1.2 Network Topologies & Connection Types

Back to Computer Networks Notes

📚Module 01: Introduction & Network ArchitecturesTopic 1.2⏱️9 min read

🎯High-Yield For:Semester Exams (All Universities) • Placement Technical Rounds • Engineering Foundations

💡 Core Intuition

🍳 The Everyday Analogy: City Transportation & Airline Hubs

Imagine designing a transport network connecting 10 major global cities:

• Strategy A (Full Point-to-Point Mesh): You pave an independent, dedicated, multi-lane private highway between every single pair of cities. City 1 gets a direct highway to City 2, City 3... through City 10.
  • The Upside: Cars traveling from City 1 to City 2 never experience traffic jams caused by commuters traveling between City 5 and City 6. If one highway collapses under a landslide, every other city pair remains completely accessible.
  • The Downside: The construction cost is staggering! Paving $\frac{10 \times 9}{2} = 45$ dedicated expressways requires massive land, asphalt, and maintenance budgets.
• Strategy B (The Airline Hub-and-Spoke / Star): You designate one central city as the supreme mega-airport (the Hub). Every other city only needs one single runway connecting it to the central hub.
  • The Upside: You only build $9$ runways instead of $45$. Adding an 11th city requires only one new flight path.
  • The Downside: The central hub is a catastrophic Single Point of Failure (SPOF). If a blizzard shuts down the central hub, global transportation completely halts.
• Strategy C (The Shared Commuter Train / Bus): You lay down a single long continuous railroad track running from East to West. Every town builds a short spur connecting to this main track.
  • The Upside: Extremely cheap and simple to build.
  • The Downside: All trains share one physical track. If an explosion destroys a rail bridge in the middle, trains on both sides are completely cut off.

Three Architectural Philosophies: Mesh vs Star vs Bus

Contrasting direct point-to-point mesh, centralized star switching, and shared bus backbones

Direct P2P

Full Mesh Topology

🕸️

Dominant Architecture / DomainPoint-to-Point Mesh

• •Dedicated physical cable between every pair of nodes.
• •Zero traffic contention; ultimate fault isolation.
• •Astronomical O(n²) cabling cost and port density.

"Dedicated expressways everywhere: maximum resilience at maximum cost."

Centralized

Star Topology

⭐

Dominant Architecture / DomainSwitch-Centric LAN

• •Every device connects via one link to a central switch.
• •Linear O(n) cabling; effortless installation and diagnosis.
• •Single Point of Failure: switch crash kills the LAN.

"Hub-and-spoke airport: centralized efficiency with single-point dependency."

Shared Medium

Bus Topology

🚌

Dominant Architecture / DomainMulti-drop Backbone

• •One continuous shared cable snaking through all stations.
• •Cheapest cable cost; terminated by 50Ω resistors at ends.
• •Backbone cut or short circuits paralyze all communication.

"Shared commuter train: cheapest infrastructure with shared vulnerability."

💻 Bridging to Computer Science

In computer networks, the geometric arrangement of devices and transmission media is called the Physical Topology. How nodes attach to transmission lines is governed by the Line Configuration (or Connection Type).

Choosing the right network topology represents a classic systems engineering trade-off balancing Cabling Cost, Hardware Port Density, Fault Tolerance, and Ease of Troubleshooting.

---

📑Table of Contents

• 🔌 Line Configuration: Point-to-Point vs Multipoint
  • 1. Point-to-Point Connection
  • 2. Multipoint (Multi-drop) Connection
• 📐 The 4 Fundamental Physical Topologies
• 🕸️ 1. Mesh Topology: Mathematical Rigor & Mechanics
  • Mathematical Formulas
  • Advantages
  • Disadvantages
• ⭐ 2. Star Topology: Central Controller Architecture
  • Operational Mechanics
  • Mathematical Formulas
  • Advantages
  • Disadvantages
• 🚌 3. Bus Topology: Shared Backbone & Multi-drop Architecture
  • Components & Physical Mechanics
  • Advantages
  • Disadvantages
• ⭕ 4. Ring Topology: Token Circulating Loop
  • Operational Mechanics
  • Advantages
  • Disadvantages
• 🌳 5. Hybrid Topologies: Hierarchical Trees & Star-Bus
• 📊 Master Comparative Topology Matrix
• 🏭 Real-World Production Context
  • Why Modern Hyperscale Cloud Data Centers Use Leaf-Spine (Clos) Topologies
• 🎯 Exam & Interview Pitfall Check

