Data transmissions

Network topologies and their applications

Server functions and selection

Client-server and peer-to-peer models

Network segmentation

Describe each topology in terms of how devices connect and how data travels. Explain how that arrangement affects its practical use.

A **network topology** describes how devices and their connections are arranged. This arrangement determines which routes data can take and what happens if equipment fails.

### Star topology

Each device in a **star topology** has its own connection to a central device, usually a switch. Devices communicate with one another through that switch.

If a device or its connecting cable fails, normally only that device is affected. A central switch failure, however, interrupts communication throughout the star. Another device is straightforward to add, provided the switch has a spare port.

### Mesh topology

In a **mesh topology**, devices connect to several other devices, giving data alternative routes through the network. Every device in a full mesh connects directly to every other device. A partial mesh has fewer connections.

These alternative routes improve reliability. If a link fails, data may still reach its destination, provided another usable route exists. The extra links come at a cost: equipment, installation and management all become more expensive.

### Hybrid topology

A **hybrid topology** combines different topology arrangements in one network. For example, separate groups of devices can form stars, with mesh links between the switches that connect those groups.

The figure compares a star, a mesh and a hybrid. In the hybrid, local stars connect through a mesh. Ordinary devices therefore need only one local connection, while the links between groups offer alternative routes.
{{a3d33ce8d98a4f5e38ae99dce293d1ceb0fd699c}}

### Comparing the design factors

**Data collisions** occur when transmissions overlap on a shared communication medium and interfere with each other. Whether collisions occur depends on the connection technology as well as the topology. A modern switched, full-duplex Ethernet connection has no collisions; shared wireless communication can experience them.

| Factor | Star | Mesh | Hybrid |
|---|---|---|---|
| Reliability | A failed individual link affects only that connection, but the central switch remains a single point of failure. | Alternative routes can keep communication working after a link fails. | Connections between groups can be resilient, although each local star still relies on its switch. |
| Transmission speed | Dedicated links allow fast communication, but switch capacity and busy outgoing links can cause bottlenecks. | Traffic can spread across several routes, though indirect routes require extra forwarding steps. | Fast local links can support heavy traffic if there is enough capacity between groups; weak connecting links cause bottlenecks. |
| Scalability | Adding devices is easy until ports or switch capacity run out. | Expanding a full mesh becomes difficult: each new device needs links to all existing devices. | New groups can be added without redesigning every existing local network. |
| Data collisions | Switched, full-duplex links avoid collisions; a shared-medium central device would not. | Dedicated, full-duplex wired links avoid collisions; wireless mesh links may share a medium. | Collision behaviour depends on the technology used in each part. |
| Cost | Usually economical, since each device needs one link to the centre. | Extra links and connection interfaces increase costs. | Cost depends on where extra links and higher-capacity equipment are installed. |

No topology guarantees a particular transmission speed. The capacity of the links matters, as do traffic levels and the performance of the connecting equipment.

### Practical applications

A switched star could suit a small design studio. It's affordable and easy to expand, and a damaged workstation cable does not disconnect the other workstations. If uninterrupted access is essential, though, the studio must also address the central switch's vulnerability.

A large training centre could use a hybrid, with stars inside teaching rooms and mesh connections between the switches serving different buildings. This puts the cost of alternative routes where a failure would affect the most users. The same principles apply in other settings; requirements determine the best choice, rather than the organisation's name or size alone.

**Key terms:** network topology, star, mesh, hybrid, reliability, transmission speed, scalability, data collisions, cost.

Three side-by-side schematic panels labelled star, mesh and hybrid. Star shows devices connected to a central switch; mesh shows devices with multiple interconnections and alternative routes; hybrid shows local stars whose switches are joined by mesh links. Use only the labels device, switch, link and alternative route in addition to the panel titles.

To describe a server, start with the service it provides. Then consider how to keep that service available, secure and responsive as demand grows.

A **server** is hardware or software that provides a service to other devices or programs over a network. A single physical computer can run several server services. Equally, several computers can share the work of running one service.

### Types of server

| Server type | Function |
|---|---|
| **Domain name server (DNS)** | Resolves domain names to IP addresses, allowing a device to locate a named network destination. |
| **Dynamic host configuration protocol (DHCP) server** | Automatically allocates IP addresses and supplies other network settings, such as the default gateway and DNS server address. These addresses are normally leased for a limited time. |
| **File server** | Stores files centrally so that authorised users or devices can access them across the network. |
| **Mail server** | Sends, receives and stores email, making messages available to their intended users. |
| **Proxy server** | Acts as an intermediary by receiving a client's request, then making a request to another server on the client's behalf. It can filter access or cache responses. |
| **Web server** | Receives requests for web resources and returns content such as pages, images or application responses. |

These roles don't necessarily need separate physical machines. In the figure, each role is distinguished by the service requested: name lookup goes to DNS and network configuration to DHCP. Requests for shared files go to the file server; email goes to the mail server. Intermediary requests go to the proxy server, while requests for web resources go to the web server.
{{b6e01f6b7ee1b7c53d83f400823ab2ec742a5872}}

