Minimizing Seek Time Across Competing Read Requests
A mechanical disk's read/write head physically moves across platters to reach requested data, and that movement (seek time) is by far the slowest part of a disk operation — disk scheduling algorithms exist specifically to minimize total head movement across a queue of pending requests, rather than serving them in arbitrary arrival order.
FCFS serves requests in arrival order — simple, but can produce a lot of unnecessary back-and-forth head movement if requests arrive in a scattered order. SSTF (Shortest Seek Time First) always services whichever pending request is physically closest to the head's current position, minimizing immediate seek distance but risking starvation of far-away requests if closer ones keep arriving. SCAN moves the head steadily in one direction, servicing requests as it passes them, then reverses at the end — like an elevator, avoiding SSTF's starvation risk while still keeping movement efficient. C-SCAN (Circular SCAN) only services requests in one direction, then jumps back to the start without servicing on the return trip, giving more uniform wait times across the disk. LOOK is SCAN's more practical refinement — reversing direction as soon as no further requests remain ahead, rather than always sweeping all the way to the physical end of the disk. C-LOOK does the same for C-SCAN: it serves requests in one direction, then jumps straight to the lowest pending request instead of the disk's edge.
Comparing total head movement (the sum of absolute differences between consecutive serviced cylinders) across all six algorithms on an identical request queue is exactly how operating systems courses demonstrate the real efficiency gap between naive and elevator-style scheduling.
Worked example
Queue 98, 183, 37, 122, 14, 124, 65, 67 with the head at 53 on a 200-cylinder disk (the textbook example from Silberschatz/Galvin):
- FCFS: 640 cylinders. |98 − 53| + |183 − 98| + |37 − 183| + … = 45 + 85 + 146 + 85 + 108 + 110 + 59 + 2 = 640
- SSTF: 236 cylinders, serving 65, 67, 37, 14, 98, 122, 124, 183
- SCAN toward 0: 236 cylinders: down to 0, then up to 183
The working table prints the |to − from| line for every move, so you can copy the whole calculation, not just the answer.
Does the C-SCAN return jump count?
This is the most common reason an answer "doesn't match". When C-SCAN reaches the end of the disk it jumps back to the other end without serving anything. Galvin and GeeksforGeeks count that jump in the total head movement; some courses treat it as free because the arm returns at high speed. The checkbox switches between the two, and the jump is marked in the working and drawn dashed on the seek graph. For the queue 176, 79, 34, 60, 92, 11, 41, 114 with the head at 50, C-SCAN is 389 cylinders with the jump counted and 190 without.
Where this fits
Disk scheduling orders requests for the disk the way CPU scheduling orders processes for the processor, and many of the same ideas, FCFS fairness against shortest-first efficiency and the risk of starvation, appear in both. Memory, the level above the disk, is covered by the page replacement and memory allocation calculators.