Chapter 1: Design computer network

Designing a computer network is a foundational skill for ICT Technicians in Kenya, underpinning efficient communication, data sharing, and resource management across various sectors. Understanding user requirements and the network environment ensures that the design meets operational demands while allowing for scalability and security. This chapter equips ICT professionals with the knowledge to analyze user needs, select appropriate network types and topologies, and identify essential components to build robust networks tailored to organizational goals.

1.1 User needs collections

Before any cable is bought or any switch is mounted, the technician must know what the people who will use the network need from it. User needs collection is the first phase of network design: gathering, checking and recording what users do, which applications they rely on, how many devices must connect, how much data moves, how sensitive that data is and how the organisation expects to grow. A ten-person architecture practice that sends large drawing files between workstations needs a very different network from a small shop that only runs a till and a printer, and guessing wrong in either direction wastes money or leaves users frustrated.

This topic builds the vocabulary and background needed to hold that conversation: what a computer network is, how networks are classified by size, how devices can be laid out, which components are involved, and finally how user requirements are identified, analysed and documented.

1.1.1 Introduction to computer networking

A computer network is a group of computers and other devices connected so that they can exchange data and share resources. The connection may be by cable, by radio or by a mixture of both. Networking lets several users work from one set of files, print on one printer, reach the internet through one connection and be managed from one place, which is why almost every business office, workshop and cyber café depends on one.

The main benefits that justify the cost of a network are listed below.

• Data communication: Messages, files and transaction data move between devices in seconds, so a sales clerk and an accountant can work on the same records without carrying disks between desks.
• Resource sharing: One printer, scanner, file store or internet line serves many users, which avoids buying duplicate equipment for each desk.
• Centralised management: User accounts, permissions, updates and backups are controlled from one place, which improves security and reduces maintenance time.
• Improved efficiency: Staff spend less time on repeated manual work, because information entered once is available to everyone authorised to see it.
• Access to the internet and cloud services: A shared connection gives every device a route to email, online banking portals, cloud storage and video calls.

Importance of networking for Kenyan organisations

Networks underpin the digital services that businesses in Kenya now take for granted. A shop that accepts mobile money payments needs a dependable connection at the till, a distributor with depots in two towns needs both sites to see the same stock records, and a company with staff working from home needs secure remote access to office files. Reliable power and affordable bandwidth are not always available, so a technician must design networks that tolerate interruptions and that use the available connection efficiently.

1.1.1.1 Definition of Computer Network terms

Network discussions use a fixed vocabulary. The terms below appear throughout this unit; later topics explain the more detailed ones, such as addressing, in full.

Network speeds are quoted in bits per second, while file sizes are quoted in bytes, and one byte is eight bits. A link of 100 Mbps (megabits per second) therefore moves at most 100 ÷ 8 = 12.5 megabytes per second. A capital B stands for bytes and a small b for bits, so MBps and Mbps are different units.

1.1.2 Computer Network types

Networks are classified mainly by the area they cover, and by who owns and operates them. The type affects the equipment, the transmission media, the speed that is realistic and the skills needed to run the network. A technician who understands the differences can recommend a cost-effective solution instead of copying a design that suits another site.

Wireless networks within these types

Wireless technology is not a separate size of network: Bluetooth forms a PAN, Wi-Fi forms a LAN, point-to-point microwave links can form a MAN, and cellular and satellite services carry WAN traffic. The common advantages are mobility, quick installation and no cabling in difficult buildings, and the common risks are unauthorised access and variable speed, which is why encryption and careful placement of access points matter.

1.1.2.1 LAN

A local area network connects devices within a limited area such as an office, a workshop, a hotel floor or a single compound. The organisation that uses it usually owns the cables and switches, so it controls who connects and how the network is secured.

