Courses / Computer Science XI (HSSC-I) – FBISE Complete SLO-Based Course
1.4 Networks: Topologies, Scalability and Reliability
Unit 1 – Computer Systems · 75 min read · 13 MCQs to check yourself
Network topologies explained in 5 minutes – bus, star, ring, mesh, tree
1. Basic terms
| Term | Meaning |
|---|---|
| Node | Any device on the network – computer, printer, switch, phone |
| Server | A computer that provides a service (files, web pages, email) to others |
| Client | A computer that requests services from a server |
| Root | The top node of a tree (hierarchical) topology |
| Topology | The arrangement (layout) of nodes and links. Physical = how cables/links are laid out; logical = how data actually flows |
| Ethernet | The most common wired LAN standard (IEEE 802.3); today usually a star using switches |
| Scalability | How easily the network can grow (more users/devices/traffic) without losing performance |
| Reliability | How well the network keeps working when a part fails (fault tolerance) |
| Availability | The percentage of time the system is up and usable |
2. The topologies
| Topology | Layout | Advantages | Disadvantages |
|---|---|---|---|
| Bus | All nodes on one backbone cable with terminators at both ends | Cheap, little cable, easy for small networks | Backbone failure stops everything; collisions grow with traffic; hard to find faults; poor scalability |
| Ring | Each node connected to two neighbours in a closed loop; data travels one way | No collisions (token passing); equal access for all nodes | One broken node/link can stop the ring (unless dual ring); adding a node disturbs the network |
| Star | Every node has its own cable to a central switch/hub | Easy to add/remove nodes; one cable fault affects only one node; easy troubleshooting; high performance | Central switch is a single point of failure; more cable than bus |
| Tree | Stars connected in a hierarchy below a root | Very scalable; suits departments/buildings | Failure of the root or a branch cable cuts off everything below it |
| Mesh | Nodes connected to many (full mesh: all) other nodes | Most reliable – many alternative paths; no traffic bottleneck | Expensive, lots of cable/links; complex. Full mesh links = (n(n-1))/2 |
| Hybrid | Mix of two or more topologies | Flexible, combines strengths | Complex design, costly |
Example: a full mesh of 6 computers needs (6 × 5)/2 = 15 links. The Internet core is a (partial) mesh, which is why it keeps working when one route fails.
Star vs ring (model paper):
| Basis | Star | Ring |
|---|---|---|
| Architecture | Central switch, point-to-point links | Closed loop, each node linked to two others |
| Scalability | High – just plug a new cable into the switch | Low – the ring must be broken to add a node |
| Reliability | One cable fault affects one node; switch is the weak point | One node or link failure can stop the whole ring |
| Data flow | Through the switch | Around the ring, node to node |
3. Access methods: CSMA and token passing
When many nodes share one medium they need rules to avoid two transmitting at once.
- CSMA/CD (Carrier Sense Multiple Access with Collision Detection) – used in classic Ethernet/bus: listen before sending (carrier sense); if two nodes send together a collision is detected; both wait a random time and resend. Works well at low traffic, slows down at high traffic.
- CSMA/CA (Collision Avoidance) – used in Wi-Fi: devices try to avoid collisions (wait, send a short request, then data) because wireless devices cannot detect collisions while sending.
- Token passing – used in ring networks: a special frame called the token circulates; only the node holding the token may send. No collisions and predictable delay, but the token can be lost and must be regenerated.
4. Scalability and reliability
Designing for reliability: redundant links (mesh, dual ring), backup switches and power supplies (UPS), RAID disks, regular backups, monitoring. Designing for scale: star/tree with switches that can be stacked, structured cabling, load balancers that spread requests across several servers, cloud resources that grow automatically.
Load testing and system testing:
- Load testing – check performance under the expected number of users (e.g. 5 000 students opening the results website).
- Stress testing – push beyond the limit to find the breaking point and see how the system fails and recovers.
- System testing – test the complete system end-to-end (network, servers, software) against requirements.
- Tools: JMeter, Locust,
ping,tracert/traceroute.
Scalability and reliability in cloud computing: cloud providers offer auto-scaling (add servers when traffic rises, remove them later – elasticity), load balancing, replication of data in several data centres/regions and failover to a backup server. Vertical scaling = a bigger server (scale up); horizontal scaling = more servers (scale out).
5. Measuring availability
Availability = uptime/(total time) × 100% = MTBF/(MTBF + MTTR) × 100% MTBF = mean time between failures; MTTR = mean time to repair.
