You are viewing a free preview of this lesson.
Subscribe to unlock all 10 lessons in this course and every other course on LearningBro.
A network is what turns a collection of isolated machines into a system that can share resources, communicate, and behave as something larger than the sum of its parts. Before you can reason about topologies, hardware, protocols or security, you need a precise vocabulary for what kind of network you are dealing with and how the machines on it relate to one another.
This lesson builds two complementary classifications. The first is by scale and coverage — LAN, WAN, PAN, MAN and SAN — which answers "how far does this network reach, and who owns the infrastructure?" The second is by organisational model — client-server versus peer-to-peer — which answers "where does control and responsibility live?" By the end you should be able to read a real-world scenario (a school, a chain of shops, a single user's bag of gadgets, a data centre) and confidently classify it on both axes, justifying your choice rather than just reciting a definition.
This lesson addresses the parts of OCR H446 section 1.3.3 dealing with the characteristics of networks and the protocols that govern them, specifically:
Throughout, the emphasis is on applied judgement: OCR examiners reward candidates who select an appropriate network type for a given context and explain the trade-off, not those who merely list features.
A computer network is a set of two or more devices (nodes) connected by a transmission medium so that they can exchange data and share resources. The shared resources might be hardware (a printer, a high-speed Internet link, a block of storage), software (a licensed application served centrally), or data (files, databases, a shared calendar).
The transmission medium itself can be wired (twisted-pair copper, coaxial, or fibre-optic cable) or wireless (radio waves, as in Wi-Fi and Bluetooth, or microwave and satellite for long hops). The choice of medium strongly shapes which network type is practical: fibre's high bandwidth and immunity to electrical interference make it the natural choice for LAN, MAN and WAN backbones, while the convenience and short range of radio make it ideal for PANs and the access edge of a LAN. You will study media in more depth in the wireless-networking and communication-methods lessons; here it is enough to know that every network needs a medium, and that the medium is independent of the classifications below.
Three properties let us classify any network:
The diagram below shows how the scale-based types nest inside one another. A PAN sits inside a LAN, LANs are joined into a MAN or WAN, and the largest WAN of all is the Internet.
graph TD
INET[Internet — global WAN of WANs]
WAN[WAN — country / continent]
MAN[MAN — city-wide]
LAN[LAN — single site]
PAN[PAN — around one person]
INET --> WAN
WAN --> MAN
MAN --> LAN
LAN --> PAN
SAN[SAN — dedicated storage fabric] -.sits beside a LAN.-> LAN
A LAN covers a small, geographically contained area such as a single building, a school site, an office floor, or a home. The defining characteristic is not just that it is small but that a single organisation or individual owns and administers all of the infrastructure — the cabling, the switches, the wireless access points. Because the organisation controls the medium end-to-end, LANs can run at very high speeds.
| Feature | Detail |
|---|---|
| Coverage | A single site — building, campus, home |
| Ownership | Wholly owned and managed by one organisation/individual |
| Speed | High — typically 100 Mbps to 10 Gbps over modern cabling |
| Latency | Very low — data travels short distances through few hops |
| Hardware | Switches, routers, wireless access points, structured cabling |
| Examples | School network, office LAN, home Wi-Fi |
A WAN spans a large geographical area — a region, a country, or the globe — connecting multiple LANs together. The crucial point for the exam is ownership: an organisation rarely owns the long-distance links between its sites. Instead it leases capacity from a telecommunications provider (leased lines, fibre backbones, satellite, undersea cable). The organisation owns the LANs at each end; the link between them is rented.
| Feature | Detail |
|---|---|
| Coverage | City, country, or worldwide |
| Ownership | Long-distance links typically leased from third parties |
| Speed | Generally lower and more variable than a LAN |
| Latency | Higher — long distances and many intermediate routers |
| Hardware | Routers, leased lines, satellite links, undersea cables |
| Examples | The Internet; a retail chain joining its London and New York offices |
A MAN sits between a LAN and a WAN in scale: it covers a city or large campus, larger than a single site but smaller than a country. A university with several buildings spread across a city, or a city council connecting its libraries and offices, would use a MAN. MANs are often built on high-speed fibre owned by the city or a regional provider.
| Feature | Detail |
|---|---|
| Coverage | A city or large multi-site campus |
| Ownership | A single large organisation or a city/regional provider |
| Speed | High — usually fibre-based |
| Examples | A multi-campus university; a council's city-wide network |
A PAN is the smallest type: a network of personal devices around a single person, typically within a few metres. PANs are almost always wireless, using short-range technologies such as Bluetooth or NFC, though USB tethering is a wired example.
| Feature | Detail |
|---|---|
| Coverage | A few metres — around one person |
| Technology | Bluetooth, NFC, infrared, USB |
| Examples | Phone linked to wireless earbuds; smartwatch syncing to a phone; a fitness tracker |
A SAN does not fit neatly on the size scale because it is classified by purpose rather than reach. It is a high-speed, dedicated network whose only job is to connect servers to consolidated, block-level storage so that the storage appears to each server as if it were a locally attached drive. SANs live in data centres and keep storage traffic off the ordinary LAN.
| Feature | Detail |
|---|---|
| Purpose | Provides shared block-level storage that looks locally attached |
| Speed | Very high — Fibre Channel or high-speed Ethernet, often 8-32 Gbps |
| Use case | Data centres, enterprise storage, server farms, virtualisation hosts |
| Key benefit | Centralised storage management, high availability, fast access |
A frequent point of confusion is the difference between a SAN and NAS (Network Attached Storage). They sound similar but operate at different levels. A SAN presents block-level storage — raw disk blocks — over a dedicated high-speed fabric, so a server treats it as if it were its own internal hard drive and puts its own file system on top. NAS, by contrast, is a single appliance attached to the ordinary LAN that serves files (not raw blocks) over standard protocols, so it is really a file server in a box. A SAN is a network dedicated to storage; NAS is a device that lives on the existing network. SANs are faster and more scalable but far more expensive and complex; NAS is cheap and simple, which is why small offices use NAS while data centres use SANs.
| Aspect | SAN | NAS |
|---|---|---|
| What it serves | Raw block-level storage | Complete files |
| Network used | Dedicated storage fabric (Fibre Channel / iSCSI) | The ordinary LAN |
| Appears to server as | A local disk | A network drive/share |
| Cost & complexity | High | Low |
| Typical user | Enterprise data centre | Home / small office |
| Type | Typical reach | Who owns the links | Relative speed |
|---|---|---|---|
| PAN | A few metres | The individual | Low-moderate |
| LAN | One site | One organisation | High |
| MAN | A city | One org / city provider | High |
| WAN | Country / global | Leased from telecoms | Lower, variable |
| SAN | Within a data centre | One organisation | Very high |
The Internet is the largest WAN in existence — a global "network of networks" in which independently owned networks agree to interconnect using a common protocol suite (TCP/IP). No single body owns it; coordination bodies such as ICANN (names and addresses), the IETF (protocols) and the W3C (web standards) manage the shared rules, but the cables and routers belong to thousands of separate operators.
A point examiners test repeatedly: the Internet is not the World Wide Web. The Internet is the infrastructure — the physical cables, routers and addressing scheme that move packets between machines. The Web is just one service that runs over that infrastructure, using HTTP/HTTPS to deliver linked documents. Email, file transfer, video streaming and online gaming are other services that use the same Internet but are not "the Web". The web-technologies lesson develops this distinction further.
It helps to picture how the Internet is physically assembled, because it explains why it is a WAN and why no one owns it. At the core sits the Internet backbone: extremely high-capacity long-distance links (continental fibre runs and undersea cables) operated by large Tier 1 providers who exchange traffic with one another for free under "peering" agreements. Smaller regional Tier 2 providers connect to the backbone, and the Internet Service Providers (ISPs) that ordinary homes and businesses pay connect to those. Networks meet and swap traffic at shared facilities called Internet Exchange Points (IXPs). When you load a page, your packets typically travel: your LAN -> your ISP -> up through the provider hierarchy -> across the backbone -> down to the destination's network. Because each link in that chain is owned by a different organisation, the Internet is the canonical example of a WAN built from leased and interconnected infrastructure rather than a single owner's cables.
This layered ownership is also why cloud computing is best understood as networking, not magic. A "cloud" service is simply a client-server arrangement in which the server lives in someone else's data centre and you reach it across the WAN. The data centre itself runs LANs and SANs internally; you, the client, see only an address to connect to. So the same three classifications in this lesson — scale, the Internet-vs-services distinction, and the client-server model — together explain what "the cloud" really is.
An intranet is a private network internal to one organisation that deliberately uses the same technologies as the Internet — TCP/IP, web servers, browsers — but restricts access to authorised members only. Staff get the convenience of web-style tools (an internal portal, document libraries, HR self-service) without exposing them to the outside world. A firewall separates the intranet from the public Internet.