• High speed: Wired LANs commonly run at 1 Gbps to each desk, and faster links join the switches together, which suits large file transfers and shared databases.
• Limited range: A twisted-pair Ethernet segment is limited to 100 m, so a LAN covers one building or a group of closely placed buildings.
• Controlled environment: Access is restricted to people on the premises, which makes the network easier to secure and manage.
• Low cost per connection: Ethernet cable, switches and network cards are inexpensive compared with long-distance links.
• Many device types: Desktops, laptops, printers, servers, cameras and phones can all share one LAN at the same time.

A wireless LAN, delivered by Wi-Fi access points, is a LAN that uses radio instead of cables. It is flexible and cheap to install where drilling is difficult, but it needs strong encryption to prevent unauthorised access, and its speed depends on signal strength and interference. Lodges and business centres commonly use it to give guests internet access.

1.1.2.2 WAN

A wide area network connects LANs that are far apart, such as the depots of a distribution company in different towns. The links are normally rented from telecommunications providers because an organisation cannot lay its own cable across towns.

• Extended coverage: A WAN joins dispersed sites so that they can share one stock system, one email system and one set of company policies.
• Lower speed for the cost: Distance, shared infrastructure and rental charges mean WAN links are slower than LAN links and cost more per megabit.
• Dependence on providers: The organisation relies on leased lines, fibre capacity, mobile data or satellite links supplied by others, and agrees the service level in a contract.
• Use of public infrastructure: A common low-cost option is a virtual private network that carries private traffic across the public internet in an encrypted tunnel.
• Complex management: Routing, security and fault tracing across several providers demand more specialised skills than a single LAN does.

The internet is a worldwide collection of interconnected provider networks and can be thought of as a very large WAN.

1.1.2.3 PAN

A personal area network connects devices around one person, typically within a few metres. It is built on short-range wireless technologies or on direct cables, and it is usually set up by the user without a technician.

• Bluetooth: Pairs a phone with earphones, a watch, a keyboard or a car audio system over a short distance.
• USB tethering and wired links: A phone connected by USB cable shares its mobile data connection with a laptop.
• Wi-Fi hotspot or Wi-Fi Direct: A phone creates a small wireless network that nearby devices join, which is a quick way to get a laptop online away from the office.
• Near-field communication (NFC): Devices a few centimetres apart exchange small amounts of data, as in contactless payments and tap-to-pair accessories.
• Wearable and smart devices: Fitness bands and similar devices connect to a phone, which relays their data to the internet.

1.1.2.4 MAN

A metropolitan area network covers a town or a city, joining several LANs that are within a few tens of kilometres of each other. It sits between a LAN and a WAN in reach and normally uses high-capacity fibre links or point-to-point radio links.

• Town-wide coverage: A security company with a control room and five guard posts spread around a town can link all of them into one network.
• Intermediate speed: A MAN is usually faster than a typical WAN link because the distances are shorter, but it is slower and more costly per connection than a LAN.
• Fibre backbone: Fibre optic cable carries data over long distances with little signal loss, so it is the usual choice for the main links.
• Shared or leased infrastructure: Many organisations buy MAN capacity from a provider instead of building their own, although a large organisation may own its fibre.
• Bridge between LAN and WAN: A MAN often connects town sites together and then hands traffic on to the wider WAN or the internet.

1.1.3 Network topologies

A topology is the pattern in which devices are connected. The physical topology describes the layout of cables and devices, while the logical topology describes the path data actually takes. The two can differ: a network cabled as a star may still behave as a ring. The topology chosen affects cost, speed, fault tolerance and how easily the network can grow, so it is one of the first design decisions.

Older Ethernet networks connected every device to a single shared cable called a bus. That layout is no longer used in new office LANs because one cable fault stopped every device, and it has been replaced by the star.

1.1.3.1 Star

In a star topology every device has its own cable to a central switch. The switch receives data from one device and forwards it to the intended device.

• Centralised control: Every connection passes through the switch, so monitoring and management happen in one place.
• Fault isolation: If one device or its cable fails, the others carry on working.
• Easy expansion: A new computer is added by running one cable to a free switch port, with no disruption to the others.
• Higher cabling cost: Each device needs its own cable back to the switch, which uses more cable than a shared bus.
• Central point of failure: If the switch fails, every device connected to it loses the network, so important sites use a spare switch or a second connection.

