IoT Connectivity and Communication Protocols

Introduction and Context

Connectivity is the architectural foundation of the Internet of Things (IoT). A sensor may measure temperature accurately, but its value is limited unless the measurement can be transported, interpreted, secured, and integrated with an application. IoT connectivity therefore involves more than selecting a radio. It includes the physical medium, link-layer behavior, addressing, network formation, transport, messaging, device discovery, security, power management, and lifecycle operations.

IoT systems operate across very different environments. A factory may require deterministic wired communication and electromagnetic resilience; a smart meter may need ten years of battery life and wide-area coverage; a wearable may require low latency and very low energy consumption; and a video camera may need high throughput. No single technology optimizes all these dimensions. Connectivity design is consequently a multi-objective engineering decision involving range, throughput, latency, energy, reliability, mobility, density, cost, spectrum, security, and regulatory constraints.

This chapter compares major wired and wireless technologies and examines application protocols used above the network layer. It also explains how device discovery enables scalable deployment and how protocol choices should be made systematically rather than by popularity.

Learning Objectives

After studying this chapter, the reader should be able to:

1. Distinguish IoT connectivity layers and explain the roles of physical, link, network, transport, and application protocols.

2. Compare wired technologies with RFID, NFC, BLE, Wi-Fi, Zigbee, Thread, LPWAN, NB-IoT, LTE-M, and 5G.

3. Select connectivity technologies using range, energy, throughput, latency, reliability, cost, and deployment criteria.

4. Explain and apply MQTT, CoAP, HTTP, and AMQP in IoT architectures.

5. Describe device discovery mechanisms and their security implications.

6. Evaluate an IoT connectivity design in a realistic engineering case.

1. A Layered View of IoT Communication

A useful abstraction is the protocol stack:

• Physical layer: carries symbols over copper, fiber, or radio.
• Link layer: provides framing, medium access, error detection, and sometimes retransmission.
• Network layer: provides addressing and routing, commonly IPv4 or IPv6.
• Transport layer: provides end-to-end delivery, such as TCP, UDP, or QUIC.
• Application layer: defines device data, commands, sessions, and service semantics.

IoT deployments also contain gateways, brokers, cloud platforms, and application programming interfaces. A constrained sensor might use IEEE 802.15.4 at the local link layer, IPv6 over a low-power adaptation layer, UDP at transport, and CoAP at application level. A cellular sensor may use IP directly over an operator network and communicate with an MQTT broker.

A connectivity decision should distinguish communication technology from application protocol. BLE, Wi-Fi, Zigbee, and 5G primarily describe access and networking. MQTT, CoAP, HTTP, and AMQP describe how application data is exchanged. They are complementary rather than mutually exclusive categories.

Important performance terms include:

• Throughput: useful data delivered per unit time, normally measured in bit/s.
• Latency: elapsed time between transmission and reception or processing.
• Jitter: variation in latency.
• Range: maximum useful separation under specified propagation and deployment conditions.
• Availability: probability that the service is operational when required.
• Energy per message: energy consumed to sense, process, transmit, receive, and sleep between transactions.
• Scalability: ability to support increasing devices, traffic, and administrative complexity.

Nominal data rate is not application throughput. Headers, contention, retransmissions, sleep schedules, coding, interference, and protocol overhead reduce useful performance. Similarly, nominal range depends strongly on antenna design, transmit power, frequency, obstacles, receiver sensitivity, and regulatory limits.

2. Wired Connectivity

2.1 Ethernet

Ethernet is widely used in industrial automation, gateways, cameras, buildings, and data centers. It provides high throughput, mature management, and strong interoperability. Power over Ethernet (PoE) can carry both data and electrical power over twisted-pair cabling, simplifying installation for access points, cameras, and sensors.

Advantages include predictable physical infrastructure, low susceptibility to radio interference, high bandwidth, and straightforward IP integration. Limitations include installation cost, inflexible cabling, vulnerability to cable damage, and difficulty connecting mobile or geographically dispersed devices. Industrial Ethernet variants and time-sensitive networking can support bounded latency and synchronization, but design must account for switch configuration, redundancy, traffic classes, and electromagnetic conditions.