---

🔌 Line Configuration: Point-to-Point vs Multipoint

Before examining geometric topologies, we must examine how two or more devices physically attach to a communication circuit:

Line Configuration: Point-to-Point vs Multipoint

Comparing dedicated private channels against shared multi-drop physical circuits

Dedicated

Point-to-Point Connection

🔗

Dominant Architecture / DomainDirect Link (Unicast)

• •Provides an exclusive, dedicated physical link between exactly two communicating devices.
• •100% of channel capacity is reserved solely for transmission between those two endpoints.
• •Zero interference or contention from third-party stations.
• •Examples: Workstation to switch port, microwave tower link, satellite ground link.

"Private direct tunnel: guaranteed bandwidth and zero contention."

Shared

Multipoint (Multi-drop) Connection

📻

Dominant Architecture / DomainShared Medium (Broadcast / Multicast)

• •More than two specific devices share a single physical transmission medium.
• •Channel capacity is shared spatially (frequency bands) or temporally (time slots).
• •Requires media access control (CSMA/CD, tokens) to prevent collisions.
• •Examples: Coaxial Ethernet bus, Wi-Fi radio airwaves, cellular base station towers.

"Public commons: high hardware efficiency requiring access arbitration."

1. Point-to-Point Connection

A Point-to-Point connection provides an exclusive, dedicated physical transmission link between exactly two communicating devices:

• The entire channel capacity of the link is permanently reserved for transmission between those two endpoints.
• Examples:
  • A workstation connected directly to a network switch port via an Ethernet patch cable.
  • A long-haul microwave relay link between two telecommunication towers.
  • A satellite link bridging two earth stations.

2. Multipoint (Multi-drop) Connection

A Multipoint connection (historically called Multi-drop) is one in which more than two specific devices share a single physical transmission link:

• The capacity of the channel is shared among multiple attached devices either:
  • Spatially: If several devices can utilize the link concurrently (e.g., frequency division multiplexing).
  • Temporally: If devices must take turns transmitting in time-shared slots (e.g., classical Ethernet CSMA/CD, Wi-Fi radio airwaves).
• Examples:
  • Classical 10BASE5 thick-net Ethernet bus running through an office corridor.
  • Wireless radio channels shared by multiple smartphones registered to a cellular base tower.

---

📐 The 4 Fundamental Physical Topologies

A Physical Topology is the geometric representation of the relationship of all physical links and connecting nodes to one another.

Comparative architectural summary of geometric configurations, link requirements, and failure modes

🕸️

Mesh Topology

O(n²) Maximum Redundancy

• Every device has a dedicated point-to-point link to every other device.
• Zero traffic congestion: each link carries its own private data load.
• Guaranteed security: physical lines prevent intermediate eavesdropping.
• Exorbitant cabling bulk and expensive hardware I/O port density.

Key Abstractions / Calls

Links: n(n-1)/2Ports/Node: n-1Total Ports: n(n-1)

⭐

Star Topology

O(n) Centralized Dominance

• Devices connect via point-to-point links only to a central hub or switch.
• Far less cabling than mesh; effortless installation and node addition.
• Robust against node link cuts: one severed cable isolates only that node.
• Single Point of Failure (SPOF): if the central switch fails, the LAN dies.

Key Abstractions / Calls

Links: nPorts/Node: 1Switch Ports: n

🚌

Bus Topology

Shared Backbone Multi-drop

• A single long continuous cable acts as a shared backbone for all nodes.
• Nodes connect via drop lines and taps; ends terminate with 50Ω resistors.
• Extremely economical cable consumption for small localized labs.
• Catastrophic failure: a single backbone break halts all network transmission.

