edited by
11,841 views
34 votes
34 votes

Consider the following statements about the timeout value used in TCP.

  1. The timeout value is set to the RTT (Round Trip Time) measured during TCP connection establishment for the entire duration of the connection.
  2. Appropriate RTT estimation algorithm is used to set the timeout value of a TCP connection.
  3. Timeout value is set to twice the propagation delay from the sender to the receiver.

Which of the following choices hold?

  1. $(i)$ is false, but $(ii)$ and $(iii)$ are true
  2. $(i)$ and $(iii)$ are false, but $(ii)$ is true
  3. $(i)$ and $(ii)$ are false, but $(iii)$ is true
  4. $(i), (ii)$ and $(iii)$ are false
edited by

5 Answers

Best answer
37 votes
37 votes
  1. (i) TCP connection established in $3$ phase between $\text{SYN}$ send and $\text{SYN}$ received $(\text{SYN},\text{SYN}+\text{ACK},\text{ACK})$. After this connection establishment, data transfer takes place. Now, $\text{FIN}$ flag is called to close the connection. $\text{FIN}$ flag can close the connection after getting the $\text{ACK}$ from the receiver. If $\text{ACK}$ is not received, a timer is set which wait for the time out. So, there is no relationship between TCP connection establishment and Timeout of $\text{RTT}.$


     
  2. This is Jacobson's algorithm (Thanks @Anirudh)
    $\text{ERTT}=p\times \text{IRTT}+(1-p)\times \text{NRTT}$
    $p$ is scaling factor
    $\text{IRTT}$ initial $\text{RTT}$
    $\text{NRTT}$ is new $\text{RTT}$
    Link here

     
  3. Actually timeout value is more than twice the propagation delay from sender to receiver. Because after connection establishment and data transfer complete, then only timeout occurs. So, if we start timer at the beginning of transaction, Time Out occurs after RTT completes and after final $\text{ACK}$ comes. So, Time Out time must be more than $\text{RTT}.$

So, only (ii) is TRUE. Answer (B).

edited by
11 votes
11 votes
Answer: B

Jacobson's algorithm is used.
7 votes
7 votes

Statement-I: It is False.
The timeout value cannot be fixed for entire duration as it will turn timer to static timer, we need dynamic timer for timeout.
Statement-II: It is True.
Basic algorithm, Jacobson's algorithm, Karl's modification; these three algorithms are to be appropriate to RTT estimation algorithm used to set timeout value dynamically.
Statement-III: It is False.
Because timeout value is set to twice the propagation delay in data link layer where hop to hop distance is known, not in TCP layer.

Answer:

Related questions

11.0k
views
4 answers
42 votes
Ishrat Jahan asked Oct 29, 2014
11,001 views
Consider a $TCP$ connection in a state where there are no outstanding $ACK$s. The sender sends two segments back to back. The sequence numbers of the first and second segments ... $ and $290$230$ and $291$60$ and $231$60$ and $230$
2.3k
views
2 answers
0 votes
Ishrat Jahan asked Oct 30, 2014
2,275 views
Consider a token ring topology with N stations (numbered 1 to N) running token ring protocol where the stations are equally spaced. When a station gets the token it is allowed to send ... = 5 ms, tp = 3 ms, N = 15 is:0.5450.6550.93750.961
3.9k
views
2 answers
3 votes
Ishrat Jahan asked Oct 30, 2014
3,943 views
Consider a token ring topology with N stations (numbered 1 to N) running token ring protocol where the stations are equally spaced. When a station gets the token it is allowed to ... when tt =3 ms, tp = 5 ms, N = 10 is0.5450.60.8570.961
12.5k
views
3 answers
28 votes
Ishrat Jahan asked Oct 30, 2014
12,468 views
You are given the following four bytes : $\begin{array}{| c | c | c | c |} \hline 10100011 & 00110111 & 11101001 & 10101011 \\\hline \end{array}$ ... $ encoding of the above four bytes?$\text{zdp}$\text{fpq}$\text{qwA}$\text{oze}$