I/O Model of Computation

Synonyms

Disk-based model

Definition

The I/O model of computation measures the efficiency of an algorithm by counting how many disk reads and writes it needs. It is widely applicable to the database environment, since most data is stored on disks and disk access typically dominates CPU time.

Key Points

For many computing-intensive applications, the appropriate model of computation is to measure CPU time. Yet in data-intensive applications, such as databases, it is more relevant to measure the number of disk I/Os [1]. This is termed the “I/O model of computation,” or disk-based model. Nowadays, most hard drives use the seek-rotate-transfer protocol [2]. In order to transfer some data from disk to memory (so as the data can be processed by the CPU), or to transfer data back to disk, the hard drive needs first to spend some “seek time” to move the read/write head to the cylinder where the data is located at. Then the “rotational delay” is spent until the sector containing the data...