2.2 Fieldbus and Serial Networks

RS-485-based systems, Controller Area Network (CAN), Modbus, PROFIBUS, and related fieldbuses remain important in industrial environments. They often provide simple, robust communication over long cable runs and support deterministic or semi-deterministic control. Their limitations include lower bandwidth than modern Ethernet, vendor-specific profiles, and more difficult Internet integration. Gateways are commonly used to translate fieldbus data into IP, MQTT, or industrial information models.

2.3 Optical Fiber

Fiber offers high bandwidth, long distance, electrical isolation, and immunity to electromagnetic interference. It is appropriate for backbone links, substations, campuses, and harsh industrial areas. Its disadvantages are higher termination complexity, bending constraints, and installation cost. Fiber is usually a backbone technology rather than a direct connection for small battery sensors.

3. Short-Range Wireless Technologies

3.1 RFID

Radio-frequency identification uses a reader and a tag. Passive tags harvest energy from the reader field and respond with stored identification or sensor information; active tags contain a battery and can provide greater range or sensing capability. RFID is effective for identification, inventory, access control, supply-chain visibility, and asset tracking.

RFID is not normally a general-purpose bidirectional network. It is optimized for rapid identification and limited data exchange. Design issues include reader collision, tag orientation, metal and liquid absorption, privacy, and unauthorized reading. Security may require cryptographic tags, protected reader infrastructure, and policies governing tag ownership and data retention.

3.2 NFC

Near Field Communication operates at very short range and supports proximity interactions such as contactless payment, pairing, access credentials, and product authentication. Its short range can be a security advantage because intentional physical proximity is required, although relay attacks and malicious tags remain possible. NFC is appropriate for commissioning and user interaction, often complementing BLE or Wi-Fi rather than replacing them.

3.3 Bluetooth Low Energy

BLE is designed for low-power short-range communication in wearables, medical devices, beacons, proximity systems, and consumer products. Devices may communicate through advertising and scanning or establish connections using services and characteristics defined by the Generic Attribute Profile (GATT). BLE supports flexible power behavior because a device can sleep and transmit brief events.

BLE is attractive for low energy, broad smartphone support, and relatively simple commissioning. Its limitations include short practical range in obstructed environments, dependence on central devices or gateways in many architectures, and possible congestion in the crowded 2.4 GHz band. BLE Mesh extends communication across managed multi-hop networks, but increases provisioning, routing, and security complexity.

3.4 Wi-Fi

Wi-Fi provides high throughput and native IP connectivity. It is suitable for cameras, appliances, industrial terminals, and devices with mains power or frequent charging. Modern generations improve spectral efficiency, scheduling, and operation in dense networks. Wi-Fi is usually a poor choice for tiny battery sensors that transmit only a few bytes because association, listening, security negotiation, and contention can consume substantial energy.

Wi-Fi design must consider coverage surveys, channel planning, roaming, authentication, network segmentation, interference, and backhaul capacity. A secure IoT deployment should place devices in dedicated networks and restrict east-west communication where possible.

4. Low-Power Mesh and Wide-Area Technologies

4.1 Zigbee

Zigbee commonly uses IEEE 802.15.4 radios and provides low-power mesh networking, addressing, security, and application profiles. Routers can forward traffic while end devices sleep. It is used in lighting, building automation, and metering.

Mesh networking can improve coverage and resilience, but routing devices require energy and careful placement. Battery devices must select parents, manage polling, and tolerate route changes. Interoperability depends not only on the radio but also on profiles, commissioning procedures, and certification.

4.2 Thread

Thread is an IPv6-based low-power mesh networking technology built on IEEE 802.15.4. It uses border routers to connect the mesh to other IP networks. This IP orientation simplifies integration with Internet services and supports low-power end devices. Thread itself is a network technology rather than a complete application model; higher-level ecosystems may use Matter or other protocols.

Thread deployments require leader and router management, commissioning credentials, border-router availability, and appropriate handling of sleepy end devices. Compared with a proprietary non-IP mesh, its IP foundation can reduce translation requirements, although operational complexity remains significant.

4.3 LPWAN

Low-power wide-area networking (LPWAN) denotes technologies optimized for small payloads, long range, low device energy consumption, and modest data rates. They are useful for meters, environmental sensors, agriculture, logistics, and infrastructure monitoring. Typical engineering characteristics include long sleep periods, infrequent uplinks, limited downlink capacity, and strong dependence on coverage and duty-cycle or spectrum rules.

