First in first out:
First in, first out, simple but not efficient
Shortest service priority allocation:
The IO at the location with the least mechanical movement is given priority, which can get the shortest seek time.
There is a problem with the above scheduling: if this request frequently appears near its current access location, and another request is far away from the current location, it will appear, and the magnetic head will move back and forth at the current location for a short distance, but it is far away from him. requests will continue to be unserved, leading to so-called starvation.
SCAN scheduling:
It has a constraint that the head moves in one direction and then moves back to the head.
c-scan scheduling:
Each time, ensure the sequence from low to high, from the beginning to the end, and service when encountering IO in the middle, and then quickly return to the start after the end and move to the end again, looping.
c-look scheduling: is the optimization of c-can
It does not reach the end point, but arrives at the last request. The so-called last request refers to the request closest to the end point. After processing, there is no new request point to go up. Stop here, and then Going up is actually meaningless, and it reverses immediately when it reaches the last request point.
Other ways:
