Network architecture

Network purposes and types

Modern digital infrastructure

Network devices and their functions

Transport and application protocols

The TCP/IP model

A **network** is a collection of connected devices that can exchange data. Through networks, devices can communicate and access services. They can also share resources such as files, printers and internet connections.

### Comparing network types

Networks vary in geographical reach, who owns them and what they are used for. A VPN differs from the other types below: it describes a protected logical connection, not a particular geographical size.

| Network type | Purpose and defining characteristics | Example |
|---|---|---|
| **Local area network (LAN)** | Connects devices in a limited area, such as a building or site. Usually, one organisation or person controls it. Connections can be wired or wireless. | Computers and a shared printer in a design workshop. |
| **Wide area network (WAN)** | Connects devices or LANs across a large geographical area. Rather than using infrastructure entirely owned by one organisation, connections commonly rely on telecommunications providers. | A repair business connecting its regional branches. |
| **Personal area network (PAN)** | Connects devices over a short distance around an individual, using wired or wireless connections. | A person's phone exchanging data with a nearby fitness sensor. |
| **Virtual private network (VPN)** | Creates a protected connection across another network, commonly the internet. Traffic travels between endpoints through an encrypted tunnel, reducing exposure to others on the underlying network. | A technician securely accessing a workshop's internal services from home. |

The figure contrasts a PAN around one person with a LAN within a workshop. It also shows a WAN between branches and a VPN tunnel across the internet. These categories can overlap: a remote device can use a VPN to access services on a LAN through a WAN.
{{b9aed8b70a11f96aae2ee3a78e00beb5ea5b25ba}}

### Characteristics in practice

With a LAN, the owner generally has more direct control over equipment and connections than with a WAN. A WAN covers greater distances and relies more heavily on external connections. That reliance can introduce delays and additional costs.

A PAN focuses on convenient connections between personal devices within a small area. By contrast, a VPN provides protected access over an existing connection. It doesn't replace that connection or guarantee that either endpoint is secure. Encryption and routing through a VPN service can also add processing and delay.

**Key terms:** local area network (LAN), wide area network (WAN), personal area network (PAN), virtual private network (VPN).

Four-panel schematic comparing network purposes: PAN connects a phone and fitness sensor around one person; LAN connects computers and a shared printer within a workshop; WAN links regional branches; VPN shows a technician at home using an encrypted tunnel across the internet to workshop internal services. Label each network type and its boundary or connection.

Digital infrastructures supply the connections and computing resources that digital services need. They often work together. A mobile device, for example, might connect through a mobile network to reach a cloud service over the internet.

### Purposes, benefits and limitations

| Infrastructure | Meaning and purpose | Benefits | Limitations |
|---|---|---|---|
| **The internet** | A global network of interconnected networks through which devices and services communicate using shared protocols. | Allows communication and access to services across organisational and geographical boundaries. | Access relies on connectivity and service providers. Performance varies, and connected services face security threats. |
| **Cloud computing** | Delivers computing resources, such as processing, storage and applications, over a network on demand. | Resources can scale with demand. Users can access services remotely without buying and maintaining all the necessary local hardware. | Users depend on connectivity and the provider. They face ongoing costs and outages, with less direct control over data and infrastructure. |
| **Distributed systems** | Collections of networked computers that coordinate their work to deliver a service or solve a problem. | Computers can share the work; replication can help keep a service running when one computer fails. | Coordinating the computers is complex. Communication delays and network failures can disrupt work or leave copies of data inconsistent. |
| **Edge computing** | Processes data close to where it is generated or used, instead of sending all of it to a distant computing centre. | Reduces response time and sends less data over wider networks. Some processing can continue if the wider connection is lost. | Edge devices may have limited processing and storage. Many separate devices need maintenance and security. |
| **Mobile networks** | Wireless communications infrastructure that connects mobile devices through a provider's coverage areas. | Lets users communicate while moving, without each device needing a fixed wired connection. | Coverage and congestion, along with signal conditions, affect reliability and speed. Access may involve data charges. |

In the figure, a mobile device communicates with a cloud service through a mobile network and the internet. A distributed system within the cloud service shares work between networked computers, while edge computing handles suitable work near the mobile device.
{{3089e017b1aa98e82256bf3f5c2020fe1bbf2053}}

### Related, but not interchangeable

The internet provides connectivity; cloud computing provides computing resources accessed remotely. A distributed system describes how multiple computers cooperate. Those computers may be in a cloud service or elsewhere. Edge computing describes where processing happens, whereas a mobile network describes how mobile devices obtain connectivity.

