Networking and Cloud Computing Foundations

Introduction

Modern organisations depend on communication networks to exchange information, provide services, coordinate work and connect users to digital resources. A university, hospital, bank or manufacturing company may operate thousands of devices across offices, data centres, homes and cloud platforms. Networking provides the communication infrastructure; cloud computing provides a way to obtain computing resources as services.

Networking and cloud computing are closely related but are not identical. A network describes the connected systems and communication mechanisms through which data moves. Cloud computing describes an operational and commercial model in which configurable computing resources—such as servers, storage, databases and applications—are made available on demand, commonly through networks and the Internet.

This chapter establishes the vocabulary and conceptual models needed to study routing, switching, network security, virtualisation, distributed systems and cloud architecture.

Learning Objectives

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

1. Explain the purposes and business value of computer networks.
2. Identify major network components and describe their roles.
3. Compare common network topologies and transmission media.
4. Explain bandwidth, throughput, latency, jitter, packet loss and availability.
5. Use common networking terminology accurately.
6. Distinguish on-premises, cloud, hybrid-cloud and multi-cloud environments.
7. Relate infrastructure choices to organisational requirements, risks and constraints.
8. Analyse a practical networking and cloud-computing design scenario.

1. The Purpose and Scope of Computer Networks

A computer network is an interconnected collection of computing and communication devices that exchange data according to agreed rules called protocols. Networks support several fundamental purposes:

• Resource sharing: Users can access shared printers, files, databases, applications and Internet connections.
• Communication: Email, voice calls, video conferencing, messaging and collaboration systems depend on networks.
• Centralised services: Authentication, directory services, backup and monitoring can be delivered from shared systems.
• Distributed processing: Applications may divide processing among clients, servers, data centres and cloud services.
• Business continuity: Replicated systems and alternative communication paths can maintain operations during failures.
• Scalability: Organisations can add users, devices and services without rebuilding every system independently.

A network has both technical and organisational dimensions. Technically, it must deliver data reliably, efficiently and securely. Organisationally, it must support business processes at an acceptable cost and comply with legal, contractual and governance requirements.

1.1 Communication as a layered activity

Network communication is complex because it involves physical signals, addressing, routing, error control, application formats and security. Layered models divide this complexity into related functions. The OSI model contains seven conceptual layers: physical, data link, network, transport, session, presentation and application. The Internet protocol suite is usually described using fewer practical layers: link, Internet, transport and application.

Layering provides modularity. For example, a web application can use HTTP at the application layer, TCP at the transport layer, IP at the Internet layer and Ethernet or Wi-Fi at the link layer. Each layer offers services to the layer above and uses services from the layer below. Encapsulation occurs when each layer adds control information to data before transmission.

The OSI model is valuable for teaching and troubleshooting, but real implementations do not always map perfectly to its boundaries. Engineers should therefore use models as analytical tools rather than assuming that every product implements seven isolated components.

2. Core Network Components

2.1 End devices and hosts

An end device, or host, is a system that originates or receives data. Examples include laptops, smartphones, servers, printers, sensors, cameras and virtual machines. A host commonly has a network interface controller (NIC), which provides wired or wireless connectivity.

2.2 Switches

A switch connects devices within a local-area network (LAN). Most modern Ethernet switches forward frames using Media Access Control (MAC) addresses. A switch learns which MAC address is reachable through each port and forwards traffic selectively rather than sending every frame to every port. This improves efficiency and reduces unnecessary traffic.

Managed switches can support virtual LANs (VLANs), link aggregation, quality-of-service policies, monitoring and access controls. A VLAN creates a logical broadcast domain even when devices use the same physical switching infrastructure.

2.3 Routers

A router connects different IP networks and selects paths for packets. It examines destination IP addresses and consults a routing table. Routes may be configured manually or learned through routing protocols. Routers also commonly perform packet filtering, network address translation, virtual private networking and traffic classification.

A switch primarily connects devices within a local network, whereas a router connects separate networks. In practice, multifunction devices may perform both roles.

2.4 Wireless access points

A wireless access point (AP) provides radio-based network access, commonly using IEEE 802.11 technologies. The AP bridges wireless clients to a wired network or another wireless infrastructure. Deployment requires attention to radio interference, channel planning, signal strength, authentication, encryption and user density.