An extranet extends a portion of an intranet to selected external users — suppliers, partners, or clients — usually over a secured connection. A manufacturer might give a parts supplier extranet access to its stock-ordering system without granting access to everything else on the intranet.
graph LR
subgraph Public
I[Internet — anyone]
end
subgraph Org[Organisation]
IN[Intranet — staff only]
EX[Extranet zone — partners + staff]
end
I -->|firewall| EX
EX --> IN
The second classification asks how responsibility is distributed. This is independent of scale — a LAN can be either model, and so can a WAN.
In a client-server network, one or more dedicated servers provide services — authentication, file storage, email, web pages, printing — and the other machines act as clients that request those services. Control is centralised: the server decides who may log in, what they may access, and holds the authoritative copy of shared data.
graph TD
S[(Server — authentication, files, backup)]
C1[Client 1] --> S
C2[Client 2] --> S
C3[Client 3] --> S
C4[Client 4] --> S
| Feature | Detail |
|---|---|
| Server | Dedicated machine managing resources and answering requests |
| Client | A device that requests services |
| Control | Centralised — administered from the server |
| Security | Easier to enforce — access rights set centrally |
| Backup | Centralised — one authoritative copy to back up |
| Cost | Higher — server hardware plus skilled administration |
| Examples | School networks, corporate file servers, websites, email systems |
"Server" is a role, not a single machine, and one organisation typically runs several specialised servers — a file server (shared storage), an authentication/domain server (verifies logins and enforces access rights), a web server (serves pages), a mail server, a print server, and a DHCP server (hands out IP addresses, covered in the TCP/IP-in-detail lesson). A useful refinement is the distinction between a thick client and a thin client. A thick client does most of the processing itself and uses the server mainly for storage and login; a thin client is a minimal terminal that pushes almost all processing onto the server and is little more than a screen, keyboard and network connection. Thin clients are cheaper to buy and easier to manage centrally (a school can replace a broken one in seconds), but they demand a powerful server and a reliable network, since the client can do nothing if the server is unavailable. This is the central trade-off of the client-server model in miniature: convenience and central control purchased at the price of dependence on the server.
In a peer-to-peer network there is no dedicated server. Every node is a peer of equal standing that can act as both client and server, sharing its own resources directly with others. There is no central authority, which makes P2P cheap and resilient but harder to secure and manage.
graph LR
P1[Peer 1] --- P2[Peer 2]
P2 --- P3[Peer 3]
P3 --- P4[Peer 4]
P4 --- P1
P1 --- P3
P2 --- P4
| Feature | Detail |
|---|---|
| Dedicated server | None — every node shares equally |
| Control | Decentralised — no single point of control or failure |
| Security | Harder — each node manages its own access |
| Backup | Each node is responsible for its own data |
| Cost | Lower — no server to buy or administer |
| Scalability | Can degrade as the number of peers grows |
| Examples | Home file/printer sharing, BitTorrent, some blockchain networks |
| Feature | Client-Server | Peer-to-Peer |
|---|---|---|
| Central management | Yes | No |
| Security | Easier to enforce centrally | Harder — distributed responsibility |
| Reliability | Server is a single point of failure | No single point of failure |
| Performance | A powerful server absorbs load | Depends on every peer's capacity |
| Cost | Higher (server + admin) | Lower (no dedicated server) |
| Scalability | Good — upgrade the server | Limited — congestion as peers grow |
| Backup | Centralised | Distributed, per device |
The exam-critical skill is matching the model to the requirement:
The bandwidth argument for P2P distribution can be made concrete. Suppose a 4 GB file is wanted by 1,000 users. In a pure client-server model with one server limited to an upload link of 100 Mbps, the total outbound bandwidth available to satisfy everyone is fixed:
server capacity=100 Mbps shared by all clientsso each client averages only 100 / 1000 = 0.1 Mbps and the crowd is starved. In a peer-to-peer model, every peer that has finished downloading a piece can upload it to others. If each of the 1,000 peers contributes even 5 Mbps of upload, the aggregate distribution capacity becomes:
— fifty times the lone server's link, and it grows as more peers join rather than shrinking. This is precisely why protocols such as BitTorrent use P2P for popular large files: the very thing that overwhelms a central server (many simultaneous downloaders) becomes the source of the network's strength.
