global WoT.
• IoT builds a network of objects, people, systems and applications. WoT integrates them into the Web.
• WoT is a flavour of application layer placed on top of the IoT network layer.
• IoT is broader. It includes systems the web cannot reach, such as conventional WSN^{18WSN (Wireless Sensor Network) — many sensor nodes that pass data to a base station by radio.} and RFID systems.

Practice — this year's assignment

Memory hook: M2M ⊂ IoT ⊃ WoT. (IoT is the largest circle.)

3. IoT addressing

3.1 The address crunch ★

• Experts estimated 20–50 billion devices by 2018.
• The cause: existing devices, smart devices and constrained nodes^{19Constrained node — a device with very little memory, processing power or energy.} all join one framework.
• Connectivity types mix together: cellular, Wi-Fi, Ethernet, plus newer ones such as BLE,^{20BLE (Bluetooth Low Energy) — a Bluetooth version made for very low power use.} DASH7,^{21DASH7 — a long-range, low-power wireless standard for sensor networks.} Insteon^{22Insteon — a home-automation networking technology (power line and radio).} and IEEE 802.15.4.^{23IEEE 802.15.4 — the standard for low-rate, low-power wireless personal area networks. ZigBee and 6LoWPAN use it.}
• The ITU vision is now true: networked devices outnumber humans on earth.
• True statement: More IoT devices → address crunch.^{24Address crunch — a shortage of free IP addresses.}

3.2 Connectivity terms ★

3.3 How gateways save addresses ★

• Nodes get local addresses that are valid only inside the gateway's domain. Two gateways can reuse the same local addresses.
• Each gateway has a unique network prefix.^{25Network prefix — the first part of an IP address. It names the network, not the device.} This prefix identifies the gateway globally. ★
• The gateway gives each field device or sensor in its area a unique (local) address. ★
• Result: less address waste. Cost: nodes must reach the Internet through the gateway.
• Routers hand out prefixes to the gateways below them.

Practice — this year's assignment

Address labels on the slides

• L = link-local address.^{26Link-local address — valid only on one network link; routers do not forward it.}
• LU = local address that is unique locally.
• ULA = unique local address.^{27ULA (Unique Local Address) — an IPv6 private address for use inside one site, not on the public Internet.}

3.4 Mobility and addressing

• When a gateway moves, its network prefix changes (for example, prefix 1 becomes prefix 2). The IoT LAN addresses stay the same. ULA / LU addresses give this stability.
• Gateways attach to a remote anchor point^{28Remote anchor point — a fixed point in the network that keeps a stable address for a moving node or gateway.} through protocols such as Mobile IPv6.^{29Mobile IPv6 — a protocol that lets a device keep the same IPv6 address as it moves between networks.} Prefix changes then do not affect them.
• Sometimes a node must reach the Internet directly. It uses tunneling^{30Tunneling — wrapping one packet inside another so it can cross a network that does not support the inner packet.} to a remote anchor point. IKEv2^{31IKEv2 (Internet Key Exchange version 2) — a protocol that sets up secure keys for IPSec tunnels.} is an example tunneling protocol.

3.5 What gateways do

• Main jobs: Internet connectivity and IoT LAN intra-connectivity.^{32Intra-connectivity — connection inside the same network.}
• Upstream prefixes come from DHCPv6.^{33DHCPv6 — a server-based method that hands out IPv6 addresses and prefixes.} The gateway passes them to nodes using SLAAC.^{34SLAAC (Stateless Address Auto-Configuration) — a device builds its own IPv6 address from a network prefix, with no server that tracks addresses.} (SLAAC = stateless addressing.)
• LU addresses stay separate from globally routable addresses. This helps when internal address stability matters most.
• Problem: an LU address cannot talk directly to the Internet or upper layers. Fix: an application-layer proxy.
• A proxy can also process data, not only pass it on.
• For nodes that cannot do heavy computing, the proxy collects data sent to the link-local multicast address and routes it globally.

Practice — this year's assignment

3.6 IPv4 to IPv6 transition

• The Internet is still mainly IPv4. IPv6 support is small.
• No universal transition fix exists. So people use makeshift solutions:
  1. IPv6-to-IPv4 translation
  2. IPv6 tunneling over IPv4
  3. Application-layer proxies (for example, data relaying)

3.7 Multi-homing ★

• Multi-homing: a node or network connects to multiple networks for better reliability.^{35Multi-homing — one node or network with more than one connection to the outside.}
• Small IoT LANs cannot get prefix allotments. They use a proxy that maps multiple IP addresses to link-local addresses.
• Other approach: a gateway assigns link-local addresses to its nodes.
• Real challenge: give the multi-homed node the right source address, destination address and routing information.
• If source and destination addresses come from the same prefix, use routing between gateways for gateway selection.
• The IETF is still working to standardize this.

3.7a IPv4 vs IPv6 ★★ (high-yield table)

The slide prints "IEF 1998" for IPv6. This is a typo for IETF. The slide also says IPv4 has no multicast. In real networks IPv4 can multicast, but write the slide answer in the exam.

