Communication Networks and Industrial Protocols

Introduction and context

A cyber-physical system (CPS) depends on communication to connect sensors, controllers, actuators, machines, supervisory software, and human operators. Communication is not merely a means of moving data. It influences whether a control loop is stable, whether an emergency stop is executed safely, whether production data can be trusted, and whether a plant can be integrated with enterprise and cloud systems.

Industrial communication differs from ordinary office networking in several important respects. Industrial traffic often has strict timing requirements, small but frequent messages, long equipment lifetimes, harsh electromagnetic environments, high availability requirements, and safety or regulatory consequences when communication fails. A suitable network must therefore be evaluated not only by throughput, but also by latency, jitter, determinism, synchronization accuracy, reliability, security, maintainability, and interoperability.

This chapter examines wired and wireless communication, timing and quality of service, fieldbuses, industrial Ethernet, OPC UA, MQTT, time-sensitive networking (TSN), and emerging 5G/6G connectivity. The emphasis is on selecting and engineering communication systems for mechatronic and manufacturing applications.

Learning objectives

By the end of this chapter, a student should be able to:

1. Explain the roles of bandwidth, latency, jitter, reliability, synchronization, and determinism in CPS communication.
2. Compare wired, wireless, fieldbus, industrial Ethernet, and cellular technologies.
3. Estimate transmission, propagation, processing, and queuing delay.
4. Distinguish communication protocols from physical media and application information models.
5. Explain the purposes of OPC UA, MQTT, TSN, and industrial wireless technologies.
6. Select a communication architecture for a practical automation problem.
7. Identify cybersecurity, safety, interoperability, and lifecycle concerns in networked control systems.

1. Communication requirements in cyber-physical systems

A communication network provides a path between information sources and destinations. In a CPS, the destination may be a programmable logic controller (PLC), a robot controller, a cloud analytics service, or an actuator. Requirements differ by function.

1.1 Fundamental performance measures

Bandwidth is the data-carrying capacity of a link, usually measured in bit/s. It is not the same as the useful application data rate, because headers, acknowledgements, retransmissions, and protocol overhead consume capacity.

Latency is the elapsed time between transmission of a message and its arrival or processing at the receiver. A useful decomposition is

where T_tx is transmission time, T_prop is propagation time, T_proc is processing time, and T_queue is queuing time. The terms are measured in seconds. This decomposition assumes that the communication path and processing stages can be represented as sequential delays; retransmission delays must be added when they occur.

For a frame of L bits sent over a link with rate R bit/s,

For a physical path of length d metres and propagation speed v metres per second,

In copper and fibre, v is commonly a substantial fraction of the speed of light, but the exact value depends on the medium. Wireless propagation is also very fast over factory-scale distances; wireless delay is often dominated by medium access, retransmission, and scheduling rather than propagation.

Jitter is variation in packet delay. If the delays of successive packets are T₁, T₂, ..., T_n, a simple sample standard deviation is

where T̄ is the sample mean and n is the number of observations. This statistic describes variation but does not guarantee a worst-case bound. Real-time engineering often requires a maximum delay or a specified percentile in addition to average latency and jitter.

Throughput is the successful useful data rate. Goodput excludes protocol overhead and retransmitted data. Packet loss is the fraction of transmitted packets that do not arrive correctly. For N_sent sent packets and N_lost lost packets,

The result is dimensionless and is often expressed as a percentage.

Availability describes the proportion of time that a service is operational. In a simple two-state model with mean time between failures (MTBF) and mean time to repair (MTTR),

This is an approximate steady-state model. It assumes statistically meaningful average failure and repair times and does not capture common-cause failures, planned maintenance, or degraded service.

1.2 Determinism and control-loop timing

A network is deterministic when its timing behaviour is sufficiently predictable for the application. Determinism does not necessarily mean zero delay; it means that delay and variation are bounded or otherwise characterized.

A sampled control loop has a sampling period T_s. The communication and computation delays consume part of the control-cycle budget. If a sensor value is sampled at time t_k, delivered after delay T_d, processed, and used by an actuator, the command may affect the plant later than intended. Excessive or variable delay can reduce phase margin and degrade stability, even when every individual packet arrives correctly.

A practical timing budget can be written as

where each term represents a stage in the cycle. This is a budget equation rather than a universal control-stability criterion. The acceptable value depends on plant dynamics, controller design, sampling frequency, and delay variation.

1.3 Synchronization

Synchronization aligns the clocks of distributed devices. It is needed for coordinated motion, timestamp correlation, distributed measurement, scheduled transmission, and event reconstruction. Clock synchronization is not the same as transporting data: a network can have high throughput while its device clocks remain poorly aligned.

Network Time Protocol (NTP) is generally suitable for less demanding time distribution. Precision Time Protocol (PTP), standardized in IEEE 1588, can achieve much finer synchronization when hardware timestamping, suitable switches, and a controlled network are used. Industrial profiles specify details such as clock roles, message rates, and permitted accuracy.

If a clock has frequency error ρ in parts per million (ppm), its approximate time error over interval t is

For example, a 20 ppm error over one hour gives 20 × 10^(−6) × 3600 = 0.072 s, or 72 ms, if no correction occurs. This illustrates why periodic synchronization is required for coordinated systems.

2. Wired and wireless communication

2.1 Wired media

Twisted-pair copper is common in industrial Ethernet because it is relatively inexpensive and can support power delivery, shielding, and convenient connectors. Its limitations include electromagnetic interference, grounding problems, distance limits, and possible damage in moving or chemically aggressive environments.

Fibre-optic communication uses light through glass or polymer media. It provides electrical isolation, high bandwidth, long distance, and strong immunity to electromagnetic interference. Installation and termination can cost more, and optical components require appropriate handling.

Industrial cabling must be selected for temperature, vibration, flexing, oil, chemicals, bending radius, ingress protection, and connector retention. Shielding is effective only when grounding and bonding are engineered correctly. Poorly designed shields can create unwanted current paths rather than eliminate interference.

2.2 Wireless media

Industrial wireless systems include Wi-Fi, Bluetooth Low Energy, IEEE 802.15.4-based systems, proprietary sensor networks, and cellular systems. Wireless reduces cabling and supports mobile assets, rotating equipment, temporary instrumentation, and difficult-to-reach locations.

Wireless introduces shared-medium contention, interference, multipath fading, obstacles, changing propagation conditions, and dependence on antenna placement. A link budget is often used to assess feasibility. In decibel form, a simplified received-power relationship is

where P_t is transmit power, G_t and G_r are antenna gains, and L represents path and other losses. All quantities are expressed in decibel-compatible units. The equation is a planning approximation; real installations require margin for fading, interference, cable loss, and regulatory limits.

Wireless is not automatically unsuitable for control. It is appropriate when the technology, coverage, redundancy, scheduling, and failure response meet the control requirement. Conversely, wired Ethernet is not automatically reliable if it has a single unmanaged switch, poor cabling, or no recovery plan.

3. Fieldbuses and industrial Ethernet

3.1 Fieldbus principles

Fieldbuses were developed to replace large point-to-point wiring systems with shared digital communication. Examples include CAN and CANopen, PROFIBUS, Modbus RTU, DeviceNet, Foundation Fieldbus, and HART. Their strengths include mature device profiles, predictable small messages, low cost, and long industrial experience.

Fieldbus design involves physical layer rules, addressing, media access, frame format, error detection, and application profiles. CAN, for example, uses message arbitration in which a dominant bit can win access without destroying the frame. Modbus RTU is a request-response protocol commonly used over serial links; it is simple but does not by itself provide modern authentication or strong semantic modeling.

Fieldbuses can remain excellent choices for legacy equipment and small control islands. Their limitations may include modest bandwidth, vendor-specific extensions, limited native security, and difficult integration with modern information systems.

3.2 Industrial Ethernet

Industrial Ethernet uses Ethernet-based physical and data-link technologies adapted for automation. Examples include PROFINET, EtherNet/IP, EtherCAT, Modbus TCP, and Ethernet POWERLINK. They generally provide higher bandwidth and easier integration with standard IT networks than traditional serial fieldbuses.

However, ordinary Ethernet is not inherently deterministic. Conventional Ethernet uses switched forwarding and may experience queueing and congestion. Industrial variants improve timing through scheduling, specialized frame handling, synchronization, redundancy, or tightly engineered network topologies.

A network design should distinguish:

• Physical layer: cables, fibre, radio, connectors, signalling, and bit transmission.
• Data-link and network behaviour: framing, switching, addressing, error detection, and access control.
• Transport behaviour: segmentation, retransmission, and end-to-end delivery where applicable.
• Application protocol: meaning and structure of data, commands, alarms, and device services.
• Information model: relationships, types, units, state, and context of industrial objects.

This distinction prevents a common error: treating “Ethernet,” “OPC UA,” and “MQTT” as interchangeable technologies. Ethernet is primarily a networking foundation; OPC UA and MQTT are higher-level communication technologies with different purposes.

4. Quality of service and network engineering

Quality of service (QoS) is the management of traffic so that important flows receive appropriate performance. QoS mechanisms include classification, priority queues, traffic shaping, policing, admission control, bandwidth reservation, and redundancy.

A simple priority scheme may assign control traffic a higher priority than video or software updates. Priority alone does not guarantee performance: if high-priority traffic is excessive, lower-priority traffic can starve, and bursts can still create delay. Traffic classes should therefore be specified with rates, deadlines, maximum frame sizes, and failure behaviour.

For a periodic stream with payload size S bits and period T, the average payload rate is

If protocol overhead is H bits per message, the link load is more accurately estimated as

This assumes one message per period and ignores retransmissions and lower-layer encoding overhead. Multiple streams must be summed and compared with usable capacity, not merely nominal link speed.

Queueing is a major source of unpredictable delay. A basic utilization measure is

where λ is mean arrival rate and μ is mean service rate, both in messages per second. The simple relationship applies to a basic queueing model and does not by itself predict industrial worst-case delay. As utilization approaches one, delay commonly increases sharply; engineered real-time systems maintain margin and use bounded scheduling rather than relying only on average utilization.

Redundancy can use parallel links, ring recovery, dual networks, duplicated messages, or redundant devices. The design must specify whether redundancy protects against a cable failure, a switch failure, a power failure, or a common software/configuration fault. Redundancy also increases testing and configuration complexity.

5. OPC UA and MQTT

5.1 OPC UA

OPC Unified Architecture (OPC UA) is a platform-independent framework for secure, structured industrial interoperability. It supports client-server communication and publish-subscribe patterns, rich information models, device and service discovery, methods, events, historical data, authentication, encryption, signing, and access control.

Its key contribution is semantic interoperability. A temperature value can be represented not merely as a numeric tag, but as an object with a type, engineering unit, timestamp, quality status, relationships, and permissions. This supports integration across vendors and organizational levels.

OPC UA is not automatically a hard real-time control protocol. Its suitability depends on the selected transport, implementation, network, deadlines, and application. Tight motion loops generally use specialized real-time protocols, while OPC UA is often valuable for supervisory control, configuration, diagnostics, and information integration.

5.2 MQTT

MQTT is a lightweight publish-subscribe messaging protocol designed around a broker. Publishers send messages to topics; subscribers receive messages for topics to which they subscribe. This decouples producers and consumers in space and time and is useful for telemetry, gateways, remote monitoring, and cloud integration.

MQTT quality-of-service levels are commonly described as:

“Exactly once” at the MQTT protocol exchange does not automatically mean exactly-once physical action. A motor command must be designed to be idempotent or protected by sequence numbers, state checks, authorization, and safe command handling. MQTT security commonly uses TLS, client authentication, access control lists, and controlled broker deployment.

OPC UA and MQTT can complement one another. An edge gateway may expose machine semantics through OPC UA locally and publish selected, normalized telemetry through MQTT to enterprise or cloud systems. This arrangement should include buffering, timestamps, data-quality flags, and clear ownership of commands.

6. Time-sensitive networking

Time-sensitive networking is a family of IEEE 802.1 standards that extends Ethernet with mechanisms for bounded latency, low jitter, synchronization, and reliability. Important mechanisms include:

• Time synchronization: a shared time base, often using IEEE 802.1AS.
• Scheduled traffic: transmission windows coordinated with a common schedule, associated with time-aware shaping.
• Credit-based shaping: controls certain traffic classes to limit burst effects.
• Frame preemption: allows selected urgent frames to interrupt lower-priority transmission.
• Stream reservation and policing: reserves and controls resources for identified flows.
• Redundancy: can duplicate traffic over independent paths and eliminate duplicates at the receiver.

TSN does not transform an arbitrary network into a deterministic system by configuration alone. Engineers must validate switch support, endpoint behaviour, schedule feasibility, clock accuracy, topology, frame sizes, traffic classes, and failure recovery. TSN is attractive because it can converge operational technology and information technology on a common Ethernet foundation, but legacy devices and unmanaged traffic can still undermine guarantees.

7. 5G and emerging 6G connectivity

Private 5G networks can provide wide-area industrial coverage, device mobility, network slicing options, high capacity, and managed quality of service. Industrial use cases include automated guided vehicles, mobile robots, remote inspection, video, and large sensor populations. 5G alone does not guarantee a control deadline: radio scheduling, core-network design, handover, interference, application processing, and backhaul must be included in the timing analysis.

Industrial 5G concepts include ultra-reliable low-latency communication (URLLC), massive machine-type communication, and time synchronization support. The actual performance depends on spectrum, deployment mode, configuration, and local conditions. Private deployment can improve control and data governance but creates responsibilities for radio planning, cybersecurity, device management, and lifecycle support.

6G remains a developing research and standardization area. Proposed directions include integrated sensing and communication, stronger positioning, AI-assisted radio resource management, higher-frequency operation, and tighter integration of computing and networking. These are development goals rather than universally available capabilities. Engineers should separate standardized, validated features from promotional claims.

8. Applied case: communication architecture for a flexible assembly cell

This is an illustrative case. A flexible assembly cell contains four servo axes, a PLC, safety I/O, a vision camera, a tool controller, an operator panel, and a manufacturing execution system (MES) connection. The cell must support synchronized motion, safe stopping, recipe management, inspection, and production reporting.

Background and problem

The original cell used a serial fieldbus for drives, a separate safety network, a USB camera connection, and manual transfer of production data. Operators experienced intermittent inspection delays, and maintenance staff could not easily correlate drive alarms with production records. The plant proposed replacing all communication with a wireless cloud connection, but this approach would not satisfy the motion and safety requirements.

Analysis

The motion loop requires periodic, bounded communication and clock synchronization. Safety communication requires a certified safety function and a defined safe-state response; ordinary message encryption is not a substitute for functional safety engineering. Vision traffic requires high throughput but can tolerate a different timing profile. MES reporting is non-real-time and can tolerate buffering.

The proposed traffic classes are:

Approach and solution

The cell uses a managed industrial Ethernet network with separated logical traffic classes. Drives and the PLC use a real-time industrial Ethernet profile with synchronized clocks. Safety functions use an appropriate certified safety protocol and hardwired or validated safe interfaces where required by the risk assessment. The camera is connected through a high-capacity switch port and its traffic is shaped so that image bursts cannot disrupt motion traffic.

An edge computer collects drive diagnostics and production events. It exposes a structured OPC UA information model to engineering and supervisory applications. Selected data is published through MQTT to the plant platform using authenticated TLS connections. The edge computer buffers data during temporary broker or upstream failures and attaches timestamps, sequence numbers, and quality indicators.

The design includes separate administrative access, least-privilege accounts, signed firmware where supported, network segmentation, backup configurations, clock monitoring, and tested recovery procedures. Wireless is used for a mobile maintenance tablet and selected noncritical sensors, not for the primary servo loop.

Outcome

The architecture preserves deterministic local control while improving integration. Inspection images no longer compete equally with motion traffic, and production events can be correlated with machine timestamps. The arrangement also permits gradual migration: legacy fieldbus devices can remain behind a gateway while new equipment uses industrial Ethernet.

Learning points

1. One protocol rarely satisfies every CPS requirement.
2. Safety, security, and real-time performance are related but distinct engineering properties.
3. Edge computing can separate time-critical control from noncritical enterprise integration.
4. Traffic classification and measurement are necessary; nominal bandwidth is insufficient.
5. A migration plan should include legacy devices, testing, training, and lifecycle support.

9. Security, safety, governance, and lifecycle issues

Connecting industrial assets expands the attack surface. Threats include unauthorized access, credential theft, malware, denial of service, manipulation of process values, unsafe commands, and compromise of engineering workstations. A robust design applies defence in depth: asset inventory, segmentation, controlled remote access, strong authentication, patch and vulnerability management, logging, backups, secure configuration, and incident response.

Security controls must respect availability and safety. Patching a controller during production may be unsafe; leaving a critical vulnerability unaddressed is also unsafe. Risk assessment should consider consequences, exposure, detectability, recovery time, and compensating controls. Standards and guidance such as IEC 62443 provide a structured basis for industrial cybersecurity, while functional safety standards address hazards and risk reduction rather than general cybersecurity.

Data governance includes ownership, retention, provenance, access rights, privacy, and cross-border transfer. Industrial data may reveal production rates, equipment weaknesses, worker activity, or proprietary recipes. AI-based predictive maintenance further requires attention to data quality, model drift, false alarms, explainability, and human authority over safety-critical decisions. An AI recommendation should not silently become an actuator command without validation, constraints, and a defined fallback state.

Interoperability also has a lifecycle dimension. A protocol may be technically open but poorly interoperable if vendors implement profiles differently. Procurement specifications should identify supported standards, conformance tests, diagnostic tools, time guarantees, firmware policy, replacement parts, and cybersecurity responsibilities.

10. Best-practice design workflow

A disciplined workflow is more reliable than choosing a fashionable protocol.

1. Characterize the application: identify data types, deadlines, cycle times, safety functions, mobility, distance, and environmental conditions.
2. Define measurable requirements: specify maximum latency, allowable jitter, loss tolerance, synchronization accuracy, availability, and recovery time.
3. Separate traffic classes: distinguish control, safety, diagnostics, video, and enterprise data.
4. Choose media and topology: evaluate copper, fibre, wireless, cellular, line, star, ring, and redundant paths.
5. Select protocols by role: choose fieldbus or industrial Ethernet for device control, OPC UA for structured interoperability, and MQTT for suitable event or telemetry distribution.
6. Engineer QoS and time: configure priorities, reservations, shaping, synchronization, and admission control.
7. Validate under worst credible conditions: test startup, congestion, cable faults, wireless interference, clock loss, switch failure, and recovery.
8. Secure and document the system: maintain diagrams, addresses, credentials, configurations, firmware records, and incident procedures.
9. Monitor continuously: collect latency, jitter, loss, clock offset, link errors, queue utilization, and application quality indicators.
10. Review the lifecycle: plan spares, upgrades, decommissioning, training, and compatibility with future systems.

Chapter Summary

Communication networks are part of the control and information architecture of a cyber-physical system. Their suitability depends on more than bandwidth: latency, jitter, determinism, packet loss, availability, synchronization, security, and recovery behaviour must be matched to the application. Transmission and propagation delay provide a useful first timing model, while queueing and retransmission often dominate practical performance.

Wired media offer stability and predictable installation when properly engineered; wireless and cellular technologies provide mobility and flexible deployment but require radio planning and resilience analysis. Fieldbuses remain valuable for mature device-level applications, while industrial Ethernet provides higher integration and capacity. OPC UA supplies structured, semantically rich interoperability, whereas MQTT provides lightweight broker-based publish-subscribe communication. TSN adds timing and reliability mechanisms to Ethernet, but requires coordinated engineering and validation. 5G can support industrial mobility and broad coverage, yet its suitability for control must be demonstrated rather than assumed.

The central engineering principle is to allocate each traffic type to a communication service whose timing, reliability, semantic, safety, and security properties are appropriate. Measurement, segmentation, synchronization, redundancy, secure lifecycle management, and tested failure responses are essential to dependable CPS operation.

Key Terms

Bandwidth: Maximum nominal data-carrying rate of a communication link.

Latency: Time taken for data to travel through transmission, propagation, processing, and queueing stages.

Jitter: Variation in communication delay.

Determinism: Predictability or boundedness of communication timing.

Throughput: Rate of successfully delivered data.

Goodput: Useful application data rate excluding protocol overhead and retransmissions.

Packet loss: Fraction of transmitted packets that fail to arrive correctly.

Clock synchronization: Alignment of time bases among distributed devices.

Fieldbus: Industrial digital network designed for device and controller communication.

Industrial Ethernet: Ethernet-based technology adapted for industrial timing, reliability, and environmental requirements.

Quality of service: Techniques for managing traffic performance and priority.

OPC UA: Secure, platform-independent framework for structured industrial interoperability.

MQTT: Lightweight broker-based publish-subscribe messaging protocol.

TSN: IEEE 802.1 family of mechanisms for time-sensitive Ethernet communication.

URLLC: Ultra-reliable low-latency communication service concept in cellular systems.

Edge computing: Processing performed near data sources to reduce delay and dependence on remote systems.

Redundancy: Duplication of resources or paths to tolerate failures.

Functional safety: Reduction of risk through safety-related systems and defined safe behaviour.

Network segmentation: Division of a network into controlled zones or logical regions.

Review Questions

5 Short Answer Questions

1. Distinguish bandwidth, throughput, and goodput.
2. What four components commonly contribute to end-to-end communication delay?
3. Why is clock synchronization important in coordinated motion?
4. What is the main architectural difference between OPC UA and MQTT?
5. Why does ordinary Ethernet not automatically provide deterministic communication?

5 Long Answer Questions

1. Compare copper, fibre-optic, and wireless communication for an industrial plant, including environmental and performance considerations.
2. Explain how QoS mechanisms can protect motion-control traffic from machine-vision traffic.
3. Describe fieldbus and industrial Ethernet strengths, limitations, and appropriate applications.
4. Explain the principal mechanisms and engineering requirements of TSN.
5. Discuss how cybersecurity and functional safety should be coordinated in a connected manufacturing cell.

3 Analytical / Application Questions

1. A sensor sends 800 bits every 20 ms, with 160 bits of protocol overhead per message. Calculate the average link load and explain whether a 100 kbit/s link is sufficient before considering retransmissions or other traffic.
2. Design a communication architecture for an automated guided vehicle system that carries safety signals, navigation data, video, and maintenance telemetry. Justify media, protocols, QoS, synchronization, and failure handling.
3. A factory wants to replace a deterministic local control network with a cloud-connected wireless system. Analyze the proposal and identify the measurements and tests required before accepting or rejecting it.

Multiple Choice Questions with Answers

Question 1

Which quantity represents variation in packet delay?

A. Bandwidth

B. Jitter

C. Goodput

D. Availability

Correct Answer: B. Jitter

Question 2

A 1,000-bit frame is transmitted over a 100,000 bit/s link. Ignoring all other delays, what is the transmission time?

A. 0.001 s

B. 0.01 s

C. 0.1 s

D. 10 s

Correct Answer: B. 0.01 s

Question 3

Which technology is primarily associated with a structured industrial information model?

A. OPC UA

B. Twisted-pair cabling

C. CAN arbitration

D. MQTT broker forwarding only

Correct Answer: A. OPC UA

Question 4

In MQTT, what does QoS level 1 provide?

A. At most once delivery

B. At least once delivery, with possible duplication

C. Guaranteed physical actuation exactly once

D. No acknowledgement under any condition

Correct Answer: B. At least once delivery, with possible duplication

Question 5

What is a principal purpose of IEEE TSN mechanisms?

A. To eliminate all application-layer data models

B. To provide bounded timing and coordinated Ethernet transmission

C. To replace every safety standard

D. To increase wireless antenna gain

Correct Answer: B. To provide bounded timing and coordinated Ethernet transmission

Question 6

A message contains 800 payload bits and 160 overhead bits and is sent every 20 ms. What is the link load?

A. 40 kbit/s

B. 48 kbit/s

C. 80 kbit/s

D. 96 kbit/s

Correct Answer: D. 96 kbit/s

Question 7

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

A. Unshielded twisted pair

B. Fibre optic

C. Coaxial copper only

D. Short-range radio only

Correct Answer: B. Fibre optic

Question 8

Why can excessive and variable network delay affect a control loop?

A. It changes the physical dimensions of the plant

B. It can reduce timing predictability and degrade control performance

C. It always increases sensor resolution

D. It eliminates the need for sampling

Correct Answer: B. It can reduce timing predictability and degrade control performance

Question 9

Which arrangement best separates local control from cloud telemetry?

A. Using MQTT as the only hard-real-time servo protocol

B. Using an edge gateway for local control data and buffered telemetry publishing

C. Sending every actuator command through a public broker

D. Removing timestamps from all messages

Correct Answer: B. Using an edge gateway for local control data and buffered telemetry publishing

Question 10

What is the main function of network segmentation in an industrial system?

A. To make all devices share one unrestricted broadcast domain

B. To divide systems into controlled zones and limit unwanted access or traffic

C. To guarantee zero packet loss without configuration

D. To replace clock synchronization

Correct Answer: B. To divide systems into controlled zones and limit unwanted access or traffic