Networks, LANs, WANs and network models
A network is two or more computers or devices connected so that they can share data and resources. This note covers why devices are networked, the characteristics of local area networks (LANs) and wide area networks (WANs), the client-server and peer-to-peer models, and thin and thick clients. Paper 1 questions on this topic are almost always scenario-based: you are given an organisation and asked to describe, compare or justify, so the key skill is linking each feature to the situation.
Why network devices?
Connecting devices together brings benefits that stand-alone computers cannot offer.
- Sharing resources: one printer, scanner, storage device or internet connection can be used by everyone, which saves cost.
- Sharing data: users can access the same files and databases from any connected computer, so everyone works with up-to-date data and there is no need to copy files on removable media.
- Communication: email, messaging, video calls and collaborative documents.
- Central management: software can be installed and updated, users managed and data backed up from one place, saving technician time.
- Security and access control: user accounts and access rights can be managed centrally.
- Software licensing: a site licence can be cheaper than individual copies.
There are drawbacks too: networks cost money to set up and need specialist staff to manage; a failure of a central device can stop everyone working; and connected devices are exposed to hacking and malware spreading from machine to machine.
LANs and WANs
A local area network (LAN) is a network covering a small geographical area, such as a single building or site, where the infrastructure is owned and managed by the organisation using it.
A wide area network (WAN) is a network covering a large geographical area, such as a country or several countries, usually made by connecting LANs together using communication links provided and owned by third parties (telecommunication companies).
| LAN | WAN | |
|---|---|---|
| Area | Small: one building or site | Large: cities, countries, worldwide |
| Ownership of the links | The organisation | Usually third parties: leased lines, telephone network, satellites |
| Typical media | Copper (twisted pair) cable, fibre, WiFi | Fibre-optic lines, telephone lines, microwave, satellite |
| Data transfer rate | High | Generally lower per connection, and more variable |
| Error rate | Low | Higher, because of the distances and number of links |
| Cost | Relatively low; mostly set-up | Ongoing cost of leasing lines or paying a provider |
| Example | A school's network | A bank linking all its branches; the internet is the largest WAN |
A company with offices in three cities typically has a LAN in each office and connects them into a WAN using leased lines or a secure connection over the internet. Each LAN has a router at its edge to connect it to the wider network.
Network models
A network model describes the roles that the computers on a network play. There are two.
The client-server model
In a client-server network, one or more servers provide services, and clients request them. The server is a computer (or a program on a computer) that waits for requests, processes them, and sends back a response.
A server is a computer or program that provides a service or resource to other computers on the network.
A client is a computer or program that requests and uses a service provided by a server.
Servers are usually specialised by role:
- a file server stores users' files centrally;
- a print server manages a queue of print jobs for shared printers;
- an email server stores and forwards email;
- a web server stores web pages and sends them to browsers;
- a database server runs the database management system and answers queries;
- an authentication server checks user names and passwords when users log on.
A very large organisation may divide its network into subnetworks (for example one per department or building), each with its own local servers for heavy everyday use (files, printing) while sharing central servers for email, authentication and the internet. This reduces traffic on the main network and means a fault in one subnetwork does not stop the others.
Benefits of client-server:
- central storage means files are backed up in one place and every user sees the latest version;
- security is centrally controlled: one place to manage user accounts, passwords and access rights;
- software updates and antivirus can be pushed from the server;
- users can log on at any client and see their own files;
- it scales well: more clients can be added, and servers can be upgraded.
Drawbacks of client-server:
- servers and network operating systems are expensive to buy and maintain, and a network manager is usually needed;
- if a server fails, clients lose access to its services (a single point of failure, unless there are backup servers);
- heavy demand can make the server or the network a bottleneck.
The peer-to-peer model
In a peer-to-peer (P2P) network there is no central server. Every computer (a peer) has equal status: each can act as both a client and a server, sharing its own files and resources directly with other peers and requesting theirs.
Benefits of peer-to-peer:
- cheap and simple to set up: no server or specialist staff is needed;
- no single point of failure: if one peer goes down, the rest can still communicate;
- for file sharing across the internet (as in BitTorrent), many peers each supply part of a file, so popular files download quickly and no single server has to carry all the traffic.
Drawbacks of peer-to-peer:
- no central backup: each user must back up their own data;
- security is weaker and harder to manage because each peer controls its own sharing and access;
- files can be hard to find, and copies on different peers may be out of date;
- a peer's performance drops when others are using its resources;
- on public P2P networks, files may be illegally shared or contain malware.
| Client-server | Peer-to-peer | |
|---|---|---|
| Central server | Yes | No |
| Role of each computer | Clients request; servers provide | Every peer can request and provide |
| Security and backup | Central and consistent | Each peer is responsible for its own |
| Cost and set-up | Higher; needs a manager | Low; easy to set up |
| Failure | Server failure affects all clients | Failure of one peer affects only its own resources |
| Suits | Schools, companies, websites, banks | Small home or office networks; internet file sharing |
Thin clients and thick clients
The terms thin and thick describe how much a client does itself.
A thin client is a client that depends heavily on a server: it has little processing power or storage of its own and relies on the server to run applications, process data and store files. It needs a constant connection to the server.
A thick client (fat client) is a client that does most of its own processing and can store data and run applications locally; it uses the server mainly for storage or shared resources and can work, at least partly, without a connection.
The terms apply to both hardware and software:
- Hardware: a thin-client terminal is a cheap device with a basic processor, little memory and no hard disk, which displays a desktop actually running on the server. A thick client is a full PC.
- Software: a browser-based application whose processing is done on the server is a thin client; an application installed and run on the user's own computer (perhaps saving files to the server) is a thick client.
| Thin client | Thick client | |
|---|---|---|
| Processing done | Mainly on the server | Mainly on the client |
| Local storage | Little or none | Has its own storage |
| Needs constant connection? | Yes | No: can work offline |
| Client hardware cost | Low | Higher |
| Server requirement | Powerful server needed | Less demanding server |
| Network traffic | High: every action goes to the server | Lower |
| Security and updates | Easy: everything central | Harder: each client must be maintained |
| Risk | Server or network failure stops all clients | Each client is more independent |
A college has one campus. A university has campuses in four different cities. Describe two differences between the networks they are likely to use.
Solution
- Area and ownership: the college network covers one site, so it is a LAN with cabling and wireless access points owned by the college. The university's network connects campuses in different cities, so it is a WAN; the links between campuses are provided by a third-party telecommunications company and leased.
- Speed and reliability: the college LAN will have high data transfer rates and low error rates because distances are short; the links in the university WAN cover long distances and pass through more equipment, so data rates between campuses are typically lower and errors more likely.
A firm of 60 accountants stores confidential client records. Staff move between desks. Justify the use of a client-server model rather than peer-to-peer.
Solution
- Client records can be held centrally on a file or database server, so all staff use the same up-to-date version, and the data can be backed up from one place.
- Security is centrally controlled: the network manager sets user accounts, passwords and access rights so that only authorised staff can see confidential records. In a peer-to-peer network each user would control access to their own machine, which is much harder to enforce.
- Because user accounts are on the server, staff can log on at any desk and access their own files.
- With 60 users, a peer-to-peer network would be slow and difficult to manage; client-server scales better.
A library wants 40 computers for the public to search the catalogue and browse the web. Explain why thin clients might be suitable, and state one disadvantage.
Solution
Thin clients are cheap, because they need only basic processors and no hard disk, which matters when 40 are needed. All software runs on the server, so it is installed, updated and secured once rather than on 40 machines. No files are stored locally, so members of the public cannot leave personal data or install malware on the terminal, and a damaged or stolen terminal can be replaced cheaply.
Disadvantage: every terminal depends on the server and the network. If the server fails or the network is congested, all 40 terminals stop working or become slow.
A dentist's surgery has four computers and one printer. The owner wants to share the printer and some files with minimal cost. Recommend a network model and justify it.
Solution
Peer-to-peer. With only four computers, a dedicated server is not justified: P2P needs no server hardware, no server operating system and no network manager, so it is cheap and easy to set up. Each computer can share the printer and selected folders directly with the others.
The owner should note the trade-off: there is no central backup, so files must be backed up on each machine, and access to sensitive patient files must be set up carefully on each computer.
Defining a LAN by speed or number of computers. The defining feature is geographical area (and ownership of the infrastructure). A LAN can have thousands of computers.
"In peer-to-peer, there are no servers." More precisely, there is no dedicated central server: every peer can act as a server for its own resources.
Calling a thin client "a slow computer". A thin client relies on the server for processing and storage. Speed depends on the server and network.
Listing generic benefits. In justification questions, every point must connect to the scenario: "staff move between desks, so central log-on means they can access their files anywhere".
Questions often say "explain two benefits" or "justify your choice". Give the feature, then the consequence for the scenario. A benefit without a reason ("better security") usually scores nothing; "central security, so the manager can set access rights to restrict confidential records" scores.
For comparison questions (LAN versus WAN, thin versus thick), write each difference as a matching pair: "a LAN is ... whereas a WAN is ...". One-sided statements ("a WAN is large") often earn only half the mark.
"Describe the role of the server" means what it does: stores and manages shared resources, processes requests from clients and returns responses, controls access.
- Networking shares resources, data and communication, allows central management, but costs money and increases security risks.
- LAN: small area, owned infrastructure, fast, low error rate. WAN: large area, third-party links, slower, higher error rate.
- Client-server: servers provide services (file, print, email, web, database, authentication); clients request them. Central control and backup; server is a single point of failure.
- Peer-to-peer: no central server; every peer can be client and server. Cheap and resilient, but weak central security and no central backup.
- Thin client: depends on the server for processing and storage; cheap, easy to manage, needs the network. Thick client: does its own processing, can work offline, more costly to maintain.
- Always justify with reference to the scenario.
Practice
- State three benefits of connecting the computers in an office to a network.
- Describe two characteristics of a WAN.
- Explain the role of a server in a client-server network, and name three different types of server.
- Give two drawbacks of a peer-to-peer network compared with client-server.
- Explain the difference between a thin client and a thick client.
- A software company provides its word processor as a web application in the browser. Explain whether this is a thin-client or thick-client approach.
- Explain why BitTorrent uses a peer-to-peer model to distribute large files.
- A school network has 800 computers across several buildings. The network manager divides it into a subnetwork per building, each with a local file server, connected to central servers. Give two benefits of this arrangement.
- A design studio's staff work on very large 3D models that need heavy processing. A manager proposes replacing their PCs with thin clients to save money. Evaluate the proposal.
- A chain of 50 shops wants every till to update stock levels in a central database at head office in real time. Describe the network the chain needs, using the terms LAN, WAN, client and server, and justify the client-server model for this purpose.
Answers
- Any three: share printers and other hardware; share files and data so everyone uses the latest version; share one internet connection; communicate by email or messaging; central backup; central software installation and updates; central security and user accounts.
- Covers a large geographical area (cities or countries); uses communication links owned by third parties (leased lines, satellites, telephone networks); generally lower data rates and higher error rates than a LAN.
- A server provides services or resources to clients: it stores shared data, waits for requests from clients, processes them and sends responses, and controls access. Types: file server, print server, web server, email server, database server, authentication server.
- No central backup, so data may be lost; security and access rights must be set on each peer, so are harder to manage and enforce; harder to find files and keep copies up to date; a peer slows down when others use its resources.
- A thin client relies on the server for processing, applications and storage, has little or no local storage and needs a constant connection. A thick client does its own processing and storage and can work without a connection, using the server mainly for shared resources.
- Thin client: the browser only displays the interface and sends user actions; the processing and storage happen on the company's servers, and it needs a connection to work.
- Each peer that has (part of) the file also uploads pieces to other peers, so the load is spread across many computers. No single server has to supply all the bandwidth; popular files download faster as more peers join, and there is no single point of failure.
- Most traffic (file access, printing) stays inside each building's subnetwork, reducing congestion on the main network; a fault in one subnetwork does not affect the others; access can be managed per building.
- Thin clients are cheaper to buy and easier to manage, but the processing would all be on the server. Heavy 3D processing for many users at once needs an extremely powerful server, and transmitting large models and rendered views would need high network bandwidth. If the server or network failed, all designers would stop work. The savings on client hardware may be outweighed by server and network costs, so thick clients (powerful PCs) are probably more suitable.
- Each shop has a LAN connecting its tills (clients) and a router; the shops' LANs are connected to head office over a WAN using leased lines or a secure internet connection. The central database runs on a database server at head office. Client-server is justified because there must be one authoritative copy of the stock levels, updated by every till; the server controls access and security, handles concurrent updates and is backed up centrally.