Practice — this year's assignment

IPv4 header fields: Version, IHL,^{37IHL (Internet Header Length) — the length of the IPv4 header.} Type of Service, Total Length, Identification, Flags, Fragment Offset,^{38Fragment offset — says where a piece of a split packet belongs in the whole packet.} Time to Live (TTL),^{39TTL (Time to Live) — a counter that limits how long a packet lives. Each router reduces it. (IPv6 calls it Hop Limit.)} Protocol, Header Checksum,^{40Checksum — a small number used to detect errors in data.} Source Address (32 bit), Destination Address (32 bit), Options, Padding.
→ IPv4 stresses reliable transmission (fields such as ToS, total length, ID, offset, TTL, checksum).

IPv6 header fields: Version, Traffic Class, Flow Label, Payload Length, Next Header, Hop Limit, Source Address (128 bit), Destination Address (128 bit).
→ IPv6 has a simpler header and focuses on addressing, not on reliability of delivery. ★
→ The Source Address field in the IPv6 header is 128 bits. ★

Extra (not on the slides, but true): The IPv4 base header is 20 bytes (minimum). The IPv6 header is a fixed 40 bytes.

4. Sensing

4.1 Definitions

• Sensor: detects (senses) changes in ambient conditions or in the state of another device or system. It then forwards or processes that information.
• Oxford: "A device which detects or measures a physical property and records, indicates, or otherwise responds to it."
• A sensor performs an input function. It senses physical changes in a system in response to a stimulus.^{41Stimulus — a change in the environment (heat, light, pressure) that a sensor detects.}
• Examples: a temperature sensor turns heat into an electrical signal. A barometer^{42Barometer — an instrument that measures air pressure.} turns atmospheric pressure into an electrical signal.

4.2 Transducer vs sensor vs actuator ★

• Transducer: converts one form of energy into another.^{43Transducer — a device that converts one type of energy into another.}
• The word "transducer" is the collective term for sensors (sense many energy forms: movement, electrical, radiant, thermal, magnetic) and actuators (switch voltages or currents).
• Sound system example: microphone (input device) turns sound into an electrical signal → amplifier amplifies it → loudspeaker (output device) turns it back into sound.

Hook: Sensor = input. Actuator = output. Transducer = both (umbrella term).

4.3 Features of a good sensor ★

1. It is sensitive only to the property it measures.
2. It is insensitive to every other property in its setting (a temperature sensor ignores light and pressure).
3. It does not change the property it measures.

Practice — this year's assignment

4.4 Resolution ★

• Resolution = the smallest change a sensor can detect in the quantity it measures.
• For a digital-output sensor, resolution = the smallest step of its digital output.
• Higher resolution → better precision.
• Accuracy does NOT depend on resolution. (Precise ≠ accurate.)

Practice — this year's assignment

4.5 Sensor classes

By output: Analog · Digital
By data type: Scalar · Vector / Multimedia

Practice — this year's assignment

4.6 Sensor types (by measured quantity) ★

Pictured on the slides: pressure sensor, ultrasonic distance sensor, tilt sensor, infrared motion sensor, analog temperature sensor, camera sensor.

4.7 Sensor errors and deviations ★★

Practice — this year's assignment

5. Actuation

5.1 Definition ★

• An actuator is a part of a machine or system that moves or controls the mechanism or system.
• It is the mechanism by which a control system acts on an environment.
• It needs two things: a control signal and a source of energy. ★
• On a control signal, it converts energy into mechanical motion.
• The control system can be a fixed mechanical or electronic system, software (printer driver, robot control), a human, or any other input.

Energy sources: electric (current, voltage) · pressure (pneumatic = air, hydraulic = fluid) · manual (for example, a crankshaft).

5.2 Types of actuators ★★

Practice — this year's assignment

5.3 Soft actuators

• Soft actuators (for example, polymer-based) handle fragile objects: fruit harvesting in agriculture; moving internal organs in biomedicine.
• They give flexible motion. Tiny molecular-level changes add up to a large deformation of the material.

Shape Memory Polymers (SMP)

• Work like our muscles. They respond to light, electrical, magnetic, heat, pH and moisture changes.
• Features: low density, high strain recovery, biocompatibility,^{66Biocompatibility — safe for contact with living tissue.} biodegradability.

Light Activated Polymers (LAP)

• A special type of SMP triggered by light.
• Instant response. Controlled remotely, with no physical contact, by changing light frequency or intensity.

Practice — this year's assignment

6. Basics of IoT networking

6.1 Convergence and components

• IoT is the convergence of many domains (source: Vermesan & Friess, 2013).
• Layers of an IoT system (top to bottom): Applications → Backend services → Internet → Local network → Device (the "Thing").

