00:01
Hello student here the first question is why is rotational latency usually not considered in the disk scheduling? how would you modify sst, f scan and c scan to include the latency optimization? so this is the question.
00:16
So here rotational latency is usually not considered in the disk scheduling because it is difficult to predict the various dependency on the location of the data on the disk.
00:28
So however if latency optimization is desired sstf shortest job first, shortest time first, shortest scan time first, c scans and scan can be modified to take into account the expected rotational latency.
00:45
One way to modify sstf to include the latency optimization is to first give the importance request that are closer to the current rotational position of the disk head.
00:56
This can be done by calculating the expected rotational latency for each request and then scheduling the request with the lowest expected rotational latency.
01:07
Another way to modify sstf to include latency optimization to use a look ahead algorithm.
01:14
A look ahead algorithm examine the next few requests in the queue and schedule the request that will optimize the overall rotational latency.
01:23
C scan and scan can be modified to include latency optimization in a similar way.
01:31
The main difference is that c scan and scan moves the disk head in a specific direction so the expected rotational latency can be calculated more accurately.
01:41
Here are some specific examples of how to modify sstf to include the latency optimization.
01:47
Calculate the expected rotational latency for each request in the queue.
01:52
Sort the request in the queue in the ascending order of the expected rotational latency.
01:57
Schedule the request at the front of the queue.
02:00
Repeat step 2 to 3 until the queue is empty.
02:03
So this algorithm will schedule the request in a way that minimizes the overall rotational latency.
02:09
So it is important to note that optimizing for the rotational latency can lead to a longer seek time.
02:14
This is because the disk head may have to move to a different cylinders in order to schedule the request with the lowest expected latency...