7. Wireless, Mobile and Internet of Things Networks

Introduction and Context

Wireless networking has changed the meaning of network access. Communication is no longer tied to a fixed cable or a permanently known endpoint. A user may move between access points, a vehicle may exchange data while travelling, and a sensor may operate for years without a wired power source. These capabilities create substantial opportunities, but they also introduce uncertainty in channel quality, device location, energy availability, interference, security, and operational control.

This chapter examines wireless, mobile, and Internet of Things (IoT) networks as related but distinct systems. Wireless LANs such as Wi-Fi provide high-speed local connectivity. Cellular systems provide wide-area mobility through carefully engineered radio access and core networks. Ad hoc and sensor networks organize communication without depending entirely on fixed infrastructure. Low-power wide-area networks support small, infrequent transmissions from constrained devices. IoT architectures connect sensing, computation, communication, analytics, and actuation into an operational system.

A central theme is trade-off. Higher data rates often require more bandwidth, energy, and favourable channel conditions. Wider coverage may require lower data rates. Mobility improves availability but complicates authentication, routing, and session continuity. Dense spectrum reuse increases capacity but also increases interference. IoT deployments therefore require architectural decisions rather than simply selecting a radio technology.

Learning Objectives

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

1. Explain wireless propagation, fading, interference, and link-budget considerations.

2. Compare Wi-Fi, cellular, ad hoc, sensor, and LPWAN architectures.

3. Interpret the roles of 4G and 5G radio access and core-network functions.

4. Explain mobility management, handover, location registration, and roaming.

5. Select suitable IoT communication protocols for different applications.

6. Analyse spectrum, capacity, latency, reliability, security, and energy trade-offs.

7. Design a reasoned network approach for a practical mobile or IoT scenario.

The Wireless Channel: Why Radio Networking Is Uncertain

In a wired network, the transmission medium is usually engineered to remain relatively stable. A wireless signal, by contrast, propagates through an environment containing walls, vehicles, vegetation, people, and other transmitters. The received signal is therefore a function of distance, frequency, antenna characteristics, obstacles, reflection, diffraction, scattering, and time.

Free-space path loss provides a first approximation of attenuation. In decibels, it can be represented as:

PL(dB) = 32.44 + 20 log10(fMHz) + 20 log10(dkm)

where f is frequency and d is distance. Real environments normally introduce additional loss, represented through a path-loss exponent. At short range, a line-of-sight link may behave close to free space; in a dense urban environment, attenuation may increase much faster with distance.

Multipath propagation occurs when copies of a signal arrive through different paths. The copies may add constructively or destructively, producing fading. Slow fading is commonly associated with shadowing by large objects, while fast fading changes over distances comparable to the wavelength or over short time intervals as a user moves. Doppler shift, caused by relative motion, changes the apparent frequency of the signal and can impair demodulation when mobility is high.

A link budget estimates whether a receiver can decode a transmission. It includes transmit power, antenna gains, cable and implementation losses, propagation loss, and receiver sensitivity. The difference between received power and receiver sensitivity is the link margin. A large margin improves robustness but may require more power or better antennas. The signal-to-interference-plus-noise ratio (SINR), rather than signal strength alone, is usually the more useful performance measure because interference may dominate thermal noise.

Wireless systems manage uncertainty through modulation and coding adaptation, retransmission, diversity, beamforming, channel estimation, scheduling, and power control. These mechanisms improve reliability, but they consume processing resources, bandwidth, or energy. The same physical channel can consequently provide very different throughput and latency under different loads and mobility conditions.

Wi-Fi as a Distributed Local-Area System

Wi-Fi, based primarily on the IEEE 802.11 family, uses a wireless LAN architecture. In infrastructure mode, stations communicate through an access point (AP), which typically connects the wireless segment to a wired distribution system. An extended service set can contain multiple APs with coordinated identifiers and overlapping coverage. A station may associate with one AP and later reassociate with another as it moves.

The wireless medium is shared. Traditional Wi-Fi uses carrier-sense multiple access with collision avoidance (CSMA/CA). A station listens before transmission, waits for a random backoff interval when the medium is busy, and may use acknowledgements and retransmissions. Because a wireless station cannot reliably detect a collision while transmitting, collision avoidance differs fundamentally from Ethernet collision detection.

Modern Wi-Fi generations improve performance through wider channels, higher-order modulation, multiple-input multiple-output (MIMO), spatial streams, scheduled access, and operation in additional frequency bands. However, nominal data rate is not user throughput. Protocol overhead, contention, interference, distance, channel width, and the number of active clients reduce application performance. Wide channels can increase peak speed but may be inefficient in dense deployments where narrower channels permit greater reuse.

