Protocols

A protocol is an agreed set of rules that controls how data is transmitted between devices on a network. The rules typically cover:

• the format of the data (how it is structured),
• the order in which messages are sent (e.g. who 'speaks' first),
• how errors are detected and dealt with.

Why is a protocol essential? Two communicating devices may be made by different manufacturers and run different software. If they did not both follow the same rules, the receiver could not make sense of the bits the sender transmits. A shared protocol is the 'common language' that makes communication possible.

Think of two people on a phone call who must agree to speak the same language, take turns, and confirm they have understood — that agreement is the protocol.

Network communication is complicated, so it is broken up into a stack of layers. Each layer has its own functionality (its own specific job).

A layer does not need to know how the other layers work internally — it only needs the interface to its neighbours:

• it provides a service to the layer above it, and
• it uses the service of the layer below it.

When you send data it travels DOWN the stack (each layer adds its own information); at the receiver the data travels UP the stack (each layer removes and uses that information).

Benefits of organising protocols into layers:

• Modular / self-contained: a large problem is split into smaller, manageable parts, so the system is easier to design, build and understand, and faults can be isolated to one layer.
• Independent layers: one layer can be changed or upgraded (for example, swapping the protocol it uses) without affecting the other layers, as long as the interface between layers stays the same.

The TCP/IP protocol suite is the set of protocols used to communicate over the internet. It is organised into four layers, named after its two core protocols (TCP and IP).

Application layer — provides network services directly to the user's application software (browser, email client, etc.). This is where the application protocols HTTP, FTP and SMTP run and where the data/request is created.

Transport layer — splits the data into segments and provides end-to-end delivery. Using TCP it adds reliability (acknowledgements, retransmission of lost data, putting segments back in order) and uses port numbers so the data reaches the correct application. (UDP is the faster, connectionless alternative.)

Internet layer — adds source and destination IP addresses to form packets and is responsible for the addressing and routing of those packets between networks across the internet, using the IP protocol.

Link layer (network access) — handles the physical transmission of the data onto the local link/medium (cable or wireless) and uses the hardware (MAC) address for delivery on the local network.

Suppose a browser requests a web page. The request passes down the stack on the sending host:

1. Application — the browser uses HTTP to create the request.
2. Transport — TCP splits the request into segments, adds port numbers, and prepares for reliable delivery.
3. Internet — IP adds the source and destination IP addresses and the packets are routed across the networks of the internet.
4. Link — the bits are physically transmitted onto the local link, using the MAC address for the next hop.

At the receiving host (the web server) the data passes back UP the same four layers in reverse:

• Link receives the bits from the medium,
• Internet checks the destination IP address,
• Transport reassembles the segments in the correct order,
• Application hands the complete request to the web server software.

The reply (the web page) then makes the same journey in the opposite direction. This is the key idea: each layer at the sender is matched by the same layer at the receiver, with the lower layers carrying the data and the upper layers giving it meaning.

All of these are application-layer protocols (BitTorrent operates at the application level too). You must know the purpose of each:

Protocol	Purpose
HTTP	Transfer web pages between a web server and a browser/client.
FTP	Transfer files to and from a remote server (upload & download).
SMTP	Send email from a client to a mail server, and relay it between mail servers.
POP3	Retrieve email — usually downloads it to one device and deletes it from the server.
IMAP	Retrieve email — keeps it on the server and synchronises across devices.
BitTorrent	Peer-to-peer (P2P) file sharing — peers swap pieces of a file with each other.

Email — get the direction right: SMTP Sends; POP3 and IMAP retrieve. The difference between the two retrieval protocols is where the mail lives:

• POP3 → downloads the mail to a single device, then usually removes it from the server (good for one device / saving server space).
• IMAP → leaves the mail on the server and synchronises it, so every device shows the same up-to-date mailbox (good for using multiple devices).

BitTorrent — not client–server: instead of everyone downloading from one central server, each peer downloads pieces of the file from other peers and uploads pieces it already has. This spreads the load and is efficient for sharing large files among many users.