The star is the standard layout for office LANs and is the layout assumed by structured cabling.

1.1.3.2 Ring

In a ring topology each device connects to exactly two neighbours so that the connections form a closed loop, and data travels around the loop from device to device.

• Sequential data flow: Each device passes data on to the next until it reaches its destination.
• Orderly access: In a token ring a device may send only when it holds a special frame called the token, which prevents collisions.
• Fault sensitivity: In a simple single ring a cable break can cut the network, which is why many rings carry data in both directions or use two rings.
• Moderate cost: A ring needs fewer cables than a mesh but more care in installation than a star.
• Modern use: Token Ring itself is obsolete, but rings remain in use where switches or fibre links are connected in a loop. Ethernet switches cabled in a ring run a protocol that blocks one link to prevent endless loops and unblocks it if another link fails.

1.1.3.3 Mesh

In a mesh topology devices are interconnected by several paths so that traffic can be rerouted when a link fails. In a full mesh every device connects directly to every other device; in a partial mesh only the important devices do.

The number of links in a full mesh of n devices is n × (n − 1) ÷ 2. For 6 devices: 6 × 5 ÷ 2 = 15 links. For 10 devices: 10 × 9 ÷ 2 = 45 links. The cabling grows quickly, which is why full mesh is limited to a few critical devices.

• High fault tolerance: If one link fails, traffic uses another path.
• Complex cabling and configuration: Many links need many ports, cables and careful routing settings.
• High cost: The extra links and ports make full mesh expensive, so it is used only where downtime is costly.
• Wireless mesh: Wi-Fi mesh systems connect access points to each other wirelessly, so coverage can be extended through a building without running cable to each unit.
• Typical use: Core links between routers in a provider network or between important switches in a server room.

1.1.3.4 Hybrid

A hybrid topology combines two or more basic topologies in one network. Most real networks are hybrids, because different parts of a site have different needs.

• Extended star: One central switch connects to switches on each floor, and each floor switch connects its own devices in a star.
• Star with a partial mesh core: Desk connections are stars, while the main switches are linked by more than one path for backup.
• Flexible growth: A new building or floor adds one more star to the existing network without redesigning the rest.
• Targeted cost: Money is spent on redundant links only where they matter, and ordinary desk connections stay simple.
• Greater complexity: The overall layout needs a clear diagram, because a fault can sit in any one of the combined parts.

1.1.3.5 Point to Point

A point-to-point topology joins exactly two devices directly with a dedicated link.

• Direct link: The two devices have the whole link to themselves, so bandwidth is not shared.
• Wireless bridge: Two directional antennas link two buildings, such as a main office and a warehouse, without laying cable between them.
• Leased line: A provider supplies a dedicated link between two sites, and the organisation pays a fixed rental.
• Router to router link: Two routers joined by a single cable or fibre form the simplest WAN connection.
• Building block: Larger topologies are made of many point-to-point links, so the concept applies to every network.

1.1.4 Components of a computer network

Every network is built from hardware that moves data and from the computers that use it. This topic introduces the five main components of a network; the selection, installation and configuration of each are taught in later chapters of the unit.

1.1.4.1 Switches/hubs

A switch is a device with many ports that connects the devices of a LAN. It reads the destination MAC address of each frame, looks it up in a table of which device is on which port, and forwards the frame only to that port. This operates at the data link layer.

A hub is an older and simpler device that repeats every signal it receives out of all its other ports. All devices on a hub share the same bandwidth and see each other's traffic, so performance drops quickly as devices are added. Hubs are obsolete, and switches have replaced them in every new installation.

Switches are either unmanaged, which work as they come with no settings, or managed, which allow VLANs, port settings and monitoring. Some switches provide Power over Ethernet (PoE), which sends electrical power along the data cable to a device such as an access point or an IP camera. The standards provide up to 15.4 W per port (IEEE 802.3af), 30 W (802.3at) and 60 W or 90 W (802.3bt).