Key Abstractions / Calls

Links: 1 BackboneTaps: nTerminators: 2

⭕

Ring Topology

Token Circulating Loop

• Each device connects in a point-to-point loop only to its 2 immediate neighbors.
• Signals travel unidirectionally; each node acts as an active signal repeater.
• Deterministic delay: token-passing access prevents signal packet collisions.
• A single cable break or unpowered station disables the entire ring.

Key Abstractions / Calls

Links: nPorts/Node: 2Direction: Unidirectional

---

🕸️ 1. Mesh Topology: Mathematical Rigor & Mechanics

In a Mesh Topology, every single network device has a dedicated point-to-point link to every other device in the network.

Architecture Flow

Step 1 of 6

Mesh Topology: Complete Point-to-Point Interconnection (4-Node Cluster)

Every autonomous node maintains n - 1 = 3 dedicated physical duplex links, totaling n(n-1)/2 = 6 cables

1Duplex Link 1 (N1 ↔ N2)

2Duplex Link 2 (N1 ↔ N3)

3Cross Link 3 (N1 ↔ N4)

4Cross Link 4 (N2 ↔ N3)

5Duplex Link 5 (N2 ↔ N4)

6Duplex Link 6 (N3 ↔ N4)

💻3 Dedicated Links

Node 1 (Workstation)

Terminal Station

🖥️3 Dedicated Links

Node 2 (Server)

Database Node

💽3 Dedicated Links

Node 3 (Storage)

SAN Storage Host

🌐3 Dedicated Links

Node 4 (Gateway)

Edge Router

💡Click or hover any card or transition arrow above to inspect deep-dive operational mechanics

Mathematical Formulas

For a network consisting of $n$ physical nodes:

1. Number of Duplex Physical Links: Every node must connect to $(n - 1)$ other nodes. Since each duplex link is shared between two endpoints, the total number of links is: $\text{Physical Links (Duplex)} = \frac{n(n - 1)}{2} = O(n^2)$

2. Number of Simplex Links: If the links are unidirectional (simplex), each direction requires a distinct physical conductor: $\text{Physical Links (Simplex)} = n(n - 1)$

3. I/O Ports Required per Device: Each device must have a dedicated physical network interface port for every peer: $\text{Ports per Device} = n - 1$

4. Total I/O Ports across Entire Network: $\text{Total Network Ports} = n \times (n - 1)$

Example Calculation:

For an enterprise server cluster with $n = 10$ nodes:

• Total Duplex Links $= \frac{10 \times (10 - 1)}{2} = \frac{90}{2} = 45\text{ physical cables}$.
• I/O Ports per Server $= 10 - 1 = 9\text{ NIC ports}$.
• Total Ports across the cluster $= 10 \times 9 = 90\text{ NIC ports}$.

Advantages

1. Zero Traffic Contention: Because each link is completely dedicated, each pair of nodes communicates without competing for bandwidth with third-party transmissions.
2. Extreme Robustness & Fault Tolerance: If a physical cable between Node 1 and Node 2 is severed, traffic between all other node pairs continues unaffected. Furthermore, Node 1 can route data to Node 2 indirectly via Node 3 or Node 4.
3. Ironclad Privacy & Physical Security: When Node 1 transmits a message to Node 2, the electromagnetic signal physically traverses only the dedicated cable connecting them. Nodes 3 and 4 cannot intercept or sniff the frames.
4. Trivial Fault Isolation: If Node 1 cannot ping Node 2, the network engineer immediately knows that the single dedicated link between Node 1 and Node 2 is faulty.