Example 1: a server was down for 10 hours in a year (8 760 hours). Availability = 8750/8760 × 100 = 99.89%. Example 2: MTBF = 990 h, MTTR = 10 h → availability = 990/1000 × 100 = 99%.
| Availability ("nines") | Downtime per year |
|---|---|
| 99% (two nines) | about 3.65 days |
| 99.9% (three nines) | about 8.76 hours |
| 99.99% (four nines) | about 52.6 minutes |
| 99.999% (five nines) | about 5.26 minutes |
mtbf, mttr = 990, 10
availability = mtbf / (mtbf + mttr) * 100
print(f"Availability = {availability:.2f}%")
print(f"Downtime per year = {(100 - availability) / 100 * 8760:.1f} hours")
Output:
Availability = 99.00%
Downtime per year = 87.6 hours
Types of network topologies with diagrams
Short and long questions
1. Compare star and ring topologies with respect to scalability, reliability and architecture.
Model answer
| Basis | Star | Ring |
|---|---|---|
| Architecture | All nodes connected to a central switch/hub | Each node connected to two neighbours in a closed loop |
| Scalability | High – plug a new cable into the switch | Low – the ring must be broken to add a node |
| Reliability | One cable fault affects only one node, but switch failure stops all | One node or link failure can stop the whole ring |
2. Differentiate between CSMA/CD and token passing.
Model answer
| CSMA/CD | Token passing |
|---|---|
| Nodes listen and send when the medium is free | Only the node holding the token can send |
| Collisions can happen and are detected; senders wait a random time | No collisions |
| Used in Ethernet / bus networks | Used in ring networks |
| Performance falls at heavy traffic | Predictable delay even at heavy traffic |
3. Define scalability, reliability and availability of a network.
Model answer
Scalability: the ability of a network to grow (more users, devices, traffic) without loss of performance. Reliability: the ability to keep working correctly when some component fails (fault tolerance). Availability: the percentage of time a system is operational: availability = uptime / total time × 100.
4. A web server ran for 8 750 hours and was down for 10 hours in a year. Calculate its availability. How can availability be improved?
Model answer
Availability = 8750 / (8750 + 10) × 100 = 8750 / 8760 × 100 ≈ 99.89%. Improve by: redundant (backup) servers with automatic failover, UPS/backup power, load balancing, RAID disks, regular maintenance and monitoring.
5. Give two advantages and two disadvantages of mesh topology.
Model answer
Advantages: (1) most reliable – many alternative paths, so one link failure does not stop communication; (2) no traffic bottleneck and good privacy, as data can go directly between nodes. Disadvantages: (1) expensive – a full mesh of n nodes needs n(n − 1)/2 links; (2) complex installation and management.
6. Differentiate between load testing and stress testing.
Model answer
Load testing checks how the system performs under the expected number of users (e.g. 2 000 students checking results at once) – response time, throughput. Stress testing pushes the system beyond its normal limit to find the breaking point and see how it fails and recovers.
7. What is network topology? Explain bus, ring, star, tree and mesh topologies with diagrams, advantages and disadvantages.
Model answer
Network topology is the arrangement of nodes (computers, printers, switches) and the links connecting them. Physical topology is the cable layout; logical topology is how data flows.
- Bus: all nodes on one backbone cable with terminators. + cheap, little cable. − backbone failure stops all; collisions; hard to troubleshoot; poor scalability.
- Ring: each node linked to two neighbours forming a loop; token passing. + no collisions, equal access. − one failure can break the ring; adding nodes disrupts the network.
- Star: all nodes connected to a central switch. + easy to add/remove nodes, one cable fault affects one node, easy troubleshooting. − the switch is a single point of failure; more cable.
- Tree: stars connected hierarchically under a root. + highly scalable, suits departments. − failure of the root/backbone disconnects branches.
- Mesh: each node connected to many/all others; links = n(n − 1)/2. + most reliable, multiple paths. − expensive and complex. (Hybrid combines two or more of these.) Diagrams: bus – line with nodes hanging; ring – circle of nodes; star – switch in centre with spokes; tree – root at top with branches; mesh – every node joined to every other.
8. Explain how a network can be prepared to run with high reliability and to scale well. Include load/system testing and cloud computing in your answer.
Model answer
Reliability: (1) redundancy – backup links (mesh/dual ring), spare switches, RAID disks, backup power (UPS/generator); (2) failover – a standby server takes over automatically; (3) regular backups and monitoring with alerts; (4) choose topologies without single points of failure. Scalability: (1) use star/tree designs with stackable switches and structured cabling; (2) load balancers distribute requests across many servers; (3) horizontal scaling (add servers) and vertical scaling (upgrade a server). Testing: load testing measures performance at expected traffic; stress testing finds the breaking point; system testing checks the whole system end-to-end against requirements – tools like JMeter simulate thousands of users. Cloud computing: auto-scaling (elasticity) adds or removes servers with demand; data is replicated in several data centres; providers guarantee availability (e.g. 99.9%) through redundancy, so even a school website can survive result-day traffic.
This lesson ends with a quiz of 13 MCQs. Open the lesson on PrepOasis to take it and save your progress.
Previous: 1.3 Systems Design: SDLC and Software Development Methodologies · Next: 1.5 Cyber Security, Attacks and Encryption
All lessons in Computer Science XI (HSSC-I) – FBISE Complete SLO-Based Course