Data Link Layer & Switching

1. What is the key functional difference between a Collision Domain and a Broadcast Domain?

• Collision domain — a network segment where two devices transmitting at the same time collide. On a hub (shared medium), all ports are one collision domain.
• Broadcast domain — the boundary within which a broadcast frame reaches every device. Broadcasts stop at routers (or VLAN boundaries).

Hub:   one collision domain + one broadcast domain
Switch (per port): each port its own collision domain, but one broadcast domain
Router: terminates the broadcast domain

Key takeaway:

• Switches break up collision domains (fewer collisions).
• Routers break up broadcast domains (less broadcast noise).

2. Which of the following accurately describes the differences between a MAC address and an IP address?

• MAC address — a 48-bit hardware address burned into the network interface card (NIC). Operates at Layer 2, permanent, unique to the hardware. It never changes, no matter which network the device joins.
• IP address — a logical address assigned by software. Operates at Layer 3 and changes depending on the network location (home Wi-Fi, office, mobile data all give different IPs).

	MAC	IP
Size	48-bit	32-bit (v4) / 128-bit (v6)
Layer	2 (Data Link)	3 (Network)
Source	Burned into NIC	Assigned by software
Changes	Never	With network

Think: MAC is your fingerprint, IP is your current address.

3. How does the Address Resolution Protocol (ARP) function on a local network?

ARP resolves a known IP address to an unknown MAC address on the local network.

The problem: to send a frame at Layer 2, you need the destination’s MAC, but you only know its IP.

How ARP solves it:

1. Device broadcasts an ARP Request: “Who has IP 192.168.1.50? Tell me your MAC.”
2. Every device hears it; the owner of that IP replies with an ARP Reply: “That’s me, my MAC is AA:BB:CC:DD:EE:FF.”
3. The requester caches this IP→MAC mapping for future use.

ARP works only within the local network — it’s what gets you from Layer 3 (IP) down to Layer 2 (MAC).

4. What catastrophic network event does Spanning Tree Protocol (STP) prevent by blocking redundant switch ports?

STP prevents broadcast storms.

In a network with redundant switch links, the switches form a loop. A broadcast frame loops forever:

Switch A ──→ Switch B ──→ Switch C ──→ Switch A ──→ (endless loop)

Each hop re-broadcasts it, so frames multiply until the network is saturated — a broadcast storm that effectively crashes the network.

STP’s solution: it detects loops and blocks redundant ports, leaving exactly one active path between any two switches. If the active path fails, STP unblocks a backup. Redundancy is preserved, loops are eliminated.

5. What is a Virtual Local Area Network (VLAN), and what are its primary administrative benefits?

A VLAN is a logical partition of a physical switch into separate, isolated virtual networks.

One physical switch divided into:
  VLAN 10 (Sales)   — port 1-5
  VLAN 20 (IT)      — port 6-10
  VLAN 30 (HR)      — port 11-15

Broadcasts and traffic stay within their VLAN — it’s as if you have separate switches.

Benefits:

• Security — departments are isolated; VLAN 20 traffic never reaches VLAN 10.
• Less broadcast noise — broadcasts only hit their own VLAN.
• Flexibility — re-organizing users is a config change, not a cable change.

6. Which variant of the Carrier Sense Multiple Access (CSMA) protocol is deployed in wireless networks, and how does it function?

Wireless uses CSMA/CA — Collision Avoidance.

Why not CSMA/CD? Wired Ethernet can detect collisions (listen while transmitting — voltage conflicts are obvious). Radios can’t easily do that: the transmitter is drowning out the receiver, and you can’t reliably “hear” a collision mid-transmission. So wireless avoids collisions instead of detecting them.

CSMA/CA process:

1. Listen — check if the channel is clear.
2. Reserve — send a short RTS (Request to Send); the AP answers with CTS (Clear to Send).
3. Transmit — only after the channel is confirmed free.
4. Wait for ACK — retry if no acknowledgment.

Client → [RTS] → AP → [CTS] → Client → data → AP → [ACK]

Proactive handshaking keeps wireless collisions rare — at the cost of extra control overhead.

7. What is the primary function of the Beaconing process within historical local network loops (like Token Ring or FDDI)?

Beaconing is a self-healing diagnostic — a node broadcasts a beacon frame to locate a cable break or failed node.

Ring:  A → B → C → D → A
B detects failure with C, sends beacon:
       "I am B, I can't reach C, the fault is between us"
Ring "wraps" around the break → traffic continues

When a node loses communication with its neighbor, it sends a beacon frame identifying the failed segment. Every other node sees it, learns the break’s location, and the ring rewraps to route around the fault. The network keeps running even with a broken link — a key design goal of ring topologies (FDDI had dual counter-rotating rings for exactly this).

8. How does error detection work — parity, checksum, and CRC?

Error detection adds redundancy so the receiver can spot corruption. Three levels of sophistication:

• Parity bit — add 1 bit so the total number of 1s is even (even parity) or odd. Catches any single-bit error but fails on two-bit errors (they cancel out). Simple but weak.
• Checksum — sum the data words, carry the sum along, and the receiver verifies the total. The IP and TCP/UDP headers use checksums. Cheap, but weaker than CRC — it can miss some multiple-bit errors.
• CRC (Cyclic Redundancy Check) — treat the data as a polynomial and divide it by a fixed generator polynomial; append the remainder as the FCS (frame check sequence). The receiver divides again and expects remainder 0. CRC catches burst errors (runs of corrupted bits) very reliably — this is what Ethernet uses.

The ladder: parity catches single bits, checksum catches simple corruption, CRC catches bursts. Error detection only tells you something is wrong — it can’t fix it. Correction (retransmission via ARQ, or FEC) is a separate mechanism.

9. What are the ARQ protocols — Stop-and-Wait, Go-Back-N, and Selective Repeat?

ARQ (Automatic Repeat reQuest) is how the data link layer turns an error-prone link into a reliable one: the sender retransmits frames the receiver didn’t acknowledge.

• Stop-and-Wait — send one frame, wait for ACK, send the next. Simple and correct, but the link is idle during every wait — terrible utilization on long/high-latency links.
• Go-Back-N — a sliding window lets the sender transmit up to N frames before needing an ACK. If a frame is lost, the sender retransmits that frame and everything after it (the receiver discards the out-of-order frames). Efficient, but one loss wastes bandwidth on retransmissions.
• Selective Repeat — also a window, but the receiver buffers out-of-order frames and only the lost frame itself is retransmitted. Best bandwidth use, but needs more receiver buffer and sequencing logic.

Protocol	Sends	On error	Window
Stop-and-Wait	1 at a time	Retransmit the 1	No
Go-Back-N	Up to N	Retransmit lost + all after	Yes
Selective Repeat	Up to N	Retransmit only lost	Yes

Interview one-liner: Stop-and-Wait is reliable but slow; Go-Back-N wastes retransmissions; Selective Repeat retransmits precisely but costs complexity. All three sit at Layer 2 (and similar logic drives TCP).