LPWAN is not appropriate for continuous audio, video, or tightly closed-loop control. Link budgets may be favorable, but latency, downlink availability, payload limits, and subscription or gateway costs must be evaluated. Application design often uses event reporting, local filtering, batching, and store-and-forward behavior.

4.4 NB-IoT and LTE-M

NB-IoT is a cellular IoT technology optimized for low data rates, deep coverage, and stationary or low-mobility devices. It is suitable for metering, environmental monitoring, and distributed sensors. LTE-M provides greater throughput, lower latency, mobility support, and voice-related capabilities than NB-IoT, generally at higher energy or module complexity in comparable use cases.

Both benefit from licensed spectrum, operator-managed coverage, SIM or eSIM-based identity, and mature cellular security. They also introduce operator dependence, subscription costs, coverage variability, and potential challenges when networks are retired or technologies evolve. Power-saving modes such as extended discontinuous reception and power-saving mode can reduce energy consumption, but application timing must accommodate delayed reachability.

4.5 5G

5G supports enhanced mobile broadband, massive machine-type communication, and ultra-reliable low-latency communication as broad service objectives. It can connect high-bandwidth cameras, mobile robots, industrial terminals, and dense sensor populations. Features such as network slicing, edge computing, private networks, and advanced positioning can support specialized enterprise requirements.

Claims of universal ultra-low latency should be treated cautiously. Actual performance depends on radio configuration, scheduling, core network location, transport, edge placement, congestion, and application processing. Private 5G may provide control and isolation, but requires spectrum arrangements, deployment expertise, security operations, and a business case superior to Wi-Fi or industrial Ethernet.

5. Comparative Technology Selection

A disciplined selection process begins with requirements rather than technology labels. Define payload size and frequency, acceptable latency and jitter, node density, mobility, coverage area, installation constraints, battery target, environmental conditions, security requirements, lifecycle, and total cost of ownership.

A useful qualitative comparison is:

• Ethernet/fiber: high throughput and reliability; low mobility; installation dependent.
• RFID/NFC: identification or proximity interaction; very short effective range; low transaction complexity.
• BLE: low-power personal-area connectivity; smartphone ecosystem; modest throughput.
• Wi-Fi: high throughput and IP integration; comparatively higher energy demand.
• Zigbee/Thread: low-power local mesh; good building-scale coverage; routing and commissioning complexity.
• LPWAN/NB-IoT: wide coverage and low energy for small infrequent messages; restricted throughput and downlink.
• LTE-M: cellular mobility and moderate IoT data; operator and subscription dependence.
• 5G: high capacity, mobility, and specialized services; greater infrastructure and management complexity.

A cost model should include modules, installation, gateways, spectrum or subscriptions, power systems, maintenance, security updates, cloud traffic, and eventual replacement. A low module price may be outweighed by battery replacement or site visits.

6. IoT Application Protocols

6.1 MQTT

Message Queuing Telemetry Transport is a lightweight publish/subscribe protocol. Devices publish messages to topics, and clients subscribe to topics through a broker. This decouples producers from consumers in space and time. MQTT quality-of-service levels include at-most-once delivery, at-least-once delivery, and exactly-once protocol semantics. Exactly-once has higher overhead and does not guarantee that an external business action is performed only once; applications still need idempotency.

MQTT supports retained messages, which allow a broker to store the latest value for a topic, and a Last Will message for abnormal disconnection notification. Security commonly uses TLS, client authentication, authorization policies, and topic isolation. Poorly designed topic hierarchies or unrestricted wildcards can expose data across tenants.

MQTT is effective for telemetry, commands, event distribution, and intermittent connections. It requires a broker and is less naturally suited than CoAP to very constrained request/response exchanges without an intermediary.

6.2 CoAP

The Constrained Application Protocol is designed for constrained nodes and lossy networks. It follows a RESTful model with resources and methods analogous to HTTP, commonly transported over UDP. Confirmable messages can be acknowledged and retransmitted; non-confirmable messages reduce overhead. CoAP supports observe relationships, allowing a client to receive resource updates without repeated polling.

