Skip to content

GATE CS 2021 Set 1 Q45 — Transport Layer: Flow Control, Congestion Control, UDP, TCP, Sockets

Explain with AI

Consider two hosts PP and QQ connected through a router RR. The maximum transfer unit (MTU)\text{(MTU)} value of the link between PP and RR is 15001500 bytes, and between RR and QQ is 820820 bytes.

A TCP\text{TCP} segment of size 14001400 bytes was transferred from PP to QQ through RR, with IP\text{IP} identification value as 0x1234.\textsf{0x1234}. Assume that the IP\text{IP} header size is 2020 bytes. Further, the packet is allowed to be fragmented, i.e., Don’t Fragment (DF)\textsf{Don’t Fragment (DF)} flag in the IP\text{IP} header is not\textit{not} set by PP.

Which of the following statements is/are correct?

Select all correct options
View worked solution

Answer: A;C

Let's analyze each statement step-by-step:

  • Statement A (Correct): The original IP datagram sent from PP has a total size of 1400 (TCP payload)+20 (IP header)=14201400 \text{ (TCP payload)} + 20 \text{ (IP header)} = 1420 bytes. At router RR, the outgoing link MTU is 820820 bytes. The maximum IP payload size per fragment is MTU−IP header=820−20=800\text{MTU} - \text{IP header} = 820 - 20 = 800 bytes (which is a multiple of 88).
    • Fragment 1: Payload = 800800 bytes, Total size = 820820 bytes.
    • Fragment 2: Remaining payload = 1400−800=6001400 - 800 = 600 bytes, Total size = 600+20=620600 + 20 = 620 bytes.
    Thus, two fragments are created, and the second fragment's size is indeed 620620 bytes.
  • Statement B (Incorrect): Routers operate at the Network Layer and do not perform retransmissions of lost fragments. Retransmission is handled end-to-end by the transport layer (TCP) at the source host PP.
  • Statement C (Correct): IP is an unreliable protocol. If any fragment of an IP datagram is lost, the destination host QQ cannot reassemble the datagram and discards all other fragments. Since TCP is unaware of IP-level fragmentation, the source PP must retransmit the entire TCP segment.
  • Statement D (Incorrect): The TCP header (which contains the destination port) is located at the very beginning of the TCP segment. Therefore, it is entirely contained within the first fragment. The second fragment contains only the remaining segment payload and cannot be used to determine the destination port.

Free practice · Your answers on this page are not saved.

Published
Report a correction
Keep exploringMore Transport Layer: Flow Control, Congestion Control, UDP, TCP, Sockets questions