Sequential Versus Random Disk Access Cost
Reading or writing a contiguous run of bytes on a storage device is far cheaper per byte than jumping between scattered locations, because seek and per-operation overhead dominate small random accesses. This asymmetry, not capacity, is what makes storage designs prefer to write in one direction and to batch. After this Concept a student can predict which of two access patterns over the same data volume will be faster on disk, and can justify a design that trades extra bytes written for fewer random seeks.
Sequential Versus Random Access Cost on Hard Disks
Hard disk data access incurs distinct cost models depending on access pattern: random access requires a seek time (moving the disk arm to the correct track) plus a rotational delay (waiting for the t…