CoAP can be secured with Datagram Transport Layer Security (DTLS) or Object Security for Constrained RESTful Environments (OSCORE), depending on the architecture. Proxies can translate between CoAP and HTTP. CoAP is suitable for local constrained networks, but engineering must address retransmission behavior, duplicate messages, congestion, timeouts, and NAT or firewall traversal.

6.3 HTTP

HTTP is ubiquitous, richly supported, and well suited to web integration, REST APIs, configuration, and firmware services. Its text-oriented representations and connection management may be excessive for very small battery devices, although HTTP/2, HTTP/3, compression, and efficient binary formats can improve performance. HTTPS is essential for confidentiality and server authentication. HTTP APIs should define authentication, authorization, rate limits, versioning, error behavior, and idempotency.

6.4 AMQP

Advanced Message Queuing Protocol is a messaging protocol designed for reliable enterprise messaging. It supports exchanges, queues, routing rules, acknowledgements, persistence, and delivery management. AMQP is useful in back-end integration, transaction-oriented workflows, and systems requiring strong broker semantics. Its feature richness and protocol overhead generally make it less suitable for the smallest battery-powered endpoints than MQTT or CoAP.

6.5 Protocol Selection

MQTT is often chosen for brokered telemetry and event distribution; CoAP for constrained RESTful devices; HTTP for broad web integration; and AMQP for enterprise messaging and reliable workflow processing. Hybrid systems are common: a sensor uses CoAP locally, a gateway translates data to MQTT, and the cloud platform exposes HTTP or AMQP interfaces.

7. Device Discovery and Commissioning

Discovery answers three questions: what devices exist, where are they reachable, and what can they do? Mechanisms include BLE advertising and GATT services, DNS Service Discovery and Multicast DNS, DHCP and IPv6 Neighbor Discovery, CoAP resource discovery, manufacturer registries, and platform-specific directories.

Discovery should expose machine-readable metadata such as device identity, resource type, units, supported operations, firmware version, and security requirements. Discovery is not equivalent to trust. An unknown device may advertise a service, but an authorized controller must still authenticate it and verify its permissions.

Commissioning normally includes secure onboarding, credential provisioning, naming, policy assignment, and verification. Best practice is to use unique device identities, certificate- or key-based authentication, protected commissioning channels, least-privilege authorization, and the ability to revoke credentials. Unauthenticated multicast discovery should not disclose sensitive information or enable unsafe control operations.

8. Security, Governance, and Regulation

IoT connectivity expands the attack surface through radios, gateways, brokers, mobile applications, cloud APIs, and maintenance interfaces. Threats include eavesdropping, spoofing, replay, denial of service, rogue firmware, credential theft, insecure defaults, and supply-chain compromise.

Controls should include secure boot, signed firmware, hardware-protected keys where appropriate, encrypted communication, mutual authentication, network segmentation, secure logging, vulnerability management, update support, and device decommissioning. Security must be proportionate to harm: a compromised light bulb and a compromised insulin pump have very different consequences.

Governance includes privacy minimization, transparent data use, retention limits, access controls, auditability, and compliance with applicable telecommunications, safety, spectrum, and data-protection rules. Radio devices may require regulatory certification, and cellular deployments may involve lawful-intercept, identity, and operator obligations. Safety-critical systems require hazard analysis and fail-safe behavior rather than reliance on connectivity alone.

9. Detailed Applied Case: Municipal Water-Leak Monitoring

Background

The following is an illustrative case. A municipality wants to monitor pressure, flow, and acoustic indicators at 4,000 remote water assets. Most sites lack mains power and fixed Internet access. Data is needed every 15 minutes, while urgent leak events should be reported promptly.

Problem

The system must operate for several years on battery power, cover underground or semi-rural locations, tolerate intermittent coverage, protect infrastructure data, and permit maintenance teams to locate and replace faulty devices. Continuous high bandwidth is unnecessary, but reliable event delivery is important.

Analysis

Wi-Fi would require local infrastructure and impose undesirable energy consumption. Ethernet is impractical because of cabling. BLE and NFC are useful for technician commissioning but do not provide wide-area backhaul. A local Zigbee or Thread mesh would require powered routers or carefully engineered site clusters. Cellular LPWAN, NB-IoT, or LTE-M is more suitable if operator coverage is adequate. NB-IoT favors small stationary messages and deep coverage; LTE-M is preferable if mobility, lower latency, or larger firmware transfers are important.