### Function, scalability, reliability and security

**Scalability** is the ability to accommodate increased demand while maintaining an acceptable service. As user numbers rise, a service may need extra processing capacity or storage. It may also need more network capacity or additional server instances.

| Factor | What must be considered |
|---|---|
| Function | Match the service to the requirement. Use a file server for central storage and DHCP to allocate network settings. |
| Scalability | Identify what grows with demand. File and mail services may need more storage; busy web and proxy services may need more processing and network capacity. |
| Reliability | Use suitable redundancy, monitoring and recovery arrangements to reduce the risk and impact of service failure. Backups help with recovery, but they don't keep a failed service running. |
| Security | Protect access, configuration and data with appropriate authentication, permissions, updates and secure communication. Choose security measures to fit the service rather than relying on one control for every server. |

### Applying the factors

Access permissions in a shared file service prevent users from opening files they should not see. Extra storage supports growth. Backups allow lost files to be recovered, while a replacement or redundant server can reduce interruption after a hardware failure.

If DNS fails, users may be unable to locate services by name even though those services are still running. Redundant DNS provision improves availability. Protecting DNS records against unauthorised changes also helps prevent users being directed to the wrong destination.

DHCP address allocation must support the number of devices joining the network. A failure can stop new devices from obtaining configuration, although devices with valid existing leases may continue communicating for a time.

A proxy can enforce access restrictions. It can also become a bottleneck or a single point of failure, so its filtering and caching functions need sufficient capacity and reliable operation.

**Key terms:** domain name server (DNS), dynamic host configuration protocol (DHCP), file server, mail server, proxy server, web server, function, scalability, reliability, security.

A service-role diagram with users or devices on the left and six service boxes on the right. Pair name lookup with DNS, network configuration with DHCP, shared files with file server, email with mail server, intermediary requests with proxy server, and web resources with web server. Show requests travelling towards each service and responses returning; do not imply that each service requires a separate physical machine.

Start by identifying what the models share. Then compare how the roles they assign affect control, cost, reliability and performance.

Clients in the **client-server model** request services or resources from servers, which provide them. Devices in the **peer-to-peer model** can take either role: they can request resources or provide them directly to other participants.

Both models let devices exchange data and share resources. What differs is how responsibility is organised, not the network's physical shape. Either model can operate over a switched star.

### How the roles differ

A client in a client-server arrangement sends a request to a designated service provider, then receives a response. With peer-to-peer, a peer may receive data from one peer while providing data to another. The figure shows this contrast in responsibility.
{{e9bdc2bfaa557981d87dd53505634b1eb9b52bfc}}

| Aspect | Client-server | Peer-to-peer |
|---|---|---|
| Control | Central administration makes it easier to manage accounts, permissions and policies consistently. | Peers distribute control; keeping rules consistent across independently managed peers is harder. |
| Cost | Dedicated infrastructure and its administration can be expensive. | Existing devices can share resources without dedicated central service infrastructure. This reduces the initial cost. |
| Reliability | If a sole server fails, its service may be interrupted for every client. Redundancy can reduce this risk. | No single provider is essential when several peers hold the resource. If its only provider disconnects, however, the resource becomes unavailable. |
| Performance | A well-resourced server can deliver predictable service, though heavy demand may turn it into a bottleneck. | Peers can share the workload. Performance, however, depends on their availability and connection speeds, as well as competing tasks. |
| Scalability | Growth may call for more powerful servers or additional servers. | New peers may bring resources as well as demand, but coordination and discovery become harder. |
| Security and backup | Central permissions and backups can be managed consistently, but a compromised server can expose many users' data. | Individual peers handle protection and backup, which makes consistent enforcement more difficult. |

### Practical applications

Web browsing normally uses client-server communication. The browser requests a resource; the service returns it. Central control suits this arrangement because the service provider manages the content and access rules. It must, though, provide enough capacity for its users.

A peer-to-peer file-sharing application can obtain different parts of a file from several peers, spreading the transfer workload and reducing dependence on one provider. Availability still depends on peers staying connected and offering the required data.

There are other applications, and some combine the models. A system might use a central service to help participants find one another, then transfer data directly between peers. Classify the particular interaction—don't assume that an entire application uses only one model.

**Key terms:** client-server, peer-to-peer, client, server, peer, benefits, drawbacks.

Two role diagrams side by side. Client-server model: Client 1 and Client 2 each send a request to the server, the designated service provider, and the server sends a response back to each client. Peer-to-peer model: there is no designated server; Peer B requests a resource from Peer A and Peer A provides the data, while Peer C requests a resource from Peer B and Peer B provides it, so Peer B both receives data from one peer and provides data to another. The diagrams show how responsibility is organised, not the network's physical shape.

Explain segmentation by showing how it works: separating traffic cuts unnecessary communication, while controlling communication between groups supports security and resource management.