These benefits depend on the setup. To tolerate failures, a distributed system needs suitable replication and coordination. Edge computing supports continued local operation only if the necessary data and processing are available locally.

### Uses in context

The **World Wide Web (WWW)** is a system of linked resources and services accessed over networks, usually through a browser. It uses the internet, but isn't the internet itself. The internet carries other services too.

At Alder School, classroom devices connect through a LAN, and an internet connection provides access to a cloud-hosted learning service. The school gains remote access without having to maintain every service locally. An internet outage, however, can prevent access to that service. These examples show infrastructure in use; they aren't an exhaustive list.

**Key terms:** internet, cloud computing, distributed systems, edge computing, mobile networks, World Wide Web (WWW).

Relationship schematic showing a mobile device connected through a mobile network and the internet to a cloud service. Inside the cloud service, show networked computers cooperating as a distributed system. Place edge computing close to the mobile device. Use these labels to distinguish connectivity, remote resources, coordinated computing and processing location.

Network devices connect equipment and direct traffic. Some translate between systems or control access. These functions stay distinct, even when one physical unit performs several of them.

### What each device does

| Device | Function |
|---|---|
| **Gateway** | Connects networks or systems, translating between different protocols or data formats where needed. A default gateway is commonly a router used to reach other networks; the word gateway doesn't always imply protocol translation. |
| **Hardware firewall** | A physical device that allows or blocks network traffic according to security rules. It can inspect information such as the source, destination and service being accessed. |
| **Modem** | Converts signals from the form used by local networking equipment to the form required by an access connection, and vice versa. A corresponding device converts the signals at the other end. |
| **Network interface card (NIC)** | Gives a device an interface to a wired or wireless network. It sends and receives data over the local connection and normally has a MAC address. |
| **Router** | Forwards packets between networks. It selects a next hop using destination IP addresses and routing information. |
| **Switch** | Connects devices within a LAN. It normally uses destination MAC addresses to forward frames to the appropriate port. |
| **Wireless access point** | Connects wireless devices to a LAN and passes traffic between those devices and the rest of the network. |

A **MAC address** identifies a network interface for local network delivery. An **IP address** is a logical address that identifies a network interface and supports delivery across networks.

### Following a connection

The figure shows a computer connected through its NIC to a switch. A wireless device joins the same LAN through a wireless access point. To reach another network, traffic passes through a hardware firewall, a router and a modem towards the internet. This is one possible arrangement; networks don't all require this order.
{{887396b99e31d6618d6659990e3563586ecc5b71}}

A switch directs traffic within the LAN, while a router directs it between networks. Adding a wireless access point gives the LAN wireless access; it doesn't automatically provide routing or internet access. One combined networking unit may perform all of these functions. They remain different functions, though.

### Mapping devices to TCP/IP layers

To associate a device with a layer, consider the information it uses or the communication function it performs. Some devices work across several layers.

| TCP/IP layer | Device mapping |
|---|---|
| **Application** | A gateway operates here when it translates application protocols. A firewall can also operate here if it inspects application content. |
| **Transport** | A hardware firewall can filter traffic using transport information, such as port numbers that identify services. |
| **Internet** | A router uses IP addresses to forward packets. A hardware firewall can also use IP addresses to filter traffic. A default gateway belongs here when it performs routing. |
| **Network interface** | NICs, ordinary switches, wireless access points and modems support communication over local links. A switch normally uses MAC addresses; a modem handles signals. |

In the four-layer TCP/IP model, the network interface layer covers both local frame handling and physical signalling. So gateways and firewalls should not be assigned one universal layer without considering what they actually do.