A professional Wi-Fi design begins with requirements: coverage, capacity, client density, application latency, roaming behaviour, security, and availability. A warehouse using scanners, a lecture hall hosting hundreds of clients, and a hospital supporting clinical devices have different radio and operational requirements. Site surveys, spectrum analysis, channel planning, transmit-power control, identity management, and monitoring are more important than simply installing additional APs.

Cellular Architecture and the 5G Transition

Cellular networks divide a geographical region into cells served by base stations. Frequency reuse allows distant cells to use related resources while limiting interference through planning and power control. A cellular system contains at least three logical domains: the user equipment, the radio access network, and the core network.

The radio access network provides the air interface, scheduling, radio resource control, and handover support. The core network authenticates subscribers, manages sessions, assigns connectivity, applies policy, and connects users to external networks. In 4G, the Evolved Packet Core integrates functions such as mobility management, serving gateways, packet gateways, and policy control. 5G introduces a service-based core with functions including access and mobility management, session management, user-plane forwarding, authentication, policy, and network exposure.

5G is not merely a faster radio interface. Its design supports enhanced mobile broadband, ultra-reliable low-latency communication, and massive machine-type communication. These are target service categories rather than guarantees available simultaneously everywhere. Performance depends on spectrum, deployment density, transport capacity, device capability, workload, and service configuration.

Important 5G concepts include massive MIMO, beam management, flexible numerology, edge computing, network slicing, and cloud-native core functions. Network slicing permits logically separated services with distinct policies and performance objectives over shared infrastructure. It requires strong isolation, orchestration, monitoring, and admission control; it is not automatically equivalent to a physically separate network.

The distinction between control plane and user plane is strategically important. The control plane supports registration, authentication, mobility, and session establishment. The user plane carries application traffic. Separating and distributing these functions allows operators to place user-plane processing closer to applications, reducing latency and backhaul load. Edge computing can support industrial control, augmented reality, and local analytics, although it adds operational complexity.

Mobility Management and Continuity of Service

Mobility management answers two questions: where is a device, and how can its communication continue while its point of attachment changes? Cellular systems maintain registration and reachability information, while Wi-Fi networks use association and reassociation mechanisms. Internet mobility may also require an address or identifier that remains stable while the point of network attachment changes.

A handover transfers an active connection from one access point or cell to another. It may be network controlled, mobile controlled, or jointly controlled. The decision can consider received signal strength, quality, predicted movement, load, service type, and policy. A handover that occurs too late may cause packet loss; one that occurs too early may cause unnecessary signalling and repeated transitions, known as ping-pong handovers.

Mobility procedures involve measurement, candidate selection, preparation, execution, path switching, and resource release. Fast mobility requires buffering, forwarding, context transfer, and authentication strategies that minimize interruption. In heterogeneous networks, a device may move between technologies with different addressing, security, and quality-of-service models. Seamless operation is therefore an architectural property, not merely a radio feature.

Roaming allows a subscriber to obtain service through another operator or administrative domain. It introduces agreements concerning authentication, charging, security, lawful access, quality, and data handling. Enterprise deployments must also consider whether device identity, location information, and telemetry are permitted to cross national or organizational boundaries.

Ad Hoc, Mesh, and Sensor Networks

An ad hoc network forms without relying on a pre-existing fixed infrastructure. Nodes may act as both hosts and routers, creating multihop paths. Reactive routing protocols discover routes when needed, reducing periodic overhead but increasing initial delay. Proactive protocols maintain routing information continuously, improving route availability at the cost of control traffic. Hybrid approaches combine the two.

Mesh networks use multiple forwarding nodes to improve coverage and resilience. They are useful where cabling is expensive or impossible, but each additional hop can reduce capacity because a radio may need to receive and retransmit traffic. Routing must account for link quality, interference, energy, and node mobility rather than hop count alone.

Wireless sensor networks consist of small nodes that sense physical conditions and communicate observations to a sink or gateway. A sensor node commonly includes a sensing unit, processor, transceiver, power source, and sometimes an actuator. Since communication often consumes more energy than local computation, aggregation, duty cycling, sleep scheduling, and event-driven reporting are central design techniques.

Sensor networks face constrained memory, intermittent connectivity, physical exposure, inaccurate clocks, and limited maintenance access. A design that maximizes raw throughput may be inferior to one that provides sufficiently accurate data while allowing multi-year battery life. Data quality, calibration, time synchronization, and failure detection are as important as routing.

LPWAN and the IoT Protocol Stack

Low-power wide-area networks (LPWANs) serve devices that transmit small amounts of data over long distances with low energy consumption. They generally trade throughput and responsiveness for coverage, device lifetime, and deployment simplicity. Technologies may use licensed cellular spectrum or unlicensed spectrum and may differ in bidirectional capability, payload limits, duty-cycle rules, and operator model.

IoT architecture is often described in layers. The device layer senses or acts. The connectivity layer transports data. The edge or gateway layer aggregates, filters, translates, and protects traffic. The platform layer manages devices, identities, digital twins, storage, and rules. The application layer turns data into operational decisions. This layered view is useful, but real systems frequently combine functions at gateways, cloud services, and devices.

At the application layer, MQTT uses a broker and publish/subscribe topics. It is well suited to intermittent devices and event-driven telemetry. CoAP provides a compact request/response model for constrained devices and commonly uses UDP, with security provided through mechanisms such as DTLS or an appropriate object-security approach. HTTP is widely supported and integrates easily with web services, but its headers, connection behaviour, and security overhead may be expensive for very constrained nodes. Message formats such as JSON are convenient but verbose; compact binary formats can reduce bandwidth and energy use.

Device management requires more than message delivery. Systems need provisioning, authentication, authorization, configuration, health monitoring, time synchronization, secure software updates, key rotation, and retirement procedures. A device that cannot be updated securely can become a long-term vulnerability.

Spectrum, Capacity, and Energy Efficiency

Spectrum is scarce and contested. Licensed spectrum provides coordinated access and predictable regulatory control, whereas unlicensed bands permit shared use but create uncertainty through contention and interference. Regulators define frequency allocations, emission limits, power limits, channelization, and usage conditions. Organizations must comply with local rules rather than assuming that a radio configuration is globally valid.

Capacity depends on bandwidth, SINR, spatial reuse, coding efficiency, protocol overhead, and the number of competing users. The Shannon relationship provides an upper-bound intuition: capacity increases with bandwidth and with the logarithm of signal-to-noise ratio. Increasing power alone therefore produces diminishing returns, particularly in interference-limited networks.

Energy efficiency should be measured as useful work per unit of energy, not merely as low transmit power. Sleep scheduling, wake-up radios, adaptive reporting, local aggregation, energy-aware routing, efficient firmware, and right-sized cryptography all contribute. A sensor that sends every raw measurement to the cloud may consume more energy and create less useful information than one that detects local anomalies and transmits only relevant events.

Security, Governance, and Professional Responsibility

Wireless exposure enlarges the attack surface. Threats include eavesdropping, rogue access points, impersonation, jamming, replay, traffic analysis, credential theft, denial of service, insecure firmware, and physical tampering. IoT systems are especially vulnerable when devices are deployed for long periods with default credentials or unsupported software.

Security should begin with threat modelling and asset identification. Strong device identity, mutual authentication, least privilege, encrypted transport, secure boot, protected key storage, signed updates, network segmentation, logging, and anomaly detection are foundational controls. Availability must be considered alongside confidentiality and integrity because interference or battery exhaustion can disable safety-related systems.

Privacy and governance are equally important. Sensors may infer occupancy, behaviour, health, or location even when they do not explicitly collect personal names. Organizations should define purpose limitation, retention, access control, consent where applicable, auditability, and incident response. Safety-critical IoT systems require validation, fail-safe behaviour, manual override, and responsibility assignment. Automation does not remove professional accountability.

Applied Case: Connected Cold-Chain Monitoring

Background

A regional pharmaceutical distributor operates refrigerated vehicles and storage rooms. Temperature excursions can damage inventory, but the organization currently relies on manual readings taken several times a day. Management wants continuous monitoring with alerts and auditable records.

Situation and Problem

Vehicles travel through urban and rural areas. Sensors must operate for months on battery power, communicate through metal containers, tolerate intermittent coverage, and report urgent excursions quickly. The system must distinguish a genuine temperature breach from a faulty sensor or temporary radio outage.

Analysis

A high-bandwidth Wi-Fi design would be unsuitable because vehicles are mobile and may lack reliable local infrastructure. A cellular LPWAN approach is attractive for wide-area coverage and small telemetry messages, but it must be evaluated against coverage gaps, latency requirements, subscription cost, payload limits, and indoor attenuation. A gateway can collect readings from several sensors using a short-range protocol and forward summaries through a cellular connection.

The application requires threshold alerts rather than continuous high-rate streaming. Local processing can detect a persistent excursion and retain samples during an outage. The gateway should timestamp messages, buffer them securely, and transmit them when connectivity returns. Sensors require unique identities, authenticated onboarding, signed firmware updates, and tamper detection.

Approach and Solution

The proposed architecture uses temperature sensors connected to a vehicle gateway. Sensors sleep between measurements and transmit only readings or exception events. The gateway performs validation, maintains a local history, and sends encrypted MQTT messages to an IoT platform. The platform stores data, applies rules, notifies responsible personnel, and exposes dashboards and compliance reports. Cellular connectivity is supplemented by store-and-forward operation when coverage is temporarily unavailable.

The deployment is tested through a pilot covering different routes, weather conditions, container loads, and battery states. Performance measures include delivery success, alert delay, false-alert rate, battery life, data completeness, and time required to replace a device.

Outcome and Learning Points

The design aligns technology with the actual information requirement: reliable, auditable events rather than maximum throughput. Local intelligence reduces traffic and improves resilience. Security and lifecycle management are treated as operational requirements, not later additions. The case also demonstrates that a communication technology cannot be selected independently of physical environment, business process, regulatory obligations, and maintenance capability.

Challenges and Common Design Mistakes

A frequent mistake is treating advertised data rate as guaranteed application performance. Another is using a single coverage measurement instead of analysing interference, mobility, indoor loss, and peak load. Adding access points or repeaters without channel planning can worsen contention. In IoT, sending excessive data to a remote cloud increases energy use and operational cost. Conversely, aggressive compression or local filtering can remove information needed for diagnosis.

Designers may also overlook failure modes. What happens when a gateway loses power, credentials expire, a device is stolen, time synchronization fails, or a firmware update is interrupted? Battery replacement, spectrum regulation, roaming charges, vendor lock-in, and end-of-life support should be included in the total-cost analysis.

Emerging Directions

Current development is moving toward integrated terrestrial and non-terrestrial connectivity, private 5G networks, Wi-Fi evolution, edge-native applications, deterministic wireless control, and more capable IoT devices. AI and machine learning are increasingly used for radio-resource optimization, traffic prediction, anomaly detection, predictive maintenance, and energy management. Their outputs must be monitored because biased training data, model drift, and opaque decisions can create operational risk.

Research continues on reconfigurable intelligent surfaces, integrated sensing and communication, semantic communications, ambient backscatter, energy harvesting, and stronger convergence between communication and computing. These technologies are promising but should be evaluated through measurable requirements rather than novelty alone.

Best-Practice Design Framework

A robust wireless or IoT project can follow this sequence:

1. Define the application, users, devices, mobility, physical environment, and failure consequences.

2. Specify measurable requirements for coverage, capacity, latency, reliability, energy, security, and lifecycle.

3. Characterize spectrum and propagation through modelling, surveys, and pilot measurements.

4. Compare architectures, including infrastructure, mesh, cellular, LPWAN, and hybrid options.

5. Design identity, security, management, data governance, and update mechanisms from the beginning.

6. Prototype under realistic load and environmental conditions.

7. Monitor service-level indicators and energy consumption after deployment.

8. Plan maintenance, incident response, scaling, interoperability, and retirement.

Chapter Summary

Wireless networks operate over variable and shared channels affected by path loss, fading, mobility, interference, and regulatory constraints. Wi-Fi provides flexible local access, while cellular networks provide managed wide-area mobility through radio access and core functions. 5G extends this model with service-oriented cores, advanced radio techniques, edge computing, and logical slicing. Ad hoc, mesh, and sensor networks emphasize distributed operation, multihop routing, and energy-aware communication. LPWAN technologies support low-volume, long-range IoT telemetry through deliberate trade-offs in throughput and responsiveness.

IoT architecture extends beyond connectivity to include devices, gateways, platforms, applications, identity, management, analytics, and actuation. Effective designs align communication choices with coverage, capacity, latency, energy, security, privacy, regulation, and lifecycle needs. Mobility management, secure updates, spectrum planning, local processing, and failure analysis are essential professional concerns.

Key Terms

Wireless propagation: The transmission of electromagnetic signals through an environment.

Path loss: Reduction in received signal power as a signal travels through space and obstacles.

Multipath fading: Signal variation caused by multiple reflected, diffracted, or scattered paths.

Doppler shift: Frequency change resulting from relative transmitter–receiver motion.

SINR: Ratio of desired signal power to interference plus noise power.

Link budget: Calculation of gains, losses, received power, and link margin.

Wi-Fi access point: Device providing wireless LAN connectivity to client stations.

Cellular cell: Geographic area served by a cellular radio site.

Handover: Transfer of an active connection between access points or cells.

Roaming: Use of network service through another administrative or operator domain.

MIMO: Use of multiple antennas to improve capacity or reliability.

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

Ad hoc network: Network formed dynamically without depending entirely on fixed infrastructure.

LPWAN: Low-power wide-area networking designed for small, energy-efficient transmissions.

MQTT: Broker-based publish/subscribe messaging protocol commonly used in IoT.

CoAP: Lightweight request/response protocol designed for constrained devices.

Edge computing: Processing data near its source to reduce latency or backhaul traffic.

Duty cycling: Alternating active and sleep states to reduce energy consumption.

Digital twin: Software representation that reflects the state or behaviour of a physical entity.

Secure boot: Startup process that executes only authenticated software.

Review Questions

Short Answer Questions

1. What factors distinguish SINR from simple received signal strength?

2. Why does multipath propagation cause fading?

3. What are the principal functions of a cellular core network?

4. Define LPWAN and identify its principal design trade-off.

5. Why is secure firmware updating important in IoT deployments?

Long Answer Questions

1. Explain wireless propagation and show how a link budget supports design decisions.

2. Compare Wi-Fi and cellular networks in terms of access control, mobility, coverage, and capacity.

3. Describe the stages and challenges of a cellular handover.

4. Explain the architecture of an IoT system from device to application layer.

5. Discuss energy-efficient communication techniques for wireless sensor networks.

Analytical / Application Questions

1. A university has a dense lecture hall with high client concentration and moderate mobility. Propose a Wi-Fi design strategy and justify channel, capacity, monitoring, and security decisions.

2. A remote agricultural sensor sends a small measurement every hour and must operate for five years. Compare a Wi-Fi, cellular, and LPWAN solution using coverage, energy, cost, and data requirements.

3. Design a resilient architecture for mobile cold-chain monitoring. Explain how you would handle outages, authentication, local buffering, alerts, battery life, and auditability.

22. Multiple Choice Questions with Answers

Question 1

Which measure best captures the effect of both interference and thermal noise on a wireless link?

A. Received signal strength only

B. SINR

C. Channel bandwidth only

D. Antenna height

Correct Answer: B. SINR

Question 2

Why can increasing Wi-Fi channel width reduce performance in a dense deployment?

A. It always lowers antenna gain

B. It prevents encryption

C. It leaves fewer non-overlapping channels for reuse

D. It eliminates acknowledgements

Correct Answer: C. It leaves fewer non-overlapping channels for reuse

Question 3

Which function primarily belongs to the cellular core network?

A. Physical antenna beam formation only

B. Subscriber authentication and session management

C. Detecting every radio collision

D. Replacing all user devices

Correct Answer: B. Subscriber authentication and session management

Question 4

A handover performed too early and repeatedly between two cells is commonly associated with which problem?

A. Packet encapsulation

B. Ping-pong handovers

C. Secure boot

D. Spectrum licensing

Correct Answer: B. Ping-pong handovers

Question 5

Which characteristic most strongly supports choosing LPWAN for a remote environmental sensor?

A. High-volume video transmission

B. Very low latency for closed-loop control

C. Small, infrequent messages and long battery life

D. Permanent high-speed local connectivity

Correct Answer: C. Small, infrequent messages and long battery life

Question 6

In MQTT, what component normally receives publications and distributes them to subscribers?

A. Broker

B. Modulator

C. Access antenna

D. Routing table

Correct Answer: A. Broker

Question 7

Why is local processing useful in the cold-chain case?

A. It removes the need for device identity

B. It allows decisions and buffering during connectivity loss

C. It guarantees unlimited battery life

D. It eliminates all sensor faults

Correct Answer: B. It allows decisions and buffering during connectivity loss

Question 8

Which measure is most directly intended to protect an IoT device from running unauthorized firmware?

A. Secure boot

B. Wider channel allocation

C. Higher transmit power

D. Random backoff

Correct Answer: A. Secure boot

Question 9

What is a principal energy-saving technique in sensor networks?

A. Keeping the radio continuously active

B. Transmitting every raw measurement immediately

C. Duty cycling and event-driven reporting

D. Maximizing packet headers

Correct Answer: C. Duty cycling and event-driven reporting

Question 10

Which statement best describes network slicing?

A. Physically replacing every shared network element

B. Creating logical service partitions over shared infrastructure

C. Disabling mobility management

D. Using only unlicensed spectrum

Correct Answer: B. Creating logical service partitions over shared infrastructure