Approach and Solution

The design uses an NB-IoT module with an eSIM, a pressure sensor, local flash storage, and a secure microcontroller. The device samples locally, applies threshold and plausibility checks, and sends periodic summaries. An exceptional pressure pattern causes an event message. MQTT over TLS is used from an aggregation service to the municipal platform; the constrained device may use a lightweight IP application protocol or a vendor cellular data interface, provided end-to-end identity and authorization are preserved.

Each device has a unique identity and signed firmware. It stores unsent data during temporary outages and uses randomized reconnection timing to avoid synchronized bursts. NFC is used by technicians for close-range commissioning and diagnostics, while a management application obtains device capabilities through authenticated discovery. The backend validates units, timestamps, sequence numbers, and duplicate messages before storing observations.

Outcome

The architecture minimizes site infrastructure, supports event-driven operation, and separates field connectivity from enterprise integration. Its main residual risks are operator coverage changes, battery aging under poor radio conditions, subscription cost, and the need to maintain secure firmware updates over a constrained link.

Learning Points

1. The best technology depends on traffic and power requirements, not merely geographical range.

2. A short-range technology such as NFC can complement rather than compete with a wide-area technology.

3. Local filtering and buffering reduce communication energy and improve resilience.

4. Delivery guarantees must be paired with application-level sequence numbers and idempotent processing.

5. Lifecycle, coverage, and security operations are as important as initial radio selection.

10. Contemporary Developments and Best Practices

Current IoT engineering increasingly combines edge computing, digital twins, standardized semantic models, and automated device lifecycle management. Edge processing reduces bandwidth and latency and can keep essential control operating during cloud outages. AI may assist with anomaly detection, adaptive sampling, radio optimization, and predictive maintenance; however, models require representative data, monitoring for drift, explainable operational thresholds, and secure update mechanisms.

Interoperability efforts increasingly favor IP-based networking, common data models, and standardized onboarding. Nonetheless, interoperability must be tested at the profile and operational level, not inferred from a shared radio standard. A practical design should:

• document requirements and failure modes before selecting a protocol;
• separate control, telemetry, management, and firmware traffic;
• design for loss, duplication, delay, and reordering;
• use authenticated encryption and rotate or revoke credentials;
• provide secure updates throughout the supported lifetime;
• measure energy under realistic radio conditions;
• test density, interference, coverage, and recovery behavior;
• plan for technology migration and vendor exit;
• validate regulatory and safety obligations;
• monitor operational metrics such as delivery rate, battery trend, latency, and firmware status.

Chapter Summary

IoT connectivity is a layered engineering problem involving media, access networks, routing, transport, application messaging, discovery, and lifecycle management. Wired Ethernet, fieldbus, and fiber provide reliability, throughput, or determinism, while RFID and NFC support identification and proximity interactions. BLE and Wi-Fi serve short-range applications with different energy and throughput profiles. Zigbee and Thread provide low-power mesh options, whereas LPWAN, NB-IoT, LTE-M, and 5G address wide-area, cellular, mobility, and capacity requirements.

MQTT provides brokered publish/subscribe messaging; CoAP offers constrained RESTful communication; HTTP provides universal web integration; and AMQP supports enterprise-grade queuing and workflows. Discovery makes devices and capabilities visible, but authentication and authorization are still required. Effective design evaluates energy, range, latency, reliability, density, security, regulation, lifecycle, and total cost rather than selecting a technology from nominal data rate alone. The water-monitoring case demonstrates how complementary technologies, local processing, buffering, secure onboarding, and application-level reliability produce a practical IoT system.

Key Terms

Connectivity: The mechanisms enabling devices to exchange data.

Throughput: Useful data delivered per unit time.

Latency: Time elapsed between sending and receiving or processing data.

Jitter: Variation in communication latency.

RFID: Radio identification using readers and electronic tags.

NFC: Very-short-range communication for proximity interactions.

BLE: Bluetooth Low Energy for low-power personal-area communication.

LPWAN: Wide-area networking optimized for low power and small payloads.

NB-IoT: Cellular technology for low-rate, deep-coverage IoT devices.

LTE-M: Cellular IoT technology supporting greater mobility and throughput than NB-IoT.

5G: Cellular generation supporting high capacity, mobility, and specialized services.

Mesh network: A network in which nodes can forward traffic for other nodes.

MQTT: Brokered publish/subscribe messaging protocol.

CoAP: Constrained RESTful application protocol, commonly using UDP.

HTTP: Web application protocol for request/response communication.

AMQP: Enterprise messaging protocol supporting queues and acknowledgements.

Device discovery: Identification of devices, endpoints, and capabilities.

Commissioning: Secure onboarding and configuration of a device.

Idempotency: Property whereby repeating an operation has the same intended effect as performing it once.

Network slicing: Logical partitioning of network resources for differentiated services.

Review Questions

Short Answer Questions

1. Why is nominal radio data rate not equivalent to application throughput?

2. State two differences between NB-IoT and LTE-M.

3. What roles do an MQTT broker and topic play?

4. Why is device discovery not sufficient for device trust?

5. Identify two advantages and two limitations of Ethernet in IoT systems.

Long Answer Questions

1. Compare BLE, Wi-Fi, Zigbee, and Thread for a smart-building deployment.

2. Explain MQTT, CoAP, HTTP, and AMQP and recommend contexts for each.

3. Discuss the engineering criteria used to select an IoT connectivity technology.

4. Explain security and governance requirements across the IoT connectivity lifecycle.

5. Describe the architecture and design reasoning in the municipal water-monitoring case.

Analytical / Application Questions

1. Design a connectivity architecture for battery-powered agricultural sensors spread across 20 km, transmitting small measurements every hour and urgent alarms immediately. Justify the radio, application protocol, buffering, and security choices.

2. A factory has Ethernet-controlled robots, Wi-Fi handheld terminals, BLE maintenance tools, and a 5G private network proposal. Analyze whether replacing all existing technologies with 5G is technically and economically justified.

3. An MQTT system reports duplicate commands after reconnection. Diagnose possible causes and propose broker, protocol, and application-level controls to prevent unsafe repeated actuation.

Multiple Choice Questions with Answers

Question 1

Which technology is primarily optimized for proximity interaction rather than wide-area telemetry?

A. NFC

B. NB-IoT

C. LTE-M

D. 5G

Correct Answer: A. NFC

Question 2

Which MQTT feature stores the latest value for a topic so a newly subscribed client can receive it?

A. Last Will message

B. Retained message

C. Topic wildcard

D. Session expiry

Correct Answer: B. Retained message

Question 3

Which protocol is specifically designed for constrained RESTful communication and commonly uses UDP?

A. AMQP

B. HTTP

C. CoAP

D. FTP

Correct Answer: C. CoAP

Question 4

Which technology is most appropriate for high-throughput IoT devices such as video cameras in a local network?

A. RFID

B. Wi-Fi

C. NFC

D. Passive BLE beaconing

Correct Answer: B. Wi-Fi

Question 5

What is the principal architectural distinction of Thread in this chapter?

A. It is an IPv6-based low-power mesh technology

B. It is a cellular technology using licensed spectrum

C. It is a proximity-only identification technology

D. It is an enterprise message-queue protocol

Correct Answer: A. It is an IPv6-based low-power mesh technology

Question 6

Which property means that repeating an operation has the same intended effect as performing it once?

A. Jitter

B. Availability

C. Idempotency

D. Throughput

Correct Answer: C. Idempotency

Question 7

Which protocol is generally most associated with enterprise queues, routing, persistence, and acknowledgements?

A. AMQP

B. BLE

C. NFC

D. CoAP

Correct Answer: A. AMQP

Question 8

For a stationary sensor sending small messages infrequently over cellular coverage, which option is generally the most directly aligned with the stated requirements?

A. NB-IoT

B. Wi-Fi

C. Ethernet

D. 5G enhanced mobile broadband

Correct Answer: A. NB-IoT

Question 9

What is a necessary step after discovering an unknown IoT device?

A. Immediately grant administrator access

B. Authenticate it and verify authorization

C. Disable all encryption

D. Assume its advertised capabilities are trustworthy

Correct Answer: B. Authenticate it and verify authorization

Question 10

Which wired medium provides electrical isolation and strong immunity to electromagnetic interference?

A. Unshielded twisted pair

B. Coaxial cable

C. Optical fiber

D. RS-485 copper cable

Correct Answer: C. Optical fiber