2.5 Modems, gateways and firewalls

A modem converts signals between a customer network and a service-provider medium such as cable, fibre or digital subscriber technology. A gateway is a broader term for a device or service that connects systems using different networks or protocols. A firewall enforces traffic-control rules between security zones. Firewalls may be network appliances, host software, cloud services or virtual functions.

2.6 Servers and network services

A server provides a service to clients. Common network services include Domain Name System (DNS), Dynamic Host Configuration Protocol (DHCP), web hosting, email, file storage, authentication, logging and time synchronisation. A server may be physical, virtual or containerised.

3. Network Classifications and Topologies

3.1 Network scope

These categories describe scale, not necessarily performance. A small LAN may have higher capacity and lower latency than a large WAN, but design and service-provider choices determine actual performance.

3.2 Physical and logical topology

A physical topology describes how equipment and cables are arranged. A logical topology describes how data flows, which may differ from the physical layout. Common topologies include:

• Bus: Devices share a common medium. Bus networks are simple conceptually but can suffer collisions, difficult fault isolation and a single shared-medium limitation.
• Star: Devices connect to a central switch or access point. This is common in Ethernet LANs; failure of one cable usually affects one device, while central-device failure can affect many devices.
• Ring: Each device connects to two neighbours. Rings can provide predictable paths, but a break may disrupt communication unless protection mechanisms exist.
• Mesh: Devices have multiple interconnections. Full mesh connectivity between n devices requires n(n-1)/2 links, so it becomes expensive as n grows. Partial mesh is more common.
• Tree or hierarchical: Multiple stars are connected in layers. This supports structured expansion but requires careful attention to uplink capacity and failure domains.
• Hybrid: Two or more topologies are combined to meet practical requirements.

Topology selection should consider cost, resilience, scalability, installation constraints, maintenance and expected traffic patterns rather than relying on a single “best” topology.

4. Transmission Media and Signals

Transmission media carry signals between network interfaces. They may be guided, using a physical cable, or unguided, using electromagnetic radiation.

4.1 Copper cable

Twisted-pair copper cable uses pairs of insulated wires twisted to reduce electromagnetic interference. Unshielded twisted pair is common in office Ethernet, while shielded variants provide additional protection in electrically noisy environments. Copper is relatively inexpensive and can support power delivery through Power over Ethernet, but its distance and interference characteristics are more limited than fibre.

Coaxial cable has a central conductor surrounded by insulation and shielding. It remains important in broadband access and some specialised systems.

4.2 Fibre-optic cable

Fibre transmits information as pulses of light through glass or plastic. Single-mode fibre supports long distances and high capacity, while multimode fibre is generally used for shorter links. Fibre is resistant to electromagnetic interference and can provide high bandwidth, but installation, connectors and optical equipment may cost more than copper.

4.3 Wireless transmission

Wireless systems use radio or microwave signals. They enable mobility and rapid deployment but are affected by distance, obstacles, interference, spectrum regulation and shared capacity. Security must be designed into wireless networks because signals extend beyond physical walls.

4.4 Signal impairment

Attenuation reduces signal strength with distance. Noise introduces unwanted energy. Interference occurs when other signals disrupt communication. Dispersion can spread signal components over time, limiting data rates in some media. Repeaters, appropriate media selection, error detection, modulation techniques and carefully designed link budgets help address these limitations.

5. Performance Concepts and Quantitative Measures

5.1 Bandwidth, throughput and goodput

Bandwidth is the maximum data-carrying capacity of a link or channel, often expressed in bits per second. It is not the same as file size, and it does not guarantee that applications will receive that rate.

Throughput is the actual rate at which data is transferred. It is influenced by protocol overhead, congestion, device processing, signal quality and competing traffic. Goodput is the useful application data rate after excluding protocol headers, retransmissions and other overhead.

For a transfer of data volume D bits completed in time t seconds, average throughput is:

The result is measured in bits per second when D is in bits and t is in seconds. This is an average measure and does not describe short-term variation.

Worked example. A backup transfers 18 gigabits in 60 seconds. Assuming the stated decimal unit, the average throughput is 18,000,000,000 ÷ 60 = 300,000,000 bits per second, or 300 Mbit/s. The link may have a higher nominal bandwidth because protocol overhead and other traffic may account for the difference.