It is worth stepping back to see why the scale classifications matter in practice, rather than memorising them as isolated facts. As a network's geographical reach grows, four measurable characteristics move together in predictable ways, and understanding these relationships lets you reason about any unfamiliar scenario.
| Characteristic | LAN (small) | WAN (large) | Why it changes with scale |
|---|---|---|---|
| Bandwidth | High | Lower / costly | Long-distance high-capacity links are expensive to build and lease |
| Latency | Very low | Higher | Signals travel further and pass through more routers (more hops) |
| Error rate | Low | Higher | More distance and more equipment means more chances for corruption |
| Security exposure | Containable | Greater | Data crosses third-party infrastructure outside your control |
| Cost of ownership | One-off | Ongoing rental | You buy LAN kit once; WAN links are rented continually |
The underlying principle is control. On a LAN you control the entire path, so you can guarantee speed and security; on a WAN your data crosses cables you neither own nor trust, so bandwidth costs money, latency is inherited from physics and routing, and security must be added (encryption, VPNs) because the medium gives you none for free. This single idea — that owning the medium gives you control, and relinquishing it forces you to compensate — explains almost every difference in the tables above and recurs throughout the networking unit. When the network-security lesson stresses encrypting data "in transit", it is responding directly to the loss of control that comes with WAN-scale communication.
A useful corollary: bandwidth and coverage tend to trade off against each other for a fixed budget. The same money buys either a very fast link over a short distance or a slower link over a long one. Designers therefore build networks hierarchically — fast, owned LANs at each site, joined by slower, leased WAN links only where distance forces it — getting high performance locally while paying for long-distance capacity only where unavoidable. You will see this hierarchy again as the "hierarchical star" in the topologies lesson.
A regional college operates from three campuses across one city. Each campus has its own network of staff and student computers. The college wants staff at any campus to log in with the same credentials and access shared files, with all coursework backed up centrally each night.
(a) Identify the type of network, classified by scale, that connects the three campuses. (1) (b) State whether the college should use a client-server or peer-to-peer model and justify your choice with reference to the scenario. (4) (c) The college also lets approved external examiners view a restricted set of student work from off-site. Name the kind of network arrangement this describes and explain how it differs from the college intranet. (3)
(a) The three campuses are across one city, so this is a MAN (Metropolitan Area Network).
(b) The college should use a client-server model. There is a dedicated server, so staff can log in with the same credentials at any campus, and files are stored centrally. Backups happen on the server each night. This is better than peer-to-peer because peer-to-peer has no central server.
(c) This is an extranet. It lets outside people such as examiners see some student work. It is different from the intranet because the intranet is only for staff inside the college.
(a) Because the network links three sites spread across a single city — larger than one building but smaller than a country — it is a MAN (Metropolitan Area Network).
(b) The college should adopt a client-server model. The scenario requires single sign-on across campuses: a centralised authentication server can hold one set of credentials and validate a staff member logging in at any site, which a peer-to-peer model cannot do because there is no central authority to check against. It also requires shared files accessible everywhere and centralised nightly backup; storing the authoritative copy of data on a dedicated file server means there is one location to back up, and access rights can be administered consistently for all three campuses from one place. A peer-to-peer model would scatter files and logins across hundreds of individual machines, making consistent access control and reliable backup impractical at this scale. The trade-off is higher cost — server hardware and skilled administration — and the server becoming a single point of failure, which the college would mitigate with redundancy, but for an organisation of this size the centralised control clearly outweighs the cost.
(c) This describes an extranet. An extranet extends a controlled slice of the organisation's internal systems to selected, authenticated external users — here, approved examiners — usually over a secured connection from off-site. It differs from the intranet, which is a wholly private network accessible only to members of the college from within: the intranet is sealed off from outsiders by the firewall, whereas the extranet deliberately opens a limited, permission-controlled window through that boundary so that named external parties can reach just the resources they have been granted, and nothing more.
The Mid-band response earns the marks for naming the correct network type and model, but its justification in (b) lists generic features ("peer-to-peer has no central server") rather than tying each feature to the college's specific requirements, capping it below full marks for AO2. The Top-band response explicitly connects single sign-on, shared files and centralised backup each to a property of the client-server model, acknowledges the cost/single-point-of-failure trade-off, and in (c) draws a precise contrast between external authenticated access (extranet) and internal-only access (intranet) — exactly the applied reasoning the AO2 marks reward.
Exam Tip: Lock down the two-part definitions. LAN = small area + single-site ownership; WAN = wide area + leased links. For client-server vs P2P scenario questions, write your answer as a chain of "the scenario needs X, therefore feature Y of model Z" — never a bare feature list. And keep "model" (who controls) firmly separate from "topology" (how it is wired); markers penalise candidates who blur them.