Sample questions
Nc Transport ReliabilityDifficulty 2
A game uses a 16-bit packet sequence number that wraps from 65535 back to 0. Which comparison rule correctly decides which of two numbers is more recent?
- aWhichever value is numerically larger is always newer
- bSubtract mod 65536, and a result under half the range means more recent✓
- cWhichever packet's UDP checksum is larger is newer
- dWhichever value is closer to zero is always newer
Explanation:Plain unsigned comparison breaks once the counter wraps. The half-range rule treats the space as circular: a is newer than b only if (a-b) mod 65536 < 32768.
Nc Transport ReliabilityDifficulty 3
16-bit sequence numbers, last accepted value is 65530. Packet #5 then arrives. Applying the half-range rule, is #5 newer or older than 65530?
- aNewer, since the wrap-around distance is only 11✓
- bOlder — 5 is numerically far smaller than 65530
- cUndecidable without the packet's arrival timestamp
- dEqual — both fall in the same 16-bit buffer window
Explanation:Distance from 65530 forward to 5 is (65536-65530)+5 = 11, far under 32768, so 5 is newer — the counter simply wrapped past 65535.
Nc Transport ReliabilityDifficulty 2
A header carries an ack field plus a 32-bit ack-bits field. If bit n of ack-bits is set, what does that mean?
- aPacket number n was dropped and must be resent now
- bThe receive buffer holds n unprocessed packets
- cThe packet numbered (ack - n) has been received✓
- dThe next outgoing packet will use number (ack + n)
Explanation:ack names the newest received sequence number; each bit of ack-bits redundantly reports one older packet counting backward, so bit n means (ack - n) was received.
Nc Transport ReliabilityDifficulty 2
The same packet is reported acknowledged inside 32 consecutive headers via ack-bits, instead of one dedicated ack packet. What is the main benefit?
- aIt halves total header size versus a single explicit ack
- bThe receiver no longer needs to track what it has seen
- cIt guarantees every ack arrives even under heavy loss
- dLosing a few headers still leaves copies of the ack✓
Explanation:Repeating the same ack across many headers means a single lost header rarely matters — another redundant copy usually still carries it through, without a dedicated retransmit.
Nc Transport ReliabilityDifficulty 1
A message was sent and no ack for it has arrived yet. What is the correct way to interpret this by itself?
- aIt only means delivery is not yet confirmed, nothing more✓
- bIt proves the packet was definitely lost on the wire
- cIt proves the receiver is alive and processing packets
- dIt means the connection has already timed out
Explanation:Absence of an ack is not proof of loss — the packet, or its ack, could simply still be in flight. Loss is only inferred once a threshold is crossed.
Nc Transport ReliabilityDifficulty 2
Packets #100, #101, #102 are sent close together. #102 arrives first, then #100, then #101 — all three arrive, just out of order. If a gap in arrival order instantly means loss, what happens?
- aNothing wrong, ack-bits only reports packets that never arrived
- b#100 and #101 get flagged as lost and needlessly resent✓
- cThe connection resets since it cannot handle any reordering
- dThe sequence space silently wraps early to skip them
Explanation:A naive gap-means-loss rule cannot tell reordering from real loss: seeing #102 before #100/#101 looks like a gap even though both are simply still in flight, causing spurious retransmits.