RET: Home

    The heads of a hard-disk drive fly on a cushion of air only a few-millionths-of-an-inch above the surface of a magnetic platter spinning at about 100 miles per hour. Compounding the problem, these heads must accurately transverse disk tracks only 300-millionths-of-an-inch apart. No matter how well designed or technologically advanced, hard drives are ultimately subject to the laws of physics, - it takes time to move and position a physical object.

Consider these key specifications when evaluating a drive's true performance characteristics:

Seek time specifies how long it takes to move the read/write head between tracks. While it might take only a couple of milliseconds to jump from one