A useful upper-bound relationship for a single transfer is the bandwidth-delay product:

Here, R is the transmission rate in bits per second and RTT is round-trip time in seconds. The result is the number of bits that could be in transit on the path. This relationship is useful for understanding why high-capacity, long-distance links may require large buffers or protocol windows. It does not by itself predict application performance.

5.2 Latency and jitter

Latency is the time required for data to travel from source to destination. Components may include processing, queuing, transmission and propagation delay. A simplified model is:

Transmission delay depends on packet size and link rate:

Both quantities must use consistent units—for example, bits and bits per second. Propagation delay depends on distance and signal velocity. Queuing delay varies with congestion and is often the dominant and least predictable component.

Round-trip time (RTT) measures travel to a destination and back. It is important for interactive applications and transport protocols. Jitter is variation in packet delay. Voice and video applications may tolerate some loss but are particularly sensitive to excessive latency and jitter.

5.3 Packet loss and error rate

Packet loss occurs when packets are discarded, corrupted beyond recovery or otherwise fail to reach the destination. Causes include congestion, faulty equipment, wireless interference and misconfiguration. Small losses may be corrected by retransmission, but retransmissions consume capacity and increase delay. Real-time traffic may experience audible or visible degradation.

5.4 Availability and reliability

Availability describes the proportion of time a service is operational and accessible according to a defined measurement. A simple availability estimate is:

MTBF is mean time between failures and MTTR is mean time to repair or restore. Both must use the same time unit. This formula assumes a repeated failure-and-repair cycle and is a model, not a guarantee.

For example, if a service has an MTBF of 9,999 hours and an MTTR of 1 hour, availability is approximately 99.99%. In practice, planned maintenance, correlated failures, dependency failures and measurement boundaries must also be considered.

Availability targets should be expressed in operational terms. A target of 99.9% annual availability permits considerably more downtime than 99.99%; therefore, service-level agreements should state the measurement period, exclusions, support response and remedies.

5.5 Capacity planning

Capacity planning estimates future demand and tests whether infrastructure can satisfy performance objectives. It should consider peak rather than only average traffic, growth, redundancy, burst behaviour and failure conditions. Overprovisioning improves headroom but increases cost; underprovisioning produces congestion and poor user experience.

6. Essential Networking Terminology

• Protocol: A defined set of rules for communication.
• Packet: A formatted unit of data at the network layer.
• Frame: A data-link-layer unit containing addressing and control information.
• MAC address: A link-layer identifier associated with a network interface.
• IP address: A logical address used for internetwork communication.
• Subnet: A logically divided IP network.
• Port: A transport-layer endpoint number identifying an application service.
• Client: A system requesting a service.
• Server: A system providing a service.
• DNS: A distributed naming system that maps domain names to records such as IP addresses.
• DHCP: A protocol that supplies hosts with configuration such as IP addresses and default gateways.
• Default gateway: A router used to reach destinations outside the local subnet.
• Routing: Selection of a path for packets between networks.
• Switching: Forwarding frames or packets within a network infrastructure.
• Firewall: A control that permits, denies or inspects traffic according to policy.
• VLAN: A logical segmentation of a switched network.
• VPN: An encrypted or logically isolated connection across an untrusted or shared network.
• QoS: Techniques for classifying and prioritising traffic.
• Redundancy: Additional components or paths intended to tolerate failures.

7. From Traditional Infrastructure to Cloud Computing

7.1 On-premises environments

In an on-premises environment, an organisation owns or directly controls computing equipment located in its facilities or a dedicated colocation site. The organisation normally manages physical security, power, cooling, hardware replacement, network connectivity, operating systems and applications.

Advantages include direct control, predictable placement, possible compliance benefits and customisation. Limitations include capital expenditure, procurement delays, facility responsibilities, specialist staffing and the risk of capacity being either insufficient or underused.

7.2 Cloud computing

Cloud computing is a model for convenient, on-demand network access to a shared pool of configurable resources that can be rapidly provisioned and released with limited provider interaction. Important characteristics include:

• on-demand self-service;
• broad network access;
• resource pooling;
• rapid elasticity; and
• measured service.

Cloud services are commonly classified as:

• Infrastructure as a Service (IaaS): Virtual machines, networks and storage are provided while the customer manages more of the operating environment.
• Platform as a Service (PaaS): The provider manages infrastructure and much of the runtime platform, allowing developers to focus on application code and data.
• Software as a Service (SaaS): The provider delivers a complete application, while the customer manages configuration, users and data responsibilities defined by the service.

These are responsibility models, not rigid product categories. Customers remain responsible for appropriate identity management, data governance, configuration and use of the service.

7.3 Public, private and community environments

A public cloud is operated for broad external customers by a provider. A private cloud is dedicated to one organisation, although it may be operated internally or by a third party. A community cloud is shared by organisations with common requirements, such as regulatory or mission needs. In practice, the important questions concern control, isolation, location, assurance, service features and cost—not labels alone.

7.4 Hybrid cloud

A hybrid cloud combines distinct on-premises and cloud environments with integration that allows data or workloads to move or operate across them. Hybrid designs may support gradual migration, data-residency requirements, local processing, disaster recovery or temporary capacity expansion.

Hybrid operation is not achieved merely by owning both a data centre and a cloud account. It requires reliable connectivity, compatible identity systems, consistent security policy, monitoring, operational processes and clearly defined data flows.

7.5 Multi-cloud

Multi-cloud uses services from two or more cloud providers. Organisations may adopt it to access specialised capabilities, negotiate commercially, meet geographic requirements or reduce dependence on one provider. However, multi-cloud can increase complexity in identity, networking, logging, skills, security policy, data transfer and cost management. It should be adopted for a clear requirement rather than as an automatic guarantee of resilience.

8. Relating Architecture to Organisational Requirements

A defensible infrastructure decision begins with requirements rather than technology enthusiasm. Important questions include:

1. Workload characteristics: Is demand stable, seasonal, bursty, latency-sensitive or computationally intensive?
2. Data sensitivity: Does the system process personal, financial, health, classified or commercially sensitive information?
3. Regulatory obligations: Are there requirements for location, retention, auditability, encryption or sector certification?
4. Availability and recovery: What are the required recovery time objective (RTO) and recovery point objective (RPO)? RTO is the maximum acceptable restoration time; RPO is the maximum acceptable data-loss interval.
5. Connectivity: Can users and systems reliably reach the service? Is offline operation necessary?
6. Cost model: Are capital expenditure, operating expenditure, licensing, egress charges and staffing understood?
7. Skills and support: Can the organisation operate the selected design securely and continuously?
8. Integration and portability: Must existing systems remain connected? Is migration to another provider realistic?

A small startup with unpredictable demand may value public-cloud elasticity. A hospital may require strong identity controls, resilient connectivity, auditable operations and carefully governed data location. A manufacturer may retain low-latency control systems locally while using cloud analytics. These are requirement-driven outcomes, not universal rules.

9. Case Study: Regional College Learning Platform

The following is an illustrative case.

Background

A regional college operates a student information system and learning platform in a campus server room. Students increasingly access online lessons from home, while staff require reliable video teaching, cloud storage and collaboration tools.

Problem

The server room has limited power and cooling capacity. Internet connectivity is adequate during ordinary periods but becomes congested during examination enrolment. Hardware replacement is slow, backups are irregular and only one administrator understands the virtualisation environment. The college must protect student records and comply with applicable privacy obligations.

Analysis

The learning platform has seasonal peaks and benefits from elastic web capacity. The student information system contains sensitive records and requires controlled access, reliable backups and auditable changes. Campus identity services are already established, but remote access is inconsistent. A complete immediate migration would create operational and integration risk.

Approach and solution

The college adopts a staged hybrid design:

1. Student records remain in a tightly controlled environment during the initial phase.
2. The public learning portal and video content are deployed using cloud services capable of scaling during enrolment and examinations.
3. A secure site-to-site connection links campus systems to the cloud environment.
4. Federated identity allows staff and students to use managed institutional accounts with multi-factor authentication.
5. Central logging records authentication, administrative actions and significant data access.
6. Backups are encrypted, tested through restoration exercises and retained according to policy.
7. Network segmentation separates student Wi-Fi, staff systems, administration and managed infrastructure.
8. Service-level objectives define acceptable latency, availability, restoration time and support response.

