The IIT Madras BS Computer Networks (Computer Networks) End Term paper sat on 21 Dec 2025, in the September 2025 term: 42 questions for 100 marks in 180 minutes. Every question is below with its answer. Take it as a timed mock test to be marked, or read it through first.
Assume a host in Mumbai wants to send a packet to a server in Delhi, approximately 1300 km away. If the signal travels at 2 × 109 m/s, what is the one-way propagation delay in milliseconds? NOTE: Assume no processing or queuing delay(in ms).
Correct answer: 0.65 (accepted within ±0.05)
Assume a host in Mumbai wants to send a packet to a server in Delhi, approximately 1510 km away. If the signal travels at 2 × 106 m/s, what is the one-way propagation delay in milliseconds? NOTE: Assume no processing or queuing delay(in ms).
Correct answer: 755 (accepted within ±5)
If the one-way latency between two cities is measured to be 35 ms, the RTT (Round Trip Time) assuming symmetric paths and no processing delay is:
Correct answer: 70 (accepted within ±5)
If the one-way latency between two cities is measured to be 15 ms, the RTT (Round Trip Time) assuming symmetric paths and no processing delay is:
Correct answer: 30 (accepted within ±2)
A transmitter sends packets of size 2100 bits at a rate of 4 Mbps from node X to node Y. The one-way delay is observed to vary between 25 ms (no queueing) and 100 ms (maximum queue), with an average of 50 ms. Assume queuing occurs only at one router along the path and there is no packet loss. What is the mean number of packets in the queue at the bottleneck link?
Correct answer: 47.5 (accepted within ±2.5)
A transmitter sends packets of size 2000 bits at a rate of 3 Mbps from node X to node Y. The one-way delay is observed to vary between 20 ms (no queueing) and 120 ms (maximum queue), with an average of 60 ms. Assume queuing occurs only at one router along the path and there is no packet loss. What is the mean number of packets in the queue at the bottleneck link?
Correct answer: 60 (accepted within ±5)
In a circuit-switched network, a source communicates with a destination through 10 intermediate switches. Each switch takes 4 ms to process a reservation request before forwarding it to the next node. The propagation delay between each pair of nodes (i.e., source/switch/destination) is 3 ms. What is the total time (in milliseconds) taken for the reservation request to reach the destination?
Correct answer: 72.5 (accepted within ±2.5)
In a circuit-switched network, a source communicates with a destination through 5 intermediate switches. Each switch takes 3 ms to process a reservation request before forwarding it to the next node. The propagation delay between each pair of nodes (i.e., source/switch/destination) is 2 ms. What is the total time (in milliseconds) taken for the reservation request to reach the destination?
Correct answer: 27.5 (accepted within ±1.5)
Answers Case Sensitive : No
Correct answer: 0x2223
Answers Case Sensitive : No
Correct answer: 0x2222
A TCP connection uses the following identifiers:
• Source IP = 192.168.1.10 • Destination IP = 10.0.0.5 • Source Port = 55000 • Destination Port = 80
Assuming that TCP port numbers range from 0 to 65535, but 0–1023 are the well-known (privileged) ports and are excluded, how many additional TCP connections can this host theoretically open to the same server (IP 10.0.0.5, port 80) using different source ports?
Correct answer: 64511
A TCP connection uses the following identifiers:
• Source IP = 192.168.1.10 • Destination IP = 10.0.0.5 • Source Port = 55000 • Destination Port = 80
Assuming that TCP port numbers range from 0 to 65535, but 0–1023 are the well-known (privileged) ports and are excluded, how many additional TCP connections can this host theoretically open to the same server (IP 10.0.0.5, port 80) using different source ports?
Correct answer: 64511
A 20 Mbps link has a propagation delay of 100 μs. The transmitter employs the Stop-and-Wait protocol. Assuming each frame is 150 bytes long, what is the maximum data rate possible (in Mbps)? Note: In two point precision
Correct answer: 4.65 (accepted within ±0.05)
A 20 Mbps link has a propagation delay of 400 μs. The transmitter employs the Stop-and-Wait protocol. Assuming each frame is 100 bytes long, what is the maximum data rate possible?
Correct answer: 0.95 (accepted within ±0.05)
If you are designing sliding window protocol (assume selective repeat) of 5 Mbps which has one way delay of 1.27 seconds. Assuming each frame carries 2kB of data, what is the minimum number of bits you need for the sequence number?
Correct answer: 11
If you are designing sliding window protocol (assume selective repeat) of 7 Mbps which has one way delay of 1.55 seconds. Assuming each frame carries 1kB of data, what is the minimum number of bits you need for the sequence number?
Correct answer: 13
In a network with a bottleneck bandwidth of 30 Mbps and a round-trip time of 50 ms,what should be the minimum congestion window size (in bytes) to fully utilize the link?
Correct answer: 187500
In a network with a bottleneck bandwidth of 25 Mbps and a round-trip time of 400 ms,what should be the minimum congestion window size (in bytes) to fully utilize the link?