Disadvantages

1. Astronomical Cabling Footprint ($O(n^2)$): As $n$ grows, the number of cables explodes quadratically. Installing 4,950 cables for a modest 100-node office is physically and architecturally impossible.
2. Prohibitive Hardware Costs: High-speed network interface cards with 10 or 20 ports per server are extraordinarily expensive.
3. Complex Installation & Cable Routing: Routing hundreds of cables through wall conduits, ceilings, and raised data center floors creates severe maintenance nightmares.

---

⭐ 2. Star Topology: Central Controller Architecture

In a Star Topology, each device has a dedicated point-to-point link only to a central network controller—traditionally called a Hub or Layer 2 Switch. Devices are never directly wired to one another.

Architecture Flow

Step 1 of 4

Star Topology: Central Switch Architecture (Radial LAN)

All stations connect exclusively to the central Layer 2 switch via dedicated point-to-point links

1Dedicated Link (Port 1)

2Dedicated Link (Port 2)

3Dedicated Link (Port 3)

4Dedicated Link (Port 4)

💻Station 1

Host 1 (Port 1)

Workstation Station

💻Station 2

Host 2 (Port 2)

Database Station

⭐Central Controller

Central Switch (L2)

High-Speed Switching Matrix

🖥️Station 3

Host 3 (Port 3)

Finance Host

🖥️Station 4

Host 4 (Port 4)

Engineering Host

💡Click or hover any card or transition arrow above to inspect deep-dive operational mechanics

Operational Mechanics

• The central switch acts as a high-speed hardware exchange matrix.
• If Host 1 wants to send an Ethernet frame to Host 2, Host 1 transmits the frame to the central switch.
• In a modern Switch, the hardware reads the destination MAC address from the frame header and forwards the frame strictly to the physical port connected to Host 2.
• In a legacy Hub, the incoming electrical signal was simply amplified and blindly broadcast out to all other ports (creating a shared collision domain).

Mathematical Formulas

For $n$ hosts:

• Total Physical Links: $n$ links ($O(n)$ linear complexity).
• Ports Required per Host: Exactly $1$ NIC port.
• Ports Required on Central Switch: $n$ ports.

Advantages

1. Linear Cost & Cabling ($O(n)$): Requires far less cable than a mesh topology. Connecting 100 computers requires only 100 cables.
2. Effortless Installation & Scalability: Adding a new computer involves running one patch cable from the desk to the central switch patch panel.
3. High Node Fault Tolerance: If Host 1's cable is cut by an office chair wheel, only Host 1 loses connectivity. Host 2, Host 3, and Host 4 continue communicating at full line speed.
4. Centralized Monitoring & Management: Enterprise network administrators can inspect port traffic, apply VLAN tags, and disable malicious ports from the central switch management console.

Disadvantages

1. The Single Point of Failure (SPOF): The entire network depends on the central controller. If the switch power supply blows out or the motherboard crashes, the entire local area network collapses completely.
2. Higher Cabling than Bus: While far cheaper than mesh, star topology requires more cabling than a bus topology because every single workstation must run an independent home-run cable all the way back to the telecommunications closet.

---

🚌 3. Bus Topology: Shared Backbone & Multi-drop Architecture

A Bus Topology is a multipoint architecture. One long continuous cable acts as a shared central backbone to link all devices in the network.

Architecture Flow

Step 1 of 4

Bus Topology: Shared Backbone & Multi-drop Architecture

Shared coaxial transmission line with drop lines, BNC T-taps, and 50Ω terminating resistors

1Left Segment Termination

2Drop Line 1 (Host 1 Tap)

3Drop Line 2 (Host 2 Tap)

4Right Segment Termination

🛑End Boundary

50Ω Terminator (Left)

End Boundary Absorber

💻Drop Line 1

Host 1 (BNC Tap)

Top Transceiver Station

🔌Shared Medium

Shared Coaxial Bus

Common Copper Channel

💻Drop Line 2

Host 2 (BNC Tap)

Bottom Transceiver Station

🛑End Boundary

50Ω Terminator (Right)

End Boundary Absorber

💡Click or hover any card or transition arrow above to inspect deep-dive operational mechanics

