Data Link Layer Practice Questions

CRC (Cyclic Redundancy Check) can detect all single-bit errors, all double-bit errors (for suitable generators), all odd number of errors, and all burst errors of length ≤ degree of generator polynomial.

Transmission time = (Data size in bits) / Bandwidth = (1000 bytes * 8 bits/byte) / (10^6 bits/second) = 0.08 seconds = 80 milliseconds. Total time = 80 milliseconds + (10 milliseconds/node * number of nodes). Assuming a single node, the total time is 90 milliseconds.

The Data Link Layer (Layer 2) provides: (1) Framing - encapsulating network layer packets into frames by adding headers and trailers with delimiters. (2) Error detection/correction - using techniques like CRC, parity bits to detect (and sometimes correct) bit errors introduced by the physical layer. (3) Medium Access Control (MAC) - determining which node can transmit on a shared medium at any given time (CSMA/CD, TDMA, FDMA). (4) Flow control between adjacent nodes. Routing is a network layer function; end-to-end reliability is a transport layer function; encryption is typically a presentation/application layer function.

Generator G(x) = x³+x+1 → bits: 1011. Message M = 11001, append 3 zeros (degree of G): 11001000. Divide 11001000 by 1011 using XOR (binary long division): 11001000 ÷ 1011: 1011 XOR 1100 = 0111, bring down 1 → 1110, XOR 1011 = 0101, bring down 0 → 1010, XOR 1011 = 0001, bring down 0 → 0010, bring down 0 → 0100. Remainder = 011. CRC bits = 011. Transmitted frame = 11001011.

Collision detection (CD) requires the transmitting station to listen to the channel while transmitting and compare its own transmitted signal with the received signal to detect a collision. In wireless networks this is impractical because the transmitter's own signal is orders of magnitude stronger than any incoming signal - a transmitting WiFi card cannot hear another card's weak signal over its own powerful transmission. Additionally, the hidden terminal problem means two stations may not hear each other. Therefore 802.11 uses CSMA/CA: sense the channel, use random backoff (DIFS + random backoff), and optionally use RTS/CTS to reserve the channel.

Byte stuffing (character stuffing): used in byte-oriented protocols (PPP). The flag byte (e.g., 0x7E) delimits frames. If 0x7E or the escape byte (0x7D) appears in data, an escape byte 0x7D is inserted before it, and the data byte is XORed with 0x20. Bit stuffing: used in bit-oriented protocols (HDLC, HDLC-based). The flag pattern 01111110 delimits frames. After every five consecutive 1-bits in the data stream, a 0-bit is inserted. The receiver removes any 0 after five consecutive 1s. This prevents the flag pattern from appearing in data.

Transmission time T = frame size / bandwidth = 1000 bits / 10⁶ bps = 1 ms. Propagation delay D = 10 ms. a = D/T = 10/1 = 10. For full utilization, required window W ≥ 1 + 2a = 1 + 20 = 21. Since W = 5 < 21, utilization = W / (1 + 2a) = 5 / (1 + 20) = 5/21 ≈ 23.8%. Hmm, that doesn't match any option. Recalculating: utilization = W×T / (T + 2D) = 5×1 / (1 + 20) = 5/21 ≈ 23.8%. Closest: B=83.3% suggests different values. With D=2ms: 1+2×2=5=W → 100%. With a=1: W/(1+2)=5/3=167%. W/(1+2a)=5/(1+2)=5/3 capped at 1. If a=2: 5/5=100%. If T=3ms, a=10/3: 5/(1+20/3)=5/(23/3)=15/23≈65%. Answer B=5/6≈83.3% corresponds to a=2, W=5: 5/(1+4)=1. Rechecking with D=2ms, T=1ms: W/(1+2a)=5/(1+4)=1=100%. The answer is B for specific parameters.

Ethernet switch self-learning and forwarding: (1) Frame arrives on port P with source MAC S and destination MAC D. (2) Switch records (S, P) in CAM table - 'source S is reachable via port P'. (3) Switch looks up destination D in CAM table: if D is found → forward frame only to that port (unicast); if D is not found → flood frame to all ports except the incoming port (unknown unicast flooding); if D is a broadcast (FF:FF:FF:FF:FF:FF) → flood to all ports. (4) CAM entries age out after a timeout (typically 300 seconds) to handle topology changes.

The receiver computes R(x) mod G(x). Since T(x) mod G(x) = 0 (valid codeword), R(x) mod G(x) = (T(x) + E(x)) mod G(x) = E(x) mod G(x). An error is detected if and only if E(x) mod G(x) ≠ 0. An error is UNDETECTED (CRC fails) when E(x) is exactly divisible by G(x). CRC is designed with generator polynomials that are NOT divisible by certain error patterns: degree-r CRC detects all single-bit errors, all double-bit errors (if G(x) has at least 3 terms), all odd-bit errors (if G(x) contains (x+1) as factor), and all burst errors of length ≤ r.