2.1.3 Network Topologies and Packet Paths
A network topology describes the pattern of links and connecting devices within a network. This pattern affects the route taken by a packet, the amount of infrastructure required, and the effect of a device or link failure.
This section focuses on four required topologies: bus, star, mesh and hybrid. You will trace a packet between two hosts and use evidence about cost, resilience, scalability and management to recommend a suitable design.
By the end of this section, you should be able to:
- describe the connection pattern used by bus, star, mesh and hybrid topologies;
- trace the route taken by a packet between two hosts in each topology;
- explain how selected device and link failures affect communication;
- compare topologies using cost, resilience, scalability and ease of management;
- select and justify a suitable topology for a given situation.
Topology and packet paths
Host: an end device that sends or receives data.
Packet: a unit of data containing control information, including information used to identify its source and destination.
When Host A sends data to Host D, the packet does not simply “move through the network”. Its route depends on how the network is arranged:
- Host A creates or forwards a packet containing destination information.
- The packet travels over one or more links according to the topology.
- Any intermediate connecting device forwards the packet towards the destination.
- Host D accepts the packet intended for it.
Exam tip
When asked how a packet travels, name the important parts of its route. For example: source host → central switch → destination host. Avoid vague statements such as “the packet goes through the network”.
Bus topology
How the network is arranged
Every host connects to the same backbone cable. There is no separate direct link between each pair of hosts and no central switch controlling the route.
Packet path from Host A to Host D
- Host A places the transmission onto the shared bus.
- The signal travels along the common cable and reaches the attached hosts.
- Hosts inspect the destination information.
- Host D accepts the packet; hosts for which it is not intended ignore it.
Benefits, limitations and failures
| Point | Effect |
|---|---|
| Single shared backbone | Uses relatively little cable, but every transmission competes for the same medium. |
| No central switch | There is no central switching device to purchase, but traffic cannot be isolated as easily. |
| One host fails | The other hosts may continue communicating if the shared cable remains intact. |
| Main cable breaks | Communication across the break is disrupted and a large part of the network may become unusable. |
| Network grows | More hosts share the same link, increasing contention and making faults harder to locate. |
Common mistake
Do not describe a bus as having a separate direct cable between each pair of hosts. All hosts share the same main link.
Star topology
Packet path from Host A to Host D
- Host A sends the packet along its dedicated link to the central switch.
- The switch examines the destination information.
- The switch forwards the packet through the link leading to Host D.
- Host D receives the packet.
If the central device were an older hub rather than a switch, it would repeat the transmission to every connected link. In a modern switched star, traffic can be directed towards the intended host.
Benefits, limitations and failures
| Point | Effect |
|---|---|
| Dedicated host links | Traffic on one host's cable does not need to use another host's cable. |
| Central switch | Packets can be directed and the network is usually easier to monitor and expand. |
| One host or outer link fails | Other hosts can normally continue communicating. |
| Central device fails | Communication across the star is interrupted because every route depends on the centre. |
| Cabling and hardware | Requires a separate link for each host and a central device, increasing setup cost. |
Common mistake
A star is resilient to the failure of one outer link, but it still has a critical central device. Do not claim that every failure leaves the network unaffected.
Mesh topology
Full mesh: every device has a direct link to every other device.
Partial mesh: only selected pairs have direct links, but alternative routes still exist.
Packet path from Host A to Host D
In a full mesh, Host A may have a direct link to Host D. In a partial mesh, the packet can pass through intermediate network devices. A route is selected, and another route may be used if a link or device becomes unavailable.
Benefits, limitations and failures
| Point | Effect |
|---|---|
| Multiple paths | Traffic may be redirected when one route fails, providing strong resilience. |
| Direct connections | Some destinations can be reached with fewer intermediate steps. |
| Large number of links | Installation, ports and maintenance can be expensive and complex. |
| Expansion | Adding a device to a full mesh requires many new links, so full meshes scale poorly. |
| Typical use | Often more appropriate for important switches or routers than for every user workstation. |
Exam tip
“Mesh is reliable” is incomplete. Develop the explanation: multiple routes exist, so a packet can be redirected when one link fails.
Hybrid topology
Packet path from Host A to Host D
Consider two star LANs joined through a resilient backbone:
- Host A sends the packet to the switch in its local star.
- The local switch forwards it to a connecting router or backbone device.
- The packet crosses the backbone using the route available there.
- The destination LAN's switch forwards the packet to Host D.
The route therefore follows different connection patterns during one journey. The packet may begin in a star, cross a mesh-like backbone, and finish in another star.