Components & Physical Mechanics

1. Backbone Cable: Traditionally thick or thin coaxial cable (such as RG-58 10BASE2).
2. Drop Line: A short cable segment running from the individual workstation to the main backbone.
3. Tap (Connector): A piercing connector or BNC T-connector that splices into the main cable to establish metallic contact with the central copper conductor.
4. Terminating Resistors (Terminators): Crucial! Both physical ends of the bus cable must be capped with a resistor (typically $50\,\Omega$ characteristic impedance) matching the cable's impedance.
   • Why? If a cable end is left open, high-frequency electrical signals hit the open circuit boundary and reflect backwards (like an echo shouting down a hallway), causing signal reflection that collides with and corrupts incoming valid frames.

Advantages

1. Minimal Cable Consumption: The backbone simply snakes down a hallway from room to room. Drop lines are short.
2. Cost-Effective for Small Setups: Extremely cheap to install initially for a small classroom or temporary laboratory.

Disadvantages

1. Difficult Fault Identification & Isolation: If the network stops working, finding which tap has oxidized or which drop line is shorted requires disconnecting stations one by one with a time-domain reflectometer (TDR).
2. Catastrophic Cable Break: A single physical cut anywhere along the central backbone cable halts all communication across the entire network. Why? The severed cut creates two unterminated cable ends, causing massive signal reflections that swamp all electrical pulses.
3. Shared Collision Domain & Scalability Limits: All devices share the same wire. Signal attenuation (weakening over distance) limits the maximum backbone length and the number of allowed taps.

---

⭕ 4. Ring Topology: Token Circulating Loop

In a Ring Topology, each device has a dedicated point-to-point connection with only the two devices on either side of it. The connections form a continuous closed physical loop.

Architecture Flow

Step 1 of 4

Ring Topology: Unidirectional Token Circulation Loop

Closed circular loop: active repeaters regenerate bitstreams and pass token authorization clockwise

11. Clockwise Frame Transit

22. Regenerated Signal

33. Confirmed Frame Return

44. Return to Sender & Release Token

💻Token Holder

Station 1 (Transmitter)

Active Token Holder

💻Active Repeater

Station 2 (Repeater)

Signal Regenerator

🖥️Destination Host

Station 3 (Target Sink)

Destination Workstation

💻Active Repeater

Station 4 (Repeater)

Signal Regenerator

💡Click or hover any card or transition arrow above to inspect deep-dive operational mechanics

Operational Mechanics

• Unidirectional Flow: Signals circulate around the ring in one specified direction (e.g., clockwise) from station to station.
• Active Regenerative Repeaters: Every station houses an active hardware repeater inside its network interface card. When a station receives a bitstream intended for another station, its repeater regenerates the weakened electrical pulses to full voltage and retransmits them to the next neighbor.
• Token Passing Protocol (IEEE 802.5): A special bit pattern called the Token circulates continuously. A station can transmit data only when it captures the free token. After its frame circles the entire ring, confirms receipt, and returns to the sender, the sender releases the token back into circulation.

Advantages

1. Deterministic Latency (Zero Collisions): Because only the token holder can transmit, packet collisions are mathematically impossible. Every station has an upper-bounded, guaranteed maximum wait time before it receives the token.
2. Simplified Fault Isolation: Modern ring networks (such as FDDI - Fiber Distributed Data Interface) use dual counter-rotating rings. If a link snaps, the two stations on either side of the break automatically wrap the primary ring onto the secondary ring, healing the loop into a single operational circuit.

Disadvantages

1. Vulnerability to Single Break: In a standard single ring, a severed cable or an unpowered computer breaks the closed loop and paralyzes the entire network.
2. Disruptive Reconfiguration: Physically inserting a new computer or unplugging a workstation breaks the ring, momentarily interrupting network traffic for all active nodes.

