Log In
24 votes

Which of the following disk scheduling strategies is likely to give the best throughput?

  1. Farthest cylinder next

  2. Nearest cylinder next

  3. First come first served

  4. Elevator algorithm

in Operating System
edited by
what is Farthest cylinder next algorithm

Option (B) and (D) both are acceptable. For more details please check comment of @Vicky rix ji.

As throughput is asked can it be interpreted as at any point of time during disk scheduling which Algo gives max request processing per second. Which implies minimum head movements.

@tusharp It should be minimum head movements not maximum as overall seek latency is minimum when head movements are minimum thus maximum throughput.

6 Answers

33 votes
Best answer
  1. Farthest cylinder next $\rightarrow$ This might be candidate for worst algorithm . This is false.
  2. Nearest cylinder next $\rightarrow$ This is output.
  3. First come first served $\rightarrow$ This will not give best throughput. It is random .
  4. Elevator algorithm $\rightarrow$ This is good but issue is that once direction is fixed we don't come back, until we go all the other way. So it does not give best throughput.

Correct Answer: $B$

edited by
I am confused between nearest cylinder next or elevator?? Because both gives good throughput. what should be the answer?? and why?? why it shouldnt be elevator since it doesnt even have starvation problem??
How to decide between nearest cylinder and elevator algo.
Option B is like SSTF ...

Hi @sushmitavaishali jhalani and Puja Mishra ji,

First of all thanks for raising good point. But could you please provide some counter example where Elevator algorithm will perform better then Nearest cylinder next ? 

In nearest cylinder next, we cant fix the direction in which go next

but in Elevator , first we fix the diection then go only that direction and then come back to opposite direction

like if we have 10,50 request and head is on 20

then nearest cylinder next , 20->10->50   total seek time 10+40=50

elevator 20->50->10  total seek time=30+40=70 ( direction is fix toward 50)

As long as the block request string is given, we cannot decide between B) and D)

B) is not always the best EX:


throughput put depends on the number of requests served per second  as we always go for the shortest reach SSTF Does this best as it tries to serve the maximum number of processes in less time 



Thank you @Vicky rix ji.

Scan is called elevator algorithm ...

For more reference go to this question.

How SCAN Algo hear move one side head and then return.

It should be Elevator algorithm since it provides more uniform waiting time than the SSTF or nearest cylinder next algorithm.

Both B AND D are correct. 

D is also correct(see @Vicky rix comment ) and as it's a GATE 1999 question multiple options are possible as answers. 


I dont think there can be 2 correct answers for this as it is asking for the BEST, how can two strategies be best simultaneously. What to write if it appears in the exam now…. Other sites like GFG have also given B

11 votes
Option B

Nearest cylinder next but it can lead to starvation
9 votes
Nearest cylinder next . in this many pages can be accessed in less time  . So throughput will be high .
3 votes

The only catch in the question is that they are asking for the most likely to give the best throughput.

Even if for some cases SCAN can have better results but the most likely to result the best throughput is Nearest cylinder Algo ;)

1 vote
A- Like always choose more distance from itself , so it can't be answer

B- its SSTF

C- its FCFS


> FCFS throughput is less then SSTF so it's also not answer

> now the fight b//w SSTF and SCAN

SSTF always serve nearest cylinder first , so necessary movements gets reduced . and in case of SCAN it goes to one end and come back ,which takes more time to complete task , so SSTF throughput is best .
–4 votes

Both the Nearest Cylinder and Elevator are high on throughput.But I will go with ELEVATOR.


Nearest Cylinder Next algo will have a overhead of computing the nearest cylinders of the queue.

Related questions

26 votes
3 answers
A multi-user, multi-processing operating system cannot be implemented on hardware that does not support Address translation DMA for disk transfer At least two modes of CPU execution (privileged and non-privileged) Demand paging
asked Sep 23, 2014 in Operating System Kathleen 6.9k views
19 votes
2 answers
System calls are usually invoked by using a software interrupt polling an indirect jump a privileged instruction
asked Sep 23, 2014 in Operating System Kathleen 7.2k views
4 votes
4 answers
Consider the following solution to the producer-consumer problem using a buffer of size 1. Assume that the initial value of count is 0. Also assume that the testing of count and assignment to count are atomic operations. Producer: Repeat Produce an item; if count = ... Producer); Consume item; Forever; Show that in this solution it is possible that both the processes are sleeping at the same time.
asked Feb 28, 2018 in Operating System jothee 1.2k views
19 votes
6 answers
A certain processor provides a 'test and set' instruction that is used as follows: TSET register, flag This instruction atomically copies flag to register and sets flag to $1$. Give pseudo-code for implementing the entry and exit code to a critical region using this instruction.
asked Sep 24, 2014 in Operating System Kathleen 2.1k views