Moving the head — the seek — is the slow part of a spinning disk, so the order you service requests matters. FCFS, SSTF, and the elevator-style SCAN/LOOK trade head travel against fairness.
On a spinning disk, reaching data means moving the read/write head to the right cylinder, and that travel — the seek time — dominates everything else. So when several requests are pending, the scheduling policy reorders them to cut total head movement. Less travel means faster service.
The main policies
1FCFS (First Come, First Served) — service requests in arrival order. Fair and simple, but the head often zig-zags across the disk.
2SSTF (Shortest Seek Time First) — always jump to the nearest pending request. Much less travel, but far-away requests can be starved.
3SCAN (the elevator) — sweep in one direction servicing everything, then reverse at the disk's edge and sweep back.
4LOOK — like SCAN, but reverse at the last request in a direction instead of riding all the way to the physical edge.
Worked example
The head starts at cylinder 53 with this request queue: 98, 183, 37, 122, 14, 124, 65, 67. Compare total head movement:
Head = 53, queue = 98 183 37 122 14 124 65 67
FCFS: 53->98->183->37->122->14->124->65->67
total head movement = 640 cylinders
SSTF: 53->65->67->37->14->98->122->124->183
total head movement = 236 cylinders
Why SSTF wins here
FCFS bounces from 183 all the way back to 37 and out to 124 — huge wasted travel. SSTF always grabs the closest request, so 53 goes to 65 then 67 (just 12 + 2 cylinders) before working outward, cutting total travel from 640 to 236.
SSTF can starve
Because SSTF always prefers the nearest request, a request far from the current head region can be postponed indefinitely if nearer requests keep arriving. SCAN/C-SCAN fix this: the steady sweep guarantees every cylinder is reached within one pass.
OperationTimeSpace
FCFS · fair, but most head travelO(n) orderO(n)
SSTF · less travel; can starve far requestsO(n) per pickO(n)
