A YouTube video streams over a cellular network. Which layer is responsible for ensuring that video data packets are delivered reliably and in order to the viewer's device?
APhysical layer — it controls the radio signals that carry the data
BNetwork layer — it routes packets across the internet from YouTube's servers to the viewer
CTransport layer — it provides end-to-end reliability between the communicating applications
DApplication layer — HTTP handles reliable delivery of video data directly
End-to-end reliability is the job of the transport layer (TCP in this case). The physical layer handles raw bit transmission over the radio link between adjacent nodes; the network layer routes packets hop by hop across the internet; the application layer (HTTP) operates in terms of requests and responses, relying on lower layers for delivery guarantees. The critical point is that 'end-to-end' means across the entire path from source to destination — spanning multiple physical links and routers — which is precisely the transport layer's scope. This is one of the most useful mental models in networking: different guarantees live at different layers.
Question 2 Multiple Choice
An engineer wants to replace the physical transmission medium from fiber optic cable to a new quantum communication link, without changing how IP routing, TCP connections, or web applications work. Is this possible under the layered model?
ANo — changing the physical medium requires updating all higher layers since they depend on implementation details of the physical transmission
BYes — each layer only interacts with its immediate neighbor through a defined service interface; higher layers don't know or care about the physical medium
COnly if the quantum link supports the same bandwidth as fiber, since higher layers assume specific throughput
DOnly if IP and TCP are redesigned to take advantage of quantum properties
This is the central benefit of protocol layering: as long as a new physical medium implements the same service interface to the data-link layer above it (the ability to transfer frames between directly connected nodes), everything higher in the stack works unchanged. This is how Ethernet, Wi-Fi, fiber, 5G, and satellite links all run under the same IP stack — they all implement the same interface. The layers above don't know what medium is below them, and by design, they don't need to.
Question 3 True / False
For the transport layer to send data correctly, it is expected to know whether the physical layer is using fiber optic cable, Wi-Fi, or a cellular radio link.
TTrue
FFalse
Answer: False
This is exactly what protocol layering prevents. Each layer interacts only with its immediate neighbors through defined interfaces: the transport layer calls on the network layer, which calls on the data-link layer, which calls on the physical layer. The physical medium is completely invisible to the transport layer — this independence is the key architectural property. It is why the internet can work across such diverse physical infrastructure without every application needing to be redesigned for each new medium.
Question 4 True / False
Encapsulation in a protocol stack means that as data passes down the sending stack, each layer adds its own header, creating a nested structure that each corresponding layer at the receiving end strips off.
TTrue
FFalse
Answer: True
As data passes from application → transport → network → data-link → physical, each layer prepends its own header (and sometimes a trailer) to the payload, wrapping it like nested envelopes. The receiving end unwraps in reverse: the physical layer passes bits up; the data-link layer strips its frame header; the network layer strips the IP header; the transport layer strips the TCP/UDP header; the application receives the original data. Each header contains information only that layer needs to do its job — port numbers for transport, IP addresses for network, MAC addresses for data-link. This structure is what lets each layer operate independently.
Question 5 Short Answer
Why does breaking a message into small packets (rather than transmitting it as one continuous stream) improve network performance and reliability?
Think about your answer, then reveal below.
Model answer: Packets enable multiple sources to share the same links simultaneously through statistical multiplexing: different packets from different users interleave on the same wire, so no user monopolizes a dedicated connection. If transmission fails, only the lost packet needs retransmission, not the entire message. Packets can take different routes through the network if a path becomes congested or fails, increasing resilience. Smaller packets also reduce head-of-line blocking: a large continuous stream would delay all other traffic until it finishes, while short packets allow other users' packets to interleave. These properties together make packet-switched networks far more efficient and robust than circuit-switched networks that dedicate a fixed path for each communication.
The contrast with circuit switching clarifies the advantage: in a circuit-switched network (like traditional telephone), a dedicated path is reserved for the duration of the call, wasting capacity whenever there's silence. Packet switching shares capacity dynamically — links carry packets from whoever has something to send. The internet's scalability to billions of simultaneous users depends fundamentally on this sharing.