Outcome

The college reduces peak congestion and avoids an immediate building expansion. However, it accepts continuing responsibility for identity, data classification, configuration, vendor management and incident response. A review process monitors cloud spending, access permissions, connectivity and exit options.

Learning points

The case demonstrates that hybrid cloud is an architectural and governance arrangement, not simply a product purchase. Network capacity, identity, security, availability and organisational skills determine whether the design succeeds. Cloud elasticity solves some capacity problems but does not remove the need for planning and operational discipline.

10. Challenges, Governance and Best Practices

10.1 Security and privacy

Networks expose systems to interception, impersonation, malware, denial-of-service attacks and unauthorised access. Controls should include least privilege, strong authentication, encryption in transit and at rest where appropriate, segmentation, secure configuration, vulnerability management, monitoring and tested incident-response procedures.

Cloud adoption changes responsibility rather than eliminating it. A provider may secure facilities and core infrastructure, while the customer remains responsible for identities, permissions, workload configuration and data handling. Contracts should define notification, audit, retention, deletion and service-continuity obligations.

10.2 Resilience and dependency

Redundancy is useful only when it avoids common failure points. Two links using the same cable route are not truly independent. Two cloud regions may still share an identity provider, software dependency or operational process. Dependency mapping and failure testing are therefore essential.

10.3 Cost and sustainability

Cloud expenditure can grow through idle resources, excessive storage, data transfer and unplanned scaling. FinOps practices—such as tagging, budgets, forecasting, rightsizing and accountability—help align consumption with value. Sustainability considerations include equipment lifespan, energy efficiency, workload placement and avoiding unnecessary duplication.

10.4 Best-practice principles

• Document requirements, assumptions and service dependencies.
• Use layered security and least privilege.
• Segment networks according to trust and business function.
• Monitor performance, capacity, availability, cost and security events.
• Automate repeatable configuration and test changes before production.
• Maintain tested backups and recovery procedures.
• Design for failure rather than assuming uninterrupted service.
• Record ownership and escalation responsibilities.
• Review provider contracts, portability and exit plans.
• Train staff and periodically reassess architecture against changing requirements.

Contemporary Developments

Software-defined networking separates network control functions from hardware forwarding, enabling programmable policy and automation. Network functions virtualisation implements capabilities such as firewalls and load balancers as software. Edge computing places processing closer to users or devices to reduce latency and limit backhaul traffic. Internet of Things deployments increase the number and variety of endpoints, making identity, segmentation and lifecycle management increasingly important.

Artificial intelligence can assist with anomaly detection, capacity forecasting and incident triage, but it should not be treated as an unquestionable authority. Training data, false positives, model drift, privacy, explainability and human approval remain important. Automated changes to production networks require safeguards, testing, audit trails and rollback mechanisms.

Chapter Summary

Networks connect hosts and services so that organisations can share resources, communicate, process information and maintain operations. Switches primarily forward frames within local networks, while routers connect different IP networks. Topology, transmission medium and protocol design influence cost, resilience, capacity and performance.

Bandwidth represents potential capacity; throughput represents achieved transfer rate; latency measures delay; jitter measures delay variation; packet loss indicates failed delivery; and availability describes operational accessibility. These measures must be interpreted together because a high-bandwidth network can still provide poor service when latency, congestion, loss or outages are excessive.

On-premises infrastructure offers direct control but requires investment and operational capacity. Cloud computing provides on-demand, pooled and measurable resources, but customers retain important responsibilities. Hybrid and multi-cloud designs can meet specific requirements, yet they introduce integration, governance and operational complexity. Sound architecture begins with workload, data, regulatory, recovery, connectivity, cost and skills requirements. Security, resilience, monitoring, tested recovery and clear responsibility are essential in every environment.

Key Terms

Network: Interconnected systems that exchange data using communication protocols.

Host: A device or virtual system that sends or receives network data.

Protocol: Rules governing communication between systems.

Switch: A device that forwards data-link frames within a network.

Router: A device that forwards packets between IP networks.

Topology: The physical or logical arrangement of network components.

Bandwidth: Maximum data-carrying capacity of a communication link.

Throughput: Actual rate at which data is transferred.

Latency: Time taken for data to travel between endpoints.

Jitter: Variation in packet delay.

