TCP congestion control in Computer Networks - Practice Problems & Coding Challenges

Challenge - 5 Problems

🧠 Conceptual

intermediate

Understanding TCP Congestion Window Behavior

What happens to the TCP congestion window size immediately after a packet loss is detected by triple duplicate acknowledgments?

A The congestion window size is reset to one segment and TCP enters slow start phase.

B The congestion window size is halved and TCP enters congestion avoidance phase.

C The congestion window size remains unchanged and TCP continues transmission normally.

D The congestion window size doubles immediately to recover lost throughput quickly.

💡 Hint

Think about how TCP reacts to moderate congestion signals without restarting completely.

When triple duplicate ACKs are received, TCP performs fast retransmit and fast recovery. It halves the congestion window size to reduce load and enters congestion avoidance to cautiously increase the window again.

📋 Factual

intermediate

TCP Slow Start Phase Behavior

During the TCP slow start phase, how does the congestion window size change with each round-trip time (RTT)?

A It increases exponentially, doubling each RTT.

B It remains constant until packet loss occurs.

C It increases linearly by one segment per RTT.

D It decreases exponentially to avoid congestion.

💡 Hint

Slow start aims to quickly find the network capacity by rapidly increasing the window.

In slow start, the congestion window doubles every RTT because it increases by one segment for every ACK received, effectively doubling the window each round.

🔍 Analysis

advanced

Effect of Network Delay on TCP Congestion Control

How does an increase in network round-trip time (RTT) affect the TCP congestion control performance?

A Higher RTT causes TCP to ignore congestion signals and maintain window size.

B Higher RTT causes faster congestion window growth, increasing throughput.

C RTT has no effect on congestion window size or throughput.

D Higher RTT causes slower congestion window growth, reducing throughput.

💡 Hint

Consider how RTT influences the feedback loop speed in TCP.

TCP relies on acknowledgments to adjust its congestion window. Higher RTT means acknowledgments arrive slower, so window growth slows, reducing throughput.

❓ Comparison

advanced

Comparing TCP Tahoe and TCP Reno

Which key difference distinguishes TCP Reno from TCP Tahoe in handling packet loss?

A TCP Tahoe ignores packet loss, TCP Reno reduces transmission rate.

B TCP Tahoe uses selective acknowledgments, TCP Reno does not.

C TCP Reno performs fast recovery after triple duplicate ACKs, while TCP Tahoe always enters slow start.

D TCP Reno resets the congestion window to zero on any loss, TCP Tahoe halves it.

💡 Hint

Think about how each version recovers from packet loss without restarting completely.

TCP Reno improves on Tahoe by using fast recovery to avoid slow start after triple duplicate ACKs, halving the window instead of resetting it.

❓ Reasoning

expert

Impact of Packet Loss Detection Method on TCP Throughput

Consider two TCP connections: one detects packet loss via timeout, the other via triple duplicate ACKs. Which connection typically achieves higher throughput and why?

A The connection detecting loss via triple duplicate ACKs achieves higher throughput because it recovers faster without reducing the congestion window drastically.

B The connection detecting loss via timeout achieves higher throughput because it waits longer before retransmitting, avoiding unnecessary retransmissions.

C Both connections achieve the same throughput since loss detection method does not affect performance.

D The connection detecting loss via timeout achieves higher throughput because it halves the congestion window more aggressively.

💡 Hint

Think about how quickly each method allows TCP to resume sending data after loss.

Triple duplicate ACKs allow fast retransmit and fast recovery, reducing the congestion window less and resuming transmission faster than timeout-based detection, which causes slow start.