**Key terms:** gateways, hardware firewalls, modems, network interface cards, routers, switches, wireless access points, application, transport, internet, network interface.

Schematic of one possible network arrangement. Show a computer with a NIC connected to a switch, and a wireless device connected through a wireless access point to the same switch. Continue towards the internet through a hardware firewall, router and modem. Label the local network LAN. Keep each function visually distinct rather than implying that every network requires separate physical devices.

A **protocol** is an agreed set of rules for communication between devices. Transport protocols manage communication between applications on different devices. Application protocols set out the exchanges used by particular services.

### Transport protocols

**Transmission control protocol (TCP)** is a connection-oriented transport protocol that delivers a byte stream reliably and in order. It sets up a connection and tracks the data sent. Acknowledgements and retransmission allow it to recover missing data, though these mechanisms add communication and processing overhead.

**User datagram protocol (UDP)** is a connectionless transport protocol. It sends datagrams without guaranteeing delivery or order, and without protection against duplication. By avoiding TCP's connection setup and built-in recovery mechanisms, UDP is useful when timely delivery matters more than recovering every missing item.

| Characteristic | TCP | UDP |
|---|---|---|
| Connection | Sets up a connection before transferring application data. | Sends without setting up a connection. |
| Delivery | Uses acknowledgements and retransmission to deliver data reliably while communication remains possible. | Does not itself acknowledge or retransmit missing datagrams. |
| Ordering | Passes data to the receiving application in the correct order. | Datagrams may arrive out of order. |
| Overhead | More control information and processing. | Less protocol overhead. |
| Suitable use | A document transfer where missing data would make the result unusable. | A live audio stream where late data may no longer be useful. |

The figure compares how the protocols handle missing data. TCP can detect that data has not been acknowledged and retransmit it; UDP has no equivalent recovery mechanism of its own. That doesn't guarantee UDP will be faster in every situation. An application using UDP may also implement its own recovery.
{{6bcdc4e766b87ee0d6189abe197a5d15ed04cb52}}

### Application protocols

**Hypertext transfer protocol (HTTP)** is an application protocol for exchanging requests and responses involving web resources. The client requests a resource or action; the server sends back a response. HTTP alone does not encrypt this exchange.

**Hypertext transfer protocol secure (HTTPS)** is HTTP carried over a connection protected using Transport Layer Security (TLS). Alongside encrypting traffic, it protects against alteration in transit and supports authentication of the server. It doesn't guarantee that the site's content is trustworthy.

**Dynamic host configuration protocol (DHCP)** is an application protocol that automatically supplies devices with network configuration. A DHCP server can lease an IP address and provide settings such as the default gateway's address. This reduces the need for manual configuration.

### Protocols working together

Application and transport protocols do different jobs. HTTP and HTTPS define web exchanges, but they do not replace the need for transport. To exchange configuration messages, DHCP uses UDP, including when a device joining the network does not yet have its own configured IP address.

When a workshop laptop joins a network, DHCP can supply its network settings. If the user then opens a protected web service, HTTPS protects the web exchange. Each is a separate communication task governed by its own protocol.

**Key terms:** transmission control protocol (TCP), user datagram protocol (UDP), hypertext transfer protocol (HTTP), hypertext transfer protocol secure (HTTPS), dynamic host configuration protocol (DHCP).

Two-panel sender–receiver sequence schematic comparing TCP and UDP. TCP shows data sent, missing data, acknowledgement and retransmission leading to ordered delivery. UDP shows datagrams sent with missing data and no built-in retransmission. Use only these labels, plus sender, receiver, TCP and UDP; do not add timings or numbered packets.

The **TCP/IP model** divides network communication tasks across four layers and describes how those layers work together. Each layer serves the one above and uses services from the one below. Agreed protocols allow different implementations to communicate.