Benefits, limitations and failures
| Point | Effect |
|---|---|
| Flexible design | Each section can use a topology suited to its local requirements. |
| Incremental expansion | A new wired, wireless or specialist LAN can be added without rebuilding every existing section. |
| Failure impact | The effect depends on the failed component and the topology used in that section. |
| Management | Different technologies and connecting devices make diagnosis and administration more complex. |
| Cost | Flexibility and resilience may require several types of equipment and skilled support. |
Common mistake
A hybrid is not one fixed geometric shape. It is a combination of network sections that use different arrangements or technologies.
Comparing the topologies
| Decision factor | Bus | Star | Mesh | Hybrid |
|---|---|---|---|---|
| Connection pattern | One shared backbone | Separate links to a centre | Multiple direct or indirect paths | Combination of structures |
| Typical packet path | Along the shared cable | Through the central device | Directly or through a selected route | Changes between network sections |
| Key vulnerability | Main cable | Central device | Cost and complexity rather than one unavoidable route | Connecting devices and complexity between sections |
| Expansion | Becomes less efficient as more hosts share the link | Usually straightforward if ports and capacity are available | Full mesh requires many additional links | Can add sections using a suitable local design |
| Typical reason to select | Very low infrastructure requirement in a simple or legacy setting | Manageable modern LAN with isolated outer-link failures | High resilience and alternative routes | Different parts of a large organisation have different needs |
Choosing and justifying a topology
A suitable recommendation must respond to the requirements in the scenario. Do not select a topology from one memorised advantage alone.
Reasoning structure
Choose the topology → identify a relevant feature → explain the consequence for the stated situation → acknowledge an important trade-off.
Examination computer room: thirty desktops, easy fault isolation and future expansion are important.
Suitable choice: star. Each computer has an individual link to a switch, so one damaged outer cable normally affects only one workstation. Extra computers can be added when switch capacity is available. The central switch is a critical component and should be reliable.
Emergency communications backbone: service should continue when one route is unavailable.
Suitable choice: mesh or partial mesh. Multiple paths allow packets to be redirected around a failed link or intermediate device. This resilience must be balanced against the higher equipment and management cost.
Multi-building college: each building has a star LAN, but the buildings need resilient interconnection.
Suitable choice: hybrid. The existing star LANs can be retained and joined through a backbone with alternative routes. Each section uses an arrangement suited to its purpose, although the combined system is more complex to manage.
Very small legacy monitoring installation: cost matters more than expansion or high availability.
A bus may be acceptable in a constrained legacy setting. One shared cable limits the amount of infrastructure. However, a backbone fault can interrupt much of the network and a star would usually be preferred for a new general-purpose LAN.
Common mistake
Avoid absolute statements such as “mesh never fails” or “star is always best”. The correct choice depends on the required resilience, cost, scale and management.
Interactive: Topology Explorer
Choose a topology and send a message. For this syllabus section, concentrate first on unicast: a packet travelling between one source host and one destination host. Multicast and broadcast remain available as extension comparisons.
Use the fault demonstration to compare what happens when a cable, host or connecting device becomes unavailable.
Practice
Packet-path questions
- Describe how a packet travels from Host A to Host D in a bus topology.
- Describe how a switch is involved when a packet travels between two hosts in a star topology.
- Explain how a mesh can continue forwarding packets after one route fails.
- Trace a possible packet journey between hosts on two different LANs in a hybrid network.
Comparison and selection
- Explain one benefit and one drawback of a bus topology.
- Explain the different effects of an outer-link failure and a central-switch failure in a star.
- Explain why a full mesh becomes difficult to scale.
- A research facility requires alternative routes between important network devices. Recommend a topology and justify it.
- A school is adding a wireless building network to several existing wired star LANs. Recommend a topology category and justify it.
Review
| Topology | Core connection idea | Packet route | Important trade-off |
|---|---|---|---|
| Bus | One shared backbone | Along the common cable | Low infrastructure, but shared capacity and backbone vulnerability |
| Star | Individual links to a centre | Source → central device → destination | Easy isolation and expansion, but central-device dependence |
| Mesh | Multiple paths | Directly or through a selected route | Strong resilience, but high link and management cost |
| Hybrid | Different structures combined | Changes as the packet crosses sections | Flexible, but more complex to design and support |
Quick self-check
- Can I identify all four required topologies from their connection patterns?
- Can I trace a packet between two hosts without using vague language?
- Can I distinguish the effects of host, link and central-device failures?
- Can I justify a topology using requirements from a scenario?