1.1.4.2 Routers

A router connects different networks and forwards packets between them using IP addresses. It keeps a routing table that tells it which path leads to each destination network. In an office, the router is the default gateway: any packet addressed outside the local network is sent to it, and it passes the packet on towards the internet service provider.

• Joining networks: The router sits between the office LAN and the provider's connection, and between separate internal networks.
• Path selection: Where several routes exist, the router chooses the best according to its routing table and configured rules.
• Address translation: Most routers translate private addresses to one public address so that many internal devices can share a single internet connection.
• Access device: A modem or a fibre terminal usually sits between the provider's line and the router. The modem converts signals between the digital form used by computers and the form carried by the line.
• Combined devices: Small offices often use one box that combines a router, a switch, a firewall and a Wi-Fi access point, while larger sites use separate devices.

1.1.4.3 Ports

The word port has two meanings in networking, and a technician must know both.

A physical port is a socket on a device where a cable is plugged in.

• RJ45 Ethernet port: The standard socket for twisted-pair cable, with link and activity lights that show whether the connection is working.
• SFP and SFP+ port: A slot that accepts a small module for fibre optic or copper links, which allows the same switch to use different media.
• Console port: A management port used with a laptop to configure a switch or router directly.
• WAN and LAN ports: On a router the WAN port faces the provider and the LAN ports face the internal network.
• Uplink port: A port used to connect one switch to another, often at a higher speed than the ordinary ports.

A switch port also has a speed, such as 10/100/1000 Mbps or 2.5, 5 or 10 Gbps, and the two ends negotiate the highest speed that both support.

A logical port is a number that identifies a service on a host, so that a computer can run several network applications at once. Web traffic uses port 80 (HTTP) or 443 (HTTPS), secure remote login uses 22 (SSH) and name lookups use 53 (DNS). Firewalls allow or block traffic by port number, which is why logical ports matter in security design.

1.1.4.4 Computers

Computers are the devices that people use and the servers that provide services. Each computer needs a network interface card (NIC) to connect to the network.

• Wired NIC: Connects through an RJ45 port to an Ethernet cable and is built into the motherboard of most desktops and laptops.
• Wireless NIC: Connects to a Wi-Fi access point and is built into laptops, phones and tablets.
• External adapter: A USB network adapter adds a wired or wireless connection to a computer that lacks one.
• Identity on the network: Each NIC has a unique MAC address, and the computer is also given an IP address to communicate across the network.
• Roles: A client requests services; a server provides them, such as storing shared files, managing print jobs or hosting a database. Servers and storage are covered in detail where network components are identified in the next chapter.

Other network devices also count as computers in this sense: network printers, IP cameras, point-of-sale terminals and smartphones all have a NIC and an IP address.

1.1.4.5 Transmission media

Transmission media are the paths that carry signals between devices. They are wired or wireless.

Twisted pair is cheap and easy to install. A copper run is limited to 100 m, and Cat 6a is needed for 10 Gbps over the full 100 m. Fibre optic cable carries light, so it is immune to electrical interference, gives very high speed and covers long distances. Wireless media remove the need for cables but share the air, so signal strength, walls and interference determine speed. The selection and installation of cables is taught under Computer Network materials in the next chapter.

1.1.5 Computer Network user requirements/needs

A user requirement is a statement of something the network must do or provide so that people can do their work. Gathering requirements follows three stages: identifying them, analysing them and documenting them. The documented requirements then become the input to the physical and logical design.

1.1.5.1 User requirements identification

Identification means finding out what users need. The technician talks to the people who will use the network and looks at the evidence of how they work now.

Methods used to identify requirements:

• Interviews: Short conversations with the owner, the office manager and a few typical users, asking what they do each day and what slows them down.
• Questionnaires: A short form given to many users when interviewing everyone is not practical.
• Observation: Watching how work flows, for example how a receptionist prints, scans and sends documents.
• Document review: Reading existing network diagrams, equipment lists, internet bills and fault logs.
• Site walk-through: Visiting each room to see where users sit, where cables could run and where power is available.