**Network segmentation** divides a network into smaller physical or logical sections. Each section can be managed separately, and communication between sections can be controlled where needed.

### Why segment a network?

There are several connected reasons to segment a network:

- **Network performance:** keeping traffic in the section that needs it reduces unnecessary processing elsewhere.
- **Reduced congestion:** when local traffic stays local, there is less competition on links connecting different sections.
- **Security:** restricting communication between sections limits access to sensitive resources and can hinder an attack's spread.
- **Efficient resource management:** separate groups can have different access rules and allocations of network capacity to match their needs.

Segmentation doesn't add link capacity. A busy link can still become congested. Communication between segments can also create a bottleneck at the device that connects them.

### Segmenting, subnetting and VLANs

| Concept | Role | Typical use |
|---|---|---|
| **Segmenting** | The general practice of dividing a network into smaller physical or logical sections. | Separate groups with different traffic patterns or security requirements. |
| **Subnetting** | Divides an IP network into smaller logical networks, each with its own network address range. Routing is required for communication between subnets. | Organise addressing and control traffic passing between groups. |
| **Virtual local area networks (VLANs)** | Divide a switched network into logically separate local networks, even when devices share physical switching equipment. Each VLAN has a separate broadcast domain. | Group devices by role without needing a separate physical switch for every group. |

A **broadcast domain** is the set of devices that can receive a broadcast transmission from one another. Keeping broadcasts inside a VLAN stops them from reaching every device on the wider switched network.

Subnetting and VLANs are related, but they aren't interchangeable. A subnet concerns IP addressing; a VLAN concerns logical separation within a switched network. Networks commonly give each VLAN a different subnet and use routing to provide communication between them.

### Worked application

A community workshop connects staff computers, visitor devices and shared tools to the same physical switching equipment. It sets up a staff VLAN, a visitor VLAN and a tools VLAN. Each has a separate subnet.

The figure shows these logical groups sharing physical switching equipment. A router connects the subnets; access rules determine which traffic can pass between them.
{{a622d3fbd3f9acc4e12d2b8a1cdf3129670274e3}}

Visitor devices can access the internet but cannot reach staff computers or shared tools. Staff computers may communicate with the tools. These rules improve security: joining the visitor group does not automatically give a device access to the other groups.

Broadcast traffic from visitor devices stays within the visitor VLAN, so staff computers and tools receive less unnecessary traffic. Local communication within a group need not cross the router. That leaves the router's capacity available for traffic that genuinely needs to move between subnets.

Resources can also be managed by group. For example, a bandwidth limit on visitor traffic can stop it from consuming capacity needed for staff work. The benefit comes from applying an appropriate policy, not simply from naming a VLAN.

### Separation needs enforcement

Creating separate subnets or VLANs does not automatically put every required security restriction in place. With unrestricted routing enabled between them, devices may still reach other groups. Access rules must allow necessary communication and block communication that is not required.

Segmentation also takes configuration and management work. Incorrect group assignments or access rules can block legitimate communication or leave unintended access open.

**Key terms:** network segmentation, segmenting, subnetting, virtual local area networks (VLANs), network performance, security, congestion, network resources.

A logical segmentation diagram showing staff computers in a staff VLAN, visitor devices in a visitor VLAN, and shared tools in a tools VLAN, all using shared physical switching equipment. Label each VLAN as having a separate subnet and connect them through a router with access rules. Show visitor devices permitted internet access but blocked from staff computers and shared tools; show staff computers permitted access to shared tools. Indicate that broadcast traffic stays within its VLAN.

Network fundamentals

State three factors, other than cost, that the company should consider when selecting a network topology.

Explain how replacing peer-to-peer file sharing with a client-server model could make access permissions for the digital patterns easier to manage.

The company proposes a hybrid topology: workstations form a switched star within each building, while the building switches have mesh connections. Explain two reasons for choosing this arrangement rather than a full mesh connecting every workstation directly to every other workstation.

State one factor the company should consider when selecting a network topology.

Describe how a staff computer accesses a rental document using the client-server model.

List three drawbacks of using a peer-to-peer model to share the company's rental documents.

Staff computers and customer devices share physical switching equipment. Explain how placing them in separate virtual local area networks (VLANs) could improve both security and network performance.

The company installs a file server for its shared documents. Explain one measure that would allow the file service to support more rental counters and one measure that would maintain access to documents if the server hardware failed.

State three purposes of network segmentation.

Explain how virtual local area networks (VLANs) and subnetting can be used together to group the devices by role while allowing only selected communication between groups.

A manager claims that introducing segmentation will improve the network without creating any new problems. Explain two potential limitations of the proposed segmentation.

Computer Science

IB Computer Science focuses on computational thinking and programming skills. Students learn about system fundamentals, network architecture, and software development, applying their knowledge to solve real-world problems through practical projects.

A2.2 Network architecture

Question 1

[3]

[4]

[4]

Question 2

[1]

[2]

[3]

[4]

[4]

Question 3

[3]

[4]

[4]