ScholaFly

CS08-02 Computer Science Watch

The four-layer TCP/IP model

Watch on YouTube

In this lesson

TCP/IP model: how to name the four layers of the TCP/IP model in order, with worked examples and the mistakes examiners report. AQA 8525, Edexcel 1CP2 and OCR J277 GCSE. Protocols and layers 2 of 6. Full Protocols and layers playlist: https://www.youtube.com/playlist?list=PLD_gG3hKyrIE More at https://scholafly.com

By the end: Name the four layers of the TCP/IP model in order - application, transport, internet, link - say what each layer does and which protocols work at it, follow a message down the layers on one computer and up them on the other, describe what a packet's header, data and footer hold, and explain why networking is built in layers at all.

What it covers

  1. 0:00 Intro: TCP/IP model: four layers
  2. 0:35 TCP/IP model: four layers: what a layer
  3. 1:37 Four layers, four jobs
  4. 3:46 Layer names
  5. 6:05 The bottom two layers
  6. 7:44 Down one stack, up the other
  7. 9:14 Why layers
  8. 10:57 TCP/IP model: four layers: the packet

Key words

About this video

TCP/IP Model: the four-layer TCP/IP model | GCSE Computer Science | CS08-02

TCP/IP model: how to name the four layers of the TCP/IP model in order, with worked examples and the mistakes examiners report.
AQA 8525, Edexcel 1CP2 and OCR J277 GCSE. Protocols and layers 2 of 6.
Full Protocols and layers playlist: https://www.youtube.com/playlist?list=PLD_gG3hKyrIE
More at https://scholafly.com

Watch first: CS08-01 - https://youtu.be/PwJYgaGSx_o
Video code: CS08-02 (search "ScholaFly CS08-02")

#GCSEComputerScience #ProtocolsAndLayers #TCPIPModelFourLayers

For teachers
This GCSE Computer Science lesson teaches the four-layer TCP/IP model. By the end, students should be able to name the four layers of the TCP/IP model in order - application, transport, internet, link - say what each layer does and which protocols work at it, follow a message down the layers on one computer and up them on the other, describe what a packet's header, data and footer hold, and explain why networking is built in layers at all. It works through six worked examples and the mistakes examiners report.

Exam board specification references:
AQA 8525
- 3.5
Edexcel 1CP2
- 4.1.7
OCR J277
- 1.3.2

Read the transcript

How can an email program written twenty years ago, and never rewritten, still work on networks that did not exist back then? You might guess someone updates it for each new network. Nobody needs to: it never knows whether mail leaves by Wi-Fi, by cable or by a phone mast. That is because the job of sending is split into four layers, and each layer does one thing.

A layer is one self-contained part of the job of communicating. Each layer does one job, uses the layer below it, and serves the layer above it. Posting a parcel works the same way. You write the address, the sorting office sorts it, and a van carries it. You never need to know which van. If your course asks for the layer names and their jobs, every one is here. If it asks only for the idea of layers and why they are used, the names are the clearest way to see it. The four layers of the TCP/IP model, from the top, nearest the user, to the bottom, nearest the wire, are application, transport, internet and link.

Start at the top. The application layer is where network applications, such as web browsers and email programs, do their work. HTTP, HTTPS, SMTP and IMAP work at the application layer. So do FTP and POP. Below it is the transport layer. It sets up the communication between the two computers, and they agree settings such as the size of the packets. The transport layer splits the data into packets and makes sure they all arrive. TCP works here. UDP sits here too, on one older course list only. Next is the internet layer. It addresses and packages the data, and routes each packet across the networks in between. IP works here, using IP addresses. At the bottom is the link layer. This is the network hardware, such as the network interface card, and the drivers the operating system uses to run it. Down here, Ethernet runs the cables and Wi-Fi runs the wireless connections. Each network card also carries its own fixed address, a MAC address. Some older books call this bottom layer the network access layer. Now three protocols, each from a different height in the stack. Which of HTTP, TCP and Wi-Fi puts the data onto the wire or into the air? Wi-Fi. It works at the link layer, beside the hardware that sends the signal. HTTP is an application protocol, and TCP is at transport.

Now the names themselves, because a name that is nearly right still scores nothing. A student, asked for two layers of the TCP/IP model, writes: Data Link layer, and Network layer. What has gone wrong with Data Link and Network in a TCP/IP answer? They come from a different set of seven, which has names of its own. Here the right names are Link and Internet. One examiner's report records it: "A very common mistake in responding to question 13.5 was to name the Data Link and Network layers in the OSI model rather than the Link and Internet layers in the TCP/IP model." The fix is four names and only these four: application, transport, internet and link. A layer name outside those four belongs to a different model. Now, from the top, can you name all four layers and a protocol for each of the top three? Application, transport, internet, link. One protocol for each of the top three: HTTP at application, TCP at transport, IP at internet. A report on a different paper notes: "Students could either complete the first two layers correctly or not identify either of the layers. There were a wide range of unrelated responses that showed a lack of understanding of the layers." The habit that fixes it: say all four, top to bottom, every time, and give the bottom two as much attention as the top two.

