Describe the four-layer TCP/IP model and the purpose of each layer
Explain why layered models are used and the benefits of protocols being standardised
Describe encapsulation and decapsulation
Compare the TCP/IP model with the OSI model
TCP/IP Layers
The Four-Layer TCP/IP Model
Application Layer
HTTP, HTTPS, FTP, SMTP, DNS, SSH
Transport Layer
TCP, UDP — ports, segmentation, reliability
Internet (Network) Layer
IP — addressing, routing, fragmentation
Link (Network Access) Layer
Ethernet, Wi-Fi — MAC addressing, physical
Why Layers?
Benefits of a Layered Model
Interoperability: any device can communicate with any other as long as both implement the same protocols at each layer. Manufacturers and developers can build independent implementations.
Abstraction: each layer only communicates with the layers directly above and below. A web developer doesn't need to understand Ethernet frames; a network engineer doesn't need to understand HTTP.
Modularity: a layer's implementation can be changed or improved without affecting other layers, as long as the interface between layers remains the same. E.g. switching from IPv4 to IPv6 doesn't break HTTP.
Troubleshooting: faults can be isolated to a specific layer, simplifying diagnosis. If packets are being routed correctly but the web page doesn't load, the issue is at the application layer, not the network layer.
Encapsulation
Encapsulation and Decapsulation
Encapsulation: as data travels down the stack on the sender's device, each layer adds its own header (and sometimes trailer) wrapping the data from the layer above. This creates the structure needed for that layer's protocol to function.
Sending (Encapsulation)
Application: data → HTTP message
Transport: HTTP + TCP header → segment
Internet: segment + IP header → packet
Link: packet + MAC header/trailer → frame
Receiving (Decapsulation)
Each layer on the receiver strips its own header and passes the remaining data up to the next layer, until the original application data is delivered.
Reliable — guarantees delivery, retransmits lost segments
Flow control and congestion control
Used for HTTP, HTTPS, email, file transfer
UDP — User Datagram Protocol
Connectionless — no handshake
Unreliable — no delivery guarantee, no retransmission
Faster and lower overhead
Used for DNS, video streaming, VoIP, online gaming
Exam Practice
OCR H446 Style · 4 marks
Explain why UDP is more suitable than TCP for a live video streaming application, despite UDP providing no guarantee of delivery.
[4 marks]
1
UDP has lower overhead than TCP — no three-way handshake is needed to establish a connection, reducing latency at the start of the stream.
1
UDP does not retransmit lost packets. In live video, a retransmitted packet would arrive too late to be useful — displaying a brief glitch is preferable to pausing playback to wait for retransmission.
1
TCP's flow and congestion control would slow transmission when the network is busy, causing buffering. UDP streams at the required rate regardless.
1
Live streaming requires real-time delivery — a dropped frame is acceptable (humans can tolerate brief quality drops), but latency caused by TCP's reliability mechanisms would break the live experience.
Common Mistakes
Don't Lose Marks
!
Saying TCP/IP has 7 layers — the OSI model has 7 layers; the TCP/IP model has 4. OCR H446 uses the TCP/IP model. Know which model is being asked about — the layers are named differently.
!
Saying UDP is faster because it skips encryption — UDP is faster because it has no connection setup, no retransmission and no flow control. Encryption is a separate concern and is independent of whether TCP or UDP is used.
!
Confusing encapsulation direction — headers are added as data moves DOWN the stack (sending); headers are stripped as data moves UP the stack (receiving). Mixing these up shows a fundamental misunderstanding of how layered models work.