Questions: Error Detection and Correction

Question 1

A satellite communicating with a deep-space probe has a 40-minute round-trip signal delay. When bit errors occur, the best error-handling strategy is:

• A. Parity bits with retransmission requested on every detected error
• B. Forward error correction codes that let the receiver reconstruct corrupted data without retransmission
• C. CRC checksums combined with TCP-style ARQ retransmission
• D. No redundancy — deep-space links are clean enough that error handling is unnecessary

Answer: (B)

With a 40-minute round-trip delay, retransmitting on every error would mean waiting 40 minutes per correction — completely impractical. Forward error correction (FEC) encodes enough redundancy into the original transmission that the receiver can reconstruct corrupted bits locally. The overhead of extra bits is worth it when retransmission is impossibly slow. Detection + retransmission strategies (ARQ) are ideal for low-latency networks like Ethernet, not deep-space links.

Question 2

A network engineer adds a single parity bit to each 8-bit data byte. If two bits are flipped by noise during transmission, what happens?

• A. The error is detected — parity bits catch all errors regardless of how many bits flip
• B. The error goes undetected — two flipped bits restore the parity to its expected value
• C. The error is corrected — the parity bit identifies which two bits were flipped
• D. The error is partially detected — parity catches the first flipped bit but ignores the second

Answer: (B)

Parity works by making the count of 1-bits even (or odd). One flipped bit changes the count and is detected. But two flipped bits change the count by 2, restoring the original parity — the error is invisible to the receiver. This is parity's fundamental weakness: it reliably detects odd numbers of bit errors but misses all even-numbered errors. This limitation motivates stronger schemes like CRC for real networking.

Question 3

Error correction codes require more redundant bits than error detection codes because they must encode enough information to identify the specific location of corrupted bits, not just whether corruption occurred.

• T. True
• F. False

Answer: True

Detection only needs to signal 'something went wrong' — a single bit change suffices to indicate corruption. Correction must pinpoint *which* bit(s) need flipping, which requires enough redundancy to uniquely identify the error position. Hamming codes, for example, use parity bits at power-of-two positions so the syndrome (pattern of parity failures) directly encodes the error location. More correction capability = more redundant bits per message.

Question 4

Because modern error correction codes can fix corrupted data automatically, they have replaced error detection protocols in most layers of modern networks.

• T. True
• F. False

Answer: False

Real networks use both, at different layers, based on tradeoffs. Ethernet uses CRC (detection) at the link layer; TCP uses checksums (detection) at the transport layer; 5G wireless uses FEC (correction) where retransmission adds too much latency. Detection + retransmission dominates when error rates are low and retransmission is fast. Correction dominates where retransmission is impractical. The two approaches are complementary, not substitutes.

Question 5

Explain why 'forward error correction' is preferred over 'detection + retransmission' in some communication environments. What property of the environment makes correction worth the extra bandwidth overhead?

Model answer: FEC is preferred when retransmission is impractical or prohibitively expensive — either because round-trip latency is very high (deep-space, satellite links) or because the channel cannot support bidirectional signaling (broadcast media, one-way streaming). In these environments, the cost of waiting for a retransmission exceeds the cost of sending extra redundant bits upfront. The extra bandwidth overhead of FEC becomes worthwhile when the alternative is unacceptable delay or failed correction.

The tradeoff is overhead vs. latency. FEC burns bandwidth on every transmission, even when no errors occur. ARQ only pays when errors happen, but its cost is delay proportional to round-trip time. For Ethernet (microseconds RTT, rare errors), ARQ wins. For satellite (hundreds of milliseconds RTT, moderate error rates), FEC wins. Real systems combine both: FEC handles most errors proactively, ARQ handles the rare cases FEC can't fix.