### Roles of the four layers

| Layer, from top to bottom | Role | Examples |
|---|---|---|
| **Application** | Sets the messages and rules that network applications use to request and provide services. | HTTP, HTTPS and DHCP. |
| **Transport** | Allows applications on different devices to communicate. Port numbers distinguish application endpoints, while the chosen transport protocol determines delivery behaviour. | TCP provides reliable, ordered delivery; UDP sends datagrams without those guarantees. |
| **Internet** | Uses logical addressing and routes packets across interconnected networks. Delivery is best effort, not guaranteed. | IP addresses identify destinations; routers forward packets towards them. |
| **Network interface** | Handles communication over the current local link. This includes framing, local addressing and transmitting data as signals. | NICs, switches, wireless access points and modems support this communication. |

### How the layers interact

During **encapsulation**, data from a higher layer is placed inside a lower layer's data unit along with control information. The receiving device reverses this process: **decapsulation** removes the control information as data moves back up the layers.

In a web exchange using TCP, the sender passes application data down through transport, internet and network interface. The data crosses a local link and may pass through routers on its way to the receiver. There, processing moves upwards in the reverse order. The figure follows both directions.
{{1f64227471bad2f41e9dcbf72f80461b15cd3b38}}

1. The application layer prepares a web request.
2. Using TCP, the transport layer divides the byte stream into segments and adds information for application endpoints, ordering and reliable delivery.
3. The internet layer puts the transport data into IP packets with source and destination IP addresses.
4. The network interface layer puts packets into frames suited to the local link, then transmits them as signals.
5. A router receives a frame and examines the IP packet inside it. It forwards the packet towards its destination in a new frame appropriate to the next link.
6. At the receiving device, the network interface layer recovers the packet. The internet layer passes its contents to transport, and TCP supplies the ordered data to the application layer.

As a packet is forwarded between links, the local frame changes. Routing belongs to the internet layer. Routers do not normally need to interpret the web request to forward it.

### Where reliability comes from

The layers work together, but they don't all guarantee delivery. IP offers best-effort packet delivery, so packets can be lost or arrive out of order. To recover missing data and deliver bytes in order, TCP uses sequence information, acknowledgements and retransmission. TCP also limits how much unacknowledged data is sent so that the receiver is not overwhelmed.

Reliability depends on the protocols in use, not just on following a layered model. An application using UDP does not automatically get TCP's reliability. Even TCP cannot make communication succeed through a permanent loss of connectivity.

Separating responsibilities also supports interoperability. A web application, for example, can use the same application protocol over a wired or wireless local link because the layers below handle the link-specific work.

**Key terms:** TCP/IP model, application, transport, internet, network interface, encapsulation, decapsulation, TCP, IP.

Two aligned four-layer stacks labelled sender and receiver, ordered application, transport, internet, network interface from top to bottom. Show encapsulation moving down the sender and decapsulation moving up the receiver. Associate application data, TCP segments, IP packets and frames with the successive layers. Connect the network interface layers through a router, with one link on each side of it, labelled local link and next link. Indicate that the router forwards IP packets using new frames on the next link.

List three network devices that could be used in the studio’s network, other than the laptops, tablets and printers.

Describe how DHCP automatically configures a designer’s laptop when it joins the network.

Explain two benefits to the studio of storing its design files using cloud computing rather than maintaining all file storage on its own premises.

List three purposes of using networks within the cooperative.

Describe how a router forwards a packet from one workshop’s network towards the other workshop’s network.

Supervisors need to access internal job records from home over the internet. Explain two benefits of using a virtual private network (VPN) for this access.

State one resource that staff laptops could share over the LAN.

Describe the function of a wireless access point in connecting the recorders to the field station's LAN.

List three limitations of using cloud computing to store the recordings.

The archive requires a complete recording with its data in the original order. Explain why TCP is more suitable than UDP for uploading these recordings.

An uploaded recording passes through a router between two networks. Explain how the internet and network interface layers of the TCP/IP model work together to forward the recording through this router.

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.

A2.1 Network fundamentals

Question 1

[3]

[4]

[4]

Question 2

[3]

[4]

[4]

Question 3

[1]

[2]

[3]

[4]

[4]