---

🌳 5. Hybrid Topologies: Hierarchical Trees & Star-Bus

In real-world enterprise environments, pure topologies are rarely deployed in isolation. Instead, architects combine them into Hybrid Topologies:

Architecture Flow

Step 1 of 6

Tree Topology: Hierarchical Branching Switch Architecture

Multi-level tree: central root switch branches into intermediate distribution switches and edge hosts

1Core Trunk A (Core ↔ Dist A)

2Core Trunk B (Core ↔ Dist B)

3Access Link A1

4Access Link A2

5Access Link B1

6Access Link B2

🌐Core Switch

Root Core Switch

Tier 1: Enterprise Core

🔀Dept A Switch

Distribution Switch A

Tier 2: Dept A Aggregator

🔀Dept B Switch

Distribution Switch B

Tier 2: Dept B Aggregator

💻Dept A Node

Host A1 (Workstation)

Tier 3: Desk Host

💻Dept A Node

Host A2 (Workstation)

Tier 3: Desk Host

🖥️Dept B Node

Host B1 (Workstation)

Tier 3: Desk Host

🖥️Dept B Node

Host B2 (Workstation)

Tier 3: Desk Host

💡Click or hover any card or transition arrow above to inspect deep-dive operational mechanics

Hierarchical Tree Topology: Enterprise 3-Tier Design

Hierarchical star aggregation structuring enterprise LANs from edge desks to core backbones

Core Tier

Core Layer (Root Switch / Router)

🌐

High-speed optical backbone switching traffic between distribution blocks at line rate.

Terabit Optical BackbonesDual Redundant CoresZero Packet Filtering

↓Aggregated High-Speed Uplinks

Distribution Tier

Distribution Layer (Aggregation Switches)

🔀

Applies routing policies, access control lists (ACLs), QoS, and inter-VLAN routing.

Departmental VLANsSecurity FirewallsRedundant Uplink Balancing

↓VLAN Trunk Links

Access Tier

Access Layer (Workstation Switches)

🔌

Connects end-user computers, Wi-Fi access points, and IP phones to the local network.

Cat6 Patch CablesPower over Ethernet (PoE)802.1X Port Security

• Tree Topology (Hierarchical Star): Workstations connect to local access switches (forming small stars), which in turn connect up to distribution switches, which aggregate into a central core switch (the root of the tree).
• Star-Ring Topology: IBM Token Ring networks physically deployed stations in a star configuration using a central wiring hub called a MSAU (Multistation Access Unit). While the physical cables radiated outward like a star, the internal circuitry of the MSAU routed the bits in a continuous circular ring!

---

📊 Master Comparative Topology Matrix

Metric / Feature	Mesh Topology	Star Topology	Bus Topology	Ring Topology
Line Configuration	Dedicated Point-to-Point	Dedicated Point-to-Point	Shared Multipoint	Dedicated Point-to-Point
Duplex Links Formula	$\frac{n(n - 1)}{2}$	$n$	$1\text{ backbone} + n\text{ drops}$	$n$
I/O Ports per Node	$n - 1$	$1$	$1\text{ (via tap)}$	$2$ (In & Out)
Cabling Consumption	Maximum ($O(n^2)$)	Moderate ($O(n)$)	Minimum	Moderate ($O(n)$)
Installation Difficulty	Extremely Difficult	Very Easy	Easy	Moderate
Fault Isolation	Trivial (isolated link)	Very Easy (switch port)	Extremely Difficult	Easy (token diagnostics)
Single Link Failure Impact	Negligible (rerouted)	Isolated to single host	Whole network dead (backbone break)	Whole network dead (single ring)
Single Point of Failure (SPOF)	None	Central Hub / Switch	Central Backbone Cable	Any Cable / Node Break
Primary Production Use	WAN Core, SAN Meshes	Modern Office LANs, Wi-Fi	Legacy 10BASE2, Avionics	Industrial Automation, FDDI

---

🏭 Real-World Production Context

Why Modern Hyperscale Cloud Data Centers Use Leaf-Spine (Clos) Topologies

For decades, enterprise data centers used a 3-Tier Tree Architecture (Access $\to$ Aggregation $\to$ Core).