The bottom two layers sit side by side, so draw a firm line between them. The internet layer is IP addresses and routing between networks. Now a student downloads a file at home over Wi-Fi. Describe what happens at the link layer during that Wi-Fi download. The laptop's network card and its driver send and receive the radio signal using the Wi-Fi protocol. MAC addresses get the data to the right device on the local network. Look at what that answer leaves out: no IP address, and no routing. Those belong one layer up. A third report says: "Candidates found this to be a challenging question, and the link layer appeared to be poorly understood. Some candidates appreciated that WiFi and Ethernet are protocols that are used at the link layer and some could go on to expand as to the types of signal that would be sent. Few candidates discussed MAC addresses and how they are used at the link layer. Many candidates confused the link layer with the network layer and IP addressing." The fix: keep IP addresses at the internet layer, and name MAC addresses whenever the link layer comes up.

So far the stack has stood still. Data moves down it on one computer and up it on the other. Follow one email. On the sender's computer, the email program hands the message to the application layer, where SMTP takes charge of it. The transport layer splits it into numbered packets. The internet layer adds the source and destination IP addresses to each one. The link layer adds what the local hardware needs, and the network card sends the signal out. On the receiving computer, four things happen. A, transport rebuilds the message from its packets. B, link takes in the signal. C, application hands over the email. D, internet checks the address. Put A, B, C and D in the right order for the receiving computer. B, D, A, C. The receiver works up from the bottom, the reverse of the way down, and each layer removes what its partner on the sender added.

Layers are there for reasons, and one course asks for those reasons by name. Go back to the email program from the start. Swap the link layer under it from Ethernet to Wi-Fi. The three layers above carry on with no change at all. Why can the three layers above ignore that swap? Each layer is independent. It only meets its neighbours through agreed rules, so one layer can be altered without touching any other. That is the first benefit. The second is that one big problem becomes several smaller ones, each easier to design, build and fix when something goes wrong. The third is that different companies can make different layers, and their products still fit together, because the rules between the layers are agreed. One examiner's report notes: "Candidates were often able to show an understanding of layers but could not say why layers are used." Then it names the most common reason given: "The most common responses referred to the need for independence between layers." The answer to give instead of a description: lead with independence, then say what it allows, one layer changed and the others left alone.

One more piece: what the transport and internet layers build, the packet itself. A packet has three parts. The header comes first. It holds control information: the source and destination IP addresses, the packet number, and similar details. The data comes next. This is the piece of the message being carried, sometimes called the payload. The footer comes last. It marks the end of the packet and carries an error check. An envelope is a fair picture: addresses and numbering on the outside, the letter inside, and a seal at the end. Now packet seven reaches the receiver before packet six does. Which part of a packet lets the receiver put those two back in order? The header. It carries the packet number, so the transport layer can rebuild the message in sequence.

Four layers, four jobs, and a few questions to see whether they have stuck. Name the four layers, starting with the one nearest the user. Application at the top, then transport, then internet, and link at the bottom. Now a harder one: the internet layer moves to a new IP version. Which layers must change? Only the internet layer. Application, transport and link carry on as they were, because each layer is independent. Try this: SMTP, TCP, IP and Ethernet. Which layer does each one work at? SMTP at application, TCP at transport, IP at internet, and Ethernet at link. Here's a different one: which part of a packet holds the error check? The footer, at the end, after the data. Now that twenty-year-old email program: why does it still work on new networks? It works at the application layer. The layers below it can change without it noticing, because each layer is independent of the others.

These are the chapter's videos so far, ending on this one. If you can rebuild the stack without looking, a thumbs up marks it done and saves you a rewatch. If it is still wobbly, save it for later. The next video, on addresses, leans on this stack, so it comes round again soon.

Next in the chapter: IP addresses and MAC addresses. Hopefully that's next in your feed, but if not, search for ScholaFly C S oh eight, oh three.

Every other video is on scholafly.com too, easy to search by video code or by keyword. Whatever you're revising, go there and have a look, or, watch the next video.

Related terms

For: AQA GCSE 8525, Edexcel GCSE 1CP2, OCR GCSE J277

On the specification

BoardSpecStatement
AQA GCSE 85253.5Computer networks
Edexcel GCSE 1CP24.1.7Explain how the 4-layer TCP/IP model (application, transport, internet, link) handles data transmission over a network.
OCR GCSE J2771.3.2Wired and wireless networks, protocols and layers
For teachers

This GCSE Computer Science lesson teaches the four-layer TCP/IP model. By the end, students should be able to name the four layers of the TCP/IP model in order - application, transport, internet, link - say what each layer does and which protocols work at it, follow a message down the layers on one computer and up them on the other, describe what a packet's header, data and footer hold, and explain why networking is built in layers at all. It works through six worked examples and the mistakes examiners report.