Questions to ask during interviews include how many people use the network now and in two years, which applications they use, whether they need internet or remote access, what they would lose if the network were down for a day, what data is confidential and what budget is available.

Requirements fall into groups.

• Functional requirements: What the network must do, such as file sharing, shared printing, internet access, email, and remote access for staff working from home.
• Performance requirements: How well it must do it, including bandwidth, low latency for video calls and minimal downtime.
• Security requirements: Who may access what, covering user authentication, access control, data encryption, firewall and antivirus protection, and the auditing of activity.
• Environmental and physical requirements: Space for equipment, reliable power, existing cabling and protection from heat, dust and humidity.
• Budget and support requirements: Purchase cost, who will maintain the network, vendor warranties, user training and the possibility of upgrades later.

1.1.5.2 User requirements analysis

Analysis turns the raw list of wishes into a clear, consistent and achievable set of requirements.

1. Group the items into functional, performance, security, physical and budget requirements, and remove duplicates.
2. Resolve conflicts. If the manager wants the shared printer reachable from every desk and the owner wants the payroll folder open to one computer only, the design must separate the two.
3. Quantify what can be measured. Convert statements such as "the internet is slow" into numbers: how many users, which applications, how many at the same time.
4. Prioritise. Label each requirement as must have, should have, could have or not now, so that the budget goes to the essentials first.
5. Check feasibility against the budget, the building and the time available, and note where a requirement cannot be met.
6. Confirm the result with the users, so that the final list matches what they actually meant.

Worked example of quantifying a requirement. A twelve-person accounting office reports these figures, which are assumed for the exercise: up to 4 video calls at once at 2 Mbps each, 12 users browsing and using cloud accounting software at 1 Mbps each, and an online backup using 5 Mbps.

• Video calls: 4 × 2 = 8 Mbps
• Browsing and cloud applications: 12 × 1 = 12 Mbps
• Backup: 5 Mbps
• Total demand: 8 + 12 + 5 = 25 Mbps
• Growth allowance of 30%: 25 × 1.3 = 32.5 Mbps

The internet package should therefore offer at least 32.5 Mbps, so the next package above that figure would be chosen.

1.1.5.3 User requirements documentation

Documentation records the agreed requirements in one document, usually called a user requirements specification, so that the client and the designer share the same understanding and the design can be checked against it later.

A good specification contains:

• Purpose and scope: What the network is for and what is included and excluded.
• Stakeholders: The people who requested and will use the network and the person who approves the work.
• Requirements list: Each requirement has a reference number, a clear description, its priority, its source and a way to confirm that it was met.
• Constraints and assumptions: Budget limits, deadlines, building restrictions and the assumptions made during analysis.
• Sign-off: The client's signature and date, and a version number so changes can be tracked.

An extract from a specification is shown below. The entries are hypothetical.

Any later change must be written down and approved before it affects the design, so that the specification always reflects what was agreed.

Chapter Summary

This chapter began by exploring how to collect user needs through understanding fundamental computer networking concepts, including key terminology and various network types such as LAN, WAN, PAN, and MAN. It then examined different network topologies like star, ring, mesh, hybrid, and point-to-point, highlighting their structural differences and applications. The essential components of a computer network were described, covering devices such as switches, routers, ports, computers, and transmission media. The process of identifying, analyzing, and documenting user requirements was emphasized as the foundation for effective network design. The chapter progressed to the development of physical and logical network designs, explaining how these designs translate user needs into practical infrastructure. An overview of network design methodologies was provided, contrasting hierarchical and flat network models, along with considerations for greenfield and brownfield site types. Site preparation activities, including floor plan design and placement of data and access points, were discussed as critical steps before implementation. Finally, the chapter outlined fundamental design goals such as scalability, availability, security, and manageability, which guide the creation of robust and efficient computer networks.