The Problem with Trees in Modern Cloud Workloads:

• In cloud environments (Google Cloud, AWS, Azure), modern microservices communicate predominantly East-West (server-to-server traffic inside the data center) rather than North-South (client-to-internet traffic).
• In a classical tree, if Server A in Rack 1 communicates with Server B in Rack 50, all packets must travel all the way up to the Core Switch root and back down. The core switch becomes a choking bottleneck, and latency varies wildly.

The Solution: Clos Leaf-Spine Topologies

Modern cloud data centers deploy a 2-Tier Clos Leaf-Spine Mesh:

1. Every Leaf Switch (Top-of-Rack switch) connects to every single Spine Switch in a partial mesh.
2. No spine switches connect to other spines; no leaf switches connect to other leaves.
3. The Result: Any server in the entire 50,000-machine data center is exactly 3 network hops away from any other server ($O(1)$ constant latency). If two spine switches explode, the Equal-Cost Multi-Path (ECMP) routing simply shifts packet flows to the remaining spines with zero downtime!

---

🎯 Exam & Interview Pitfall Check

Core Conceptual Questions

Question 1: "A company wants to connect 8 branch offices in a fully connected Mesh topology. Calculate: (a) The total number of duplex links required, (b) The number of I/O ports required per branch office router, and (c) The total number of ports across the entire organization."
Step-by-Step Solution:

• Given: $n = 8$ nodes.
• (a) Total Duplex Links: $\text{Links} = \frac{n(n - 1)}{2} = \frac{8 \times 7}{2} = \frac{56}{2} = 28\text{ duplex links}$
• (b) Ports per Device: $\text{Ports} = n - 1 = 8 - 1 = 7\text{ I/O ports}$
• (c) Total Ports across Network: $\text{Total Ports} = n \times (n - 1) = 8 \times 7 = 56\text{ ports}$

Question 2: "Explain why terminating resistors are mandatory at both ends of a Bus topology cable. What physical phenomenon occurs if one terminator is removed?"
Key Focus Points:

• Transmission cables have a characteristic impedance (e.g., $50\,\Omega$).
• When an electromagnetic wave reaches an abrupt physical boundary (an open or short circuit at an unterminated end), energy cannot simply disappear. By basic physics, the wave reflects back along the copper wire.
• This reflected signal collides with outbound valid signals, distorting voltages and corrupting frames. Terminators absorb the electrical wave's energy completely without reflection.

Question 3: "Compare Star and Mesh topologies in terms of robustness and cost."
Key Focus Points:

• Robustness: Mesh is superior because it contains multiple redundant paths and zero central single point of failure. Star is vulnerable to central switch hardware failure.
• Cost: Star is dramatically cheaper ($O(n)$ cables, $1$ port per host) compared to Mesh ($O(n^2)$ cables, $n-1$ ports per host).

Common Interview Traps

Trap 1: "Is a legacy Ethernet Hub a Star or a Bus topology?"
Answer: It is physically a Star, but logically a Bus!
Physically, cables run from every computer to a central box (forming a physical star). However, inside the hub's physical circuitry, all input pins are electrically soldered together onto a shared common electrical bus. When one computer transmits, the hub repeats the signal to all ports simultaneously. Hence: Physical Star, Logical Bus.

Trap 2: "If a question asks for the number of links in a Mesh topology without mentioning duplex or simplex, what should you assume?"
Answer: Standard academic computer networking literature (Forouzan, Tanenbaum) always assumes duplex (bidirectional) links by default, using the formula $\frac{n(n - 1)}{2}$. However, in a university exam or interview, always explicitly state: "Assuming duplex links: $\frac{n(n - 1)}{2}$; if simplex links were required: $n(n - 1)$." This demonstrates complete mastery.

---

🗺️View Complete Computer Networks Roadmap & Syllabus

← Previous Topic1.1 Network Fundamentals & Data CommunicationNext Topic →1.3 The OSI Reference Model & 7-Layer Architecture