Functional components of IoT ★

1. Interaction and communication with other IoT devices
2. Processing and analysis of operations
3. Internet interaction
4. Handling web services of applications
5. Integrating application services
6. A user interface to access IoT

Example implementation: Sensor mote^{67Mote — a small wireless sensor node.} → processor → radio → gateway → Internet → proxy server (via WebSocket^{68WebSocket — a web protocol that keeps a two-way connection open between client and server.}) → cloud server → analytics and actuation.

Practice — this year's assignment

6.2 IoT categories

IoT Service-Oriented Architecture (SOA)^{69SOA (Service-Oriented Architecture) — software built as separate services that talk over a network.} (Xu, He & Li, 2014): a layered model with sensing, networking, service and interface layers. Know its name and source.

IoT and associated technologies: Big data · SDN^{70SDN (Software-Defined Networking) — network control moves from devices into central software.} · Cloud · 3G/4G/5G · Smart grid · CPS · IoV^{71IoV (Internet of Vehicles) — IoT applied to connected cars and traffic systems.} · Telemedicine^{72Telemedicine — medical care given at a distance by telecom links.} · M2M.

6.3 Technical deviations from the regular web ★ (image slide)

Hook: IoT is light. It uses UDP, 6LoWPAN and 802.15.4 because devices have little power and memory.

6.4 IoT challenges ★

Security · Scalability · Energy efficiency · Bandwidth management · Modeling and analysis · Interfacing · Interoperability^{77Interoperability — the ability of different systems to work together.} · Data storage · Data analytics · Complexity management (for example, SDN)

6.5 Design considerations

• Communication with the outside world decides the network architecture.
• The choice of communication technology decides device hardware needs and cost.
• No single networking paradigm fits all needs, because IoT has so many applications.

6.6 Complexity and wireless issues

Complexity of networks: network growth · interference among devices · network management · heterogeneity^{78Heterogeneity — a mix of many different types of devices and networks.} · protocol standardization.

Wireless networks: traffic and load management · variations (Wireless Body Area Networks and other Personal Area Networks^{79PAN (Personal Area Network) — a network of a few metres around one person. Body Area Network is a PAN on or near the body.}) · interoperability · network management · overlay networks.^{80Overlay network — a virtual network built on top of another network.}

Scalability: flexibility within the Internet · IoT integration · large-scale deployment · real-time connectivity of billions of devices.

7. Exam Focus

7.1 Facts that MCQs love

• IoT full form: Internet of Things.
• IoT is a union of: low-power embedded systems, cloud computing, big data, machine learning, networking. (Not "high-power" embedded systems or supercomputing.)
• ITU 2005 report named the term. Enablers: RFID, nanotechnology, sensors, smart networks.
• IoT characteristic that is wrong: ambiguous naming and addressing. The right one is unambiguous.
• More devices → address crunch. (True.)
• Gateway assigns a unique address to each field device in its area, and holds a unique network prefix.
• Gateway connects the LAN to a WAN and the Internet, forwarding packets at the IP layer.
• Multi-homing = connection to multiple networks for reliability. (Not "transparent roaming" and not "multi-clouding".)
• IPv6 source address = 128 bits. IPv4 = 32 bits.
• IPv6: fixed header, no checksum, no options, IPSec compulsory, multicast, hex notation.
• Resolution ≠ accuracy.
• Actuator needs: control signal + energy source.
• Sensor is the input device. Actuator is the output device.
• Hydraulic = liquid, large force. Pneumatic = air, quick response. Electric = cheap, clean, fast. Mechanical = rotary → linear.
• Everyday devices (fan, light, refrigerator) can all be IoT "things". If the option is "none of these" for "which is not a thing", the answer is usually None of these.
• Embedded systems are the core of IoT devices. (True.)

7.2 Extra facts that past papers asked (small revision, not in these slides)

• A sensor node has: a sensing unit, a processing unit, a power unit (and often a communication unit). "All of these" is the usual answer.
• Common OS in IoT: Contiki, TinyOS, RIOT. Regular desktop OSs are not the norm.
• Low-rate wireless protocol: IEEE 802.15.4 (base of ZigBee and 6LoWPAN).
• Typical IoT sensors: temperature, humidity, pressure, motion (PIR), proximity, gas, light.

7.3 Ten-line last-minute recap

1. IoT = physical objects + embedded tech + connectivity.
2. Term from the ITU 2005 report. Enablers: RFID, nano, sensors, smart networks.
3. M2M ⊂ IoT. WoT = IoT over web technology (REST).
4. Growth causes an address crunch. IPv6 (128 bit) is the fix.
5. Gateway = LAN ↔ WAN router with a unique prefix. Proxy = application-layer work.
6. Multi-homing = many networks for reliability.
7. Sensor: analog / digital; scalar / vector. Errors: offset, drift, noise, hysteresis, quantization, aliasing.
8. Resolution = smallest detectable change. Accuracy is independent.
9. Actuator: hydraulic, pneumatic, electric, thermal/magnetic, mechanical, soft (SMP, LAP).
10. IoT stack: 802.15.4 → 6LoWPAN/IPv6 → UDP/DTLS → MQTT/CoAP → JSON/CBOR.

Footnotes