COMP90007 Chap.1 Network Architecture and Layered Services
Network Architecture and Layered Services
A network answer is strongest when it identifies the kind of connection before naming a technology. A client-server exchange has an application-level request and reply, but those messages travel through several layers and links. Broadcast delivery exposes a packet to every station on one shared medium, point-to-point delivery joins individual pairs, and multicast selects a subset.
Topology describes how devices and links are arranged; it does not by itself state the service or access rule. The chapter therefore separates physical arrangement, transmission scope and protocol behaviour. It then uses service, interface and protocol as three different architectural relations.
The OSI model provides a principled seven-layer reference vocabulary, TCP/IP is tied to a successful protocol suite, and the subject's hybrid view keeps physical and data-link functions distinct while following the deployed Internet stack. The goal is not to recite layer names. It is to trace what changes as data is encapsulated, passed across an interface, communicated between peers and delivered upward at the destination.
What this chapter covers
- 01
Client
- 02
Packet
- 03
Encapsulation
- 04
Requests, replies, clients, servers and packets
- 05
Broadcast, point-to-point and multicast transmission
- 06
Mesh, bus, star and ring topology
- 07
Services, interfaces and peer protocols
- 08
OSI, TCP/IP and hybrid reference models
- 09
Encapsulation and end-to-end tracing
Locate three relations in a layered exchange
- 1The application service is the capability requested by the browser, such as retrieving the named resource.
- 1An interface is used locally when the application hands data to transport or transport hands a packet to the network layer.
- 1A peer protocol coordinates corresponding entities, such as transport at the two endpoints or the wireless link entities on one hop.
- 1The wireless technology changes the first hop but does not turn the Web application into a link-layer protocol.
Key terms
- Client
- A network device or program that requests a service from a server.
- Packet
- A formatted message transferred between network devices, carrying control information and payload appropriate to its layer.
- Encapsulation
- Adding layer-specific control information around data as it moves down a protocol stack.
Network Architecture and Layered Services FAQ
Does a topology determine the protocol?
No. A bus creates a shared-medium coordination problem and a star identifies a central attachment pattern, but several access or forwarding protocols could operate over either arrangement. State the physical relationship and the behavioural rule separately.
Why is the Web not the Internet?
The Internet is an interconnected network infrastructure and protocol environment. The Web is a distributed application that runs over that infrastructure. Separating them prevents application behaviour from being mistaken for a property of every Internet packet.
Exam move
Redraw an exchange twice: first as physical devices and links, then as a layered stack at each endpoint. Label every arrow as service access, interface use or peer communication. The mismatch between the two drawings is where most vocabulary errors become visible. Build a packet-trace notebook with one row for each boundary crossed by a request.
Record the application message, the service requested from the next layer, the local interface operation, the peer protocol and the data object carried on the current link. At every hop, mark what stays end-to-end and what is replaced locally. Repeat the trace for a reply so client and server roles are tied to the exchange rather than to a permanent machine type.
Draw broadcast, point-to-point and multicast delivery separately, then add bus, star, ring and mesh arrangements without assuming that topology determines access behaviour. For each drawing, state who can receive a transmission and which coordination problem remains. Compare OSI, TCP/IP and the subject's hybrid model by mapping functions, not by forcing identical layer counts.
When two models group functions differently, explain which service boundary becomes less visible and why the deployed protocol suite may still work. Use a counterfactual check: remove one layer, merge two layers or change one link technology, and predict which interfaces and peer relations must change.
End each weekly trace with a fault placed at one layer—wrong name resolution, missing route, damaged frame or noisy signal—and reject explanations that belong to another layer. This produces a reusable diagnosis method while keeping service, interface and protocol distinct.
Before revising, ask whether every arrow represents communication between peers or access to a local service; relabel any arrow that silently mixes those relations. Validate the trace with a second application that uses a different transport pattern. Preserve the same layer vocabulary while noting exactly which endpoint behaviour changes and which lower-layer services remain reusable.
Keep the final diagram beside its assumptions so another reader can reproduce the route.
Working through Network Architecture and Layered Services in COMP90007? Sia is AskSia’s AI Computer Science tutor — ask any COMP90007 Network Architecture and Layered Services question and get a clear, step-by-step explanation grounded in how COMP90007 is taught and assessed. Read this chapter free, then take your hardest questions to Sia.
COMP30023 Computer Systems · COMP90038 Algorithms and Complexity · COMP20008 Elements of Data Processing
