If networks spread viruses, why does every school connect its PCs?

Why connect computers?

• A computer network is two or more computers (or other devices) connected so they can share data and resources.
• Joining them brings gains, and it brings costs and risks:

Backups, archives and disaster recovery

• Backups: every computer's data can be copied automatically, from one place, to a server or off-site store.
• Archiving: old files that are rarely used can be moved to cheaper storage on the network but still found later.
• Disaster recovery: if a building floods or a server fails, data and services can be restored from copies kept at another site.

Security, reach and running costs

• Security: staff can manage user accounts, passwords and access rights for everyone from one place.
• Globalisation: people in different countries can share data and work on the same files.
• Cost: switches, cables and servers must be bought, and they all use power.
• Maintenance: a network needs skilled technicians to keep it running, update it and fix faults.

One box fails and a whole office goes offline. How?

The star topology

• A topology is the layout of how the devices on a network are connected.
• In a star topology, every device has its own cable to one central device, usually a switch. All data passes through the centre, which sends it on to the right device.
• The central device can also be a hub or a computer.

The mesh topology

• In a mesh topology, devices connect directly to several others, so there are many paths between them.
• In a full mesh every device connects to every other one. With 4 devices that is 6 links.
• If a link fails, data is sent along another path: the network self-heals.

The bus topology

• In a bus topology, every device connects to one shared main cable, the backbone, with a terminator at each end.
• Data sent by one device travels along the whole backbone; only the device it is addressed to keeps it.
• It is cheap and needs little cable, but if the backbone breaks the whole network fails, and it slows down as devices are added because of data collisions.

The ring topology

• In a ring topology, each device connects to exactly two others, forming a closed loop.
• Data travels round the ring in one direction, passing through each device until it reaches its destination.
• There are no collisions, so it copes well with heavy traffic. But one broken link or device can stop the whole ring, and adding a device means breaking the loop.

How does your data find one computer among billions online?

IP addresses

• An IP address identifies a device on a network, so data can be sent to it. Every packet carries a destination IP address.

Routers

• A router connects networks together, for example a home or school LAN to the internet.
• The internet is a vast number of networks joined by routers.
• A router reads each packet's destination IP address and sends it to the next router on a route towards it. Hop by hop, the packet reaches the right network.
• A router stores the addresses of the computers on its network.

Devices that build a LAN

• A NIC lets a device connect to a network, wired or wireless.
• A switch connects the devices on a LAN. It reads where each packet is going and sends it only to that device.
• A wireless access point (WAP) lets wireless devices join a wired network.
• A hub copies every packet to all devices. A bridge joins two networks that use the same protocols, e.g. two LANs.

How do two computers that have never met agree how to talk?

Protocols

• A protocol is a set of rules for transferring data between devices.
• It sets things like how data is formatted, how devices are addressed and how errors are spotted.
• Both devices must follow the same protocol, whoever made them. Different jobs use different protocols.
• A standard is an agreed set of rules for an area of computing. Standards let hardware and software from different makers, and different countries, work together.

Protocols for the web and email

• HTTP: a browser requests web pages from a web server, which sends them.
• HTTPS: HTTP with encryption, so intercepted data can't be read.
• FTP: transfers files between computers.
• SMTP: sends email from an app to a mail server, and between mail servers.

Protocols that move the data

• TCP splits data into numbered packets, checks they all arrive, has missing ones resent and puts them back in order.
• IP gives each packet its source and destination IP addresses and routes it across networks.
• Ethernet is the set of rules for wired LANs; Wi-Fi (802.11) for wireless LANs.

Why use layers?

• Sending data is split into layers. Each layer does one part of the job and each protocol belongs to one layer.
• Sent data passes down the layers, each adding its own information. Received data passes back up.
• Benefits: each layer is a smaller problem; one layer can change without changing the others (swap a cable for Wi-Fi and the browser still works); makers can build for one layer.

The 7-layer OSI model

How does a video arrive in order if its pieces take different routes?

Splitting data into packets

• Before data is sent across a network, it is broken into small pieces called packets.
• Each packet is sent on its own and the receiving computer joins them back together.
• Small packets share the network's links fairly: no single huge file blocks a link. If one packet is lost or damaged, only that packet needs to be sent again.

Inside a packet

• The header holds the destination address, the source address (where it came from) and the packet number, used to put the data back in order.
• The payload is the actual data being carried.
• A checksum is included so the receiver can check the data has not been corrupted.

Packet switching

• In packet switching, each packet is sent across the network separately.
• Routers read each packet's destination address and send it on towards its destination, one hop at a time.
• Each packet may take a different route, depending on which links are free or working.
• So packets can arrive out of order. Once the last packet arrives, the receiver uses the packet numbers to reorder them.

Routing costs

• Routers look for the lowest-cost route to the destination.
• Each link between nodes is given a cost, based on things like its length and how fast the nodes on it are.
• The cost of a route is the total of the costs of the links it uses.
• The lowest-cost route is not always the one with the fewest hops.

Routers only use numbers. How does a typed web address find its site?

Reading a web address

• A URL (uniform resource locator) is the text address of a web page.
• Its parts: the scheme (https), subdomain (www), second level domain (example), top level domain (.org) and sub directory (/physics/). The page waves.html sits in that sub directory.

From name to number

• Every device on the internet is found by its IP address, e.g. 203.0.113.25. Names are easier to remember.
• The DNS (Domain Name System) matches domain names to IP addresses.
• The browser sends the domain name to a DNS server, which returns the IP address.
• If that server doesn't hold the name, it asks another DNS server. If none does, an error is returned.

Fetching the page

• A web server stores web pages and sends them to clients, such as your browser, that request them.
• The browser sends an HTTP request to the server's IP address. The server replies with the page's HTML, which the browser renders into the page you see.
• HTTPS is the secure version: data is encrypted, so intercepted data can't be read.

What a browser does

• A web browser's main job is to render HTML and display web pages.
• It also stores bookmarks, records history, allows multiple tabs, and provides navigation tools and an address bar.

Big ideas

• A network connects devices so they can share data and resources. It brings gains (sharing, central backup and management) and costs (hardware, maintenance, security risks).
• Networks allow central backups, archiving and disaster recovery, but cost money to build and maintain.
• A star links every device to a central switch: a cut cable affects one device, but if the switch fails the whole network fails.
• A mesh gives devices many paths, so it self-heals, but it needs a lot of cable and is complex to set up.
• A bus shares one backbone and a ring passes data round a loop: both are cheap, but one break can stop the whole network.
• An IP address identifies a device on a network. Routers read each packet's destination IP address and forward it hop by hop.
• Switches send data to one device and a WAP adds wireless devices; hubs copy data to every device and bridges join similar networks.
• A protocol is a set of rules for transferring data. Both devices must follow the same one, and different jobs use different protocols.
• The 7-layer OSI model runs from application (7) down to physical (1); each layer has its own purpose and protocols.
• Standards let hardware and software from different makers work together.
• Data is split into packets; each has a header (destination address, source address, packet number) and a payload.
• In packet switching, routers send each packet separately, so packets may take different routes and arrive out of order; the receiver reorders them using the packet numbers.
• Routers choose the lowest-cost route: add up the link costs along each route.
• A URL holds a domain name; a DNS server converts it into the web server's IP address.
• The browser requests the page with HTTP or HTTPS (encrypted) and renders the HTML that comes back.