Correct answer: 1250000
A TCP flow has a congestion window of 86 KB and an RTT of 450 ms. Calculate the average throughput in Mbps. Note: In two point precision
Correct answer: 1.55 (accepted within ±0.05)
A TCP flow has a congestion window of 54 KB and an RTT of 510 ms. Calculate the average throughput in Mbps. Note: In two point precision
Correct answer: 0.875 (accepted within ±0.025)
In TCP BBR’s bottleneck bandwidth estimation, the ACK arrival rate measured over an epoch gives a delivery rate of 84 Mbps. If the minimum measured RTT (RTTmin) is 200 ms, what is BBR’s current estimate of the bottleneck bandwidth (in Mbps)?
Correct answer: 84
In TCP BBR’s bottleneck bandwidth estimation, the ACK arrival rate measured over an epoch gives a delivery rate of 48 Mbps. If the minimum measured RTT (RTTmin) is 100 ms, what is BBR’s current estimate of the bottleneck bandwidth (in Mbps)?
Correct answer: 48
A TCP Reno connection detects congestion when CWND = 200 KB. After triple duplicate ACK, CWND is reduced to 1/4, and additive increase resumes with +3 KB per RTT. Compute the CWND after 5 RTTs post-loss.(in KB)
Correct answer: 65
A TCP Reno connection detects congestion when CWND = 100 KB. After triple duplicate ACK, CWND is reduced to 3/4, and additive increase resumes with +1 KB per RTT. Compute the CWND after 3 RTTs post-loss.(in KB)
Correct answer: 78
An ISP manages a network block of 200.8.0.0/12. it partitions and assigns a customer a smaller block 200.8.4.128/18. Calculate the number of network bits that were added to the prefix length during this delegation process and then find how many total subnets could the ISP theoretically create from the original /12 block if it subdivides it all the way down to /18 subnets?
Correct answer: 64
An ISP manages a network block of 200.8.0.0/10. it partitions and assigns a customer a smaller block 200.8.4.128/22. Calculate the number of network bits that were added to the prefix length during this delegation process and then find how many total subnets could the ISP theoretically create from the original /10 block if it subdivides it all the way down to /22 subnets?
Correct answer: 4096
How many distinct prefixes can a /12 block be divided into when it is subnetted up to /16 networks?
Correct answer: 16
How many distinct prefixes can a /10 block be divided into when it is subnetted up to /18 networks?
Correct answer: 256
An organization has 1900 devices connected in a private network. Each NAT router can handle up to 90,000 simultaneous connections and can serve 200 devices comfortably. If they use NAT (one public IP per NAT router), how many public IP addresses are required?
Correct answer: 10
An organization has 2,200 devices connected in a private network. Each NAT router can handle up to 80,000 simultaneous connections and can serve 290 devices comfortably. If they use NAT (one public IP per NAT router), how many public IP addresses are required?
Correct answer: 8
When using Dijkstra’s algorithm to find shortest path in a graph, which of the following statement is not true?
It can find shortest path within the same graph data structure
Every time a new node is visited, we choose the node with smallest known distance/cost (weight) to visit first
Shortest path always passes through least number of vertices
The graph needs to have a non-negative weight on every edge
Correct answer
Shortest path always passes through least number of vertices
When using Dijkstra’s algorithm to find shortest path in a graph, which of the following statement is not true?
It can find shortest path within the same graph data structure
Every time a new node is visited, we choose the node with smallest know distance/cost (weight) to visit first
Shortest path always passes through least number of vertices
The graph needs to have a non-negative weight on every edge
Correct answer
Shortest path always passes through least number of vertices
The browser connects (TCP handshake visible) but the webpage never loads. What is the MOST likely issue?
Correct answer
The browser connects (TCP handshake visible) but the webpage never loads. What is the MOST likely issue?
Correct answer
The server prints the GET request in console but the browser keeps “Waiting for localhost...”.Why?
The server didn’t call accept()
The server didn’t send file data back using send()
Browser sent POST instead of GET
TCP packet was dropped
Correct answer
The server didn’t send file data back using send()
The server prints the GET request in console but the browser keeps “Waiting for localhost...”. Why?
The server didn’t call accept()
The server didn’t send file data back using send()
Browser sent POST instead of GET
TCP packet was dropped
Correct answer
The server didn’t send file data back using send()
Correct answer
Correct answer
8 bytes
16 bytes
20 bytes
32 bytes
Correct answer
32 bytes
8 bytes
16 bytes
20 bytes
32 bytes
Correct answer
32 bytes
The browser expects an HTTP response, but the server closes the connection without sending one.
The buffer size is too large, causing a deadlock.
read() cannot be used with HTTP traffic; recv() must be used.
The server must send a TCP FIN before printf.
Correct answer
The browser expects an HTTP response, but the server closes the connection without sending one.
The browser expects an HTTP response, but the server closes the connection without sending one.
The buffer size is too large, causing a deadlock.
read() cannot be used with HTTP traffic; recv() must be used.
The server must send a TCP FIN before printf.
Correct answer
The browser expects an HTTP response, but the server closes the connection without sending one.