Packet loss: Failure of packets to reach their intended destination.

Availability: Proportion of time a service is operational and accessible.

VLAN: Logical segmentation of a switched network.

Firewall: A control that filters or inspects traffic according to policy.

On-premises: Infrastructure operated in facilities controlled by the organisation.

Cloud computing: On-demand access to pooled, configurable computing resources.

Hybrid cloud: Integrated use of on-premises and cloud environments.

Multi-cloud: Use of services from multiple cloud providers.

RTO: Maximum acceptable time to restore a service.

RPO: Maximum acceptable period of data loss measured in time.

Review Questions

Short Answer Questions

1. What are two purposes served by computer networks?
2. Distinguish the primary roles of a switch and a router.
3. Define bandwidth, throughput and latency.
4. What is the difference between a physical and a logical topology?
5. Name two characteristics of cloud computing.

Long Answer Questions

1. Explain how the OSI model and Internet protocol suite help analyse network communication. Discuss why models should not be treated as exact descriptions of every implementation.
2. Compare star, ring, bus and mesh topologies in terms of cost, scalability, fault tolerance and maintenance.
3. Explain the differences among IaaS, PaaS and SaaS, including the changing responsibilities of the provider and customer.
4. Discuss the technical and governance challenges of operating a hybrid-cloud environment.
5. Explain how bandwidth, latency, jitter, packet loss and availability jointly influence application performance.

Analytical / Application Questions

1. A company requires low-latency control of industrial machinery, elastic analytics during monthly reporting and strict protection of operational data. Propose an infrastructure approach and justify which functions should remain local and which could use cloud services.
2. A network has a nominal rate of 500 Mbit/s but transfers 25 gigabits in 60 seconds. Calculate the average throughput and identify two reasons it may be below the nominal rate.
3. An organisation plans to adopt two cloud providers solely to “guarantee resilience.” Critically evaluate this reasoning and identify the technical, operational and governance conditions needed for multi-cloud to improve resilience.

Multiple Choice Questions with Answers

Question 1

Which device primarily forwards Ethernet frames within a local-area network?

A. Router

B. Switch

C. Modem

D. Firewall

Correct Answer: B. Switch

Question 2

A link transfers 12 gigabits in 30 seconds. What is its average throughput, using decimal units?

A. 40 Mbit/s

B. 240 Mbit/s

C. 400 Mbit/s

D. 3.6 Gbit/s

Correct Answer: C. 400 Mbit/s

Question 3

Which measure describes variation in packet delay?

A. Bandwidth

B. Jitter

C. Availability

D. Goodput

Correct Answer: B. Jitter

Question 4

Which topology normally connects devices to a central switch?

A. Bus

B. Ring

C. Star

D. Full mesh

Correct Answer: C. Star

Question 5

Which cloud service model generally gives the customer the greatest control over virtual machines and operating systems?

A. SaaS

B. PaaS

C. IaaS

D. Managed application service

Correct Answer: C. IaaS

Question 6

Which statement best describes a hybrid cloud?

A. A network using only wireless devices

B. A single private data centre with no external connection

C. Integrated use of on-premises and cloud environments

D. A cloud service supplied by two departments

Correct Answer: C. Integrated use of on-premises and cloud environments

Question 7

What is the main purpose of DNS?

A. To encrypt all network traffic

B. To map names to records such as IP addresses

C. To assign MAC addresses permanently

D. To measure electrical interference

Correct Answer: B. To map names to records such as IP addresses

Question 8

If MTBF is 999 hours and MTTR is 1 hour, the availability estimated by MTBF ÷ (MTBF + MTTR) is approximately:

A. 90%

B. 99%

C. 99.9%

D. 99.99%

Correct Answer: C. 99.9%

Question 9

Which factor is especially important when selecting wireless transmission?

A. Radio interference and spectrum conditions

B. Only the colour of network cabinets

C. The number of server passwords

D. The file format used by databases

Correct Answer: A. Radio interference and spectrum conditions

Question 10

Why can multi-cloud increase operational complexity?

A. It eliminates the need for identity management

B. It guarantees identical tools across providers

C. It may require different networking, monitoring and security practices

D. It prevents all data movement

Correct Answer: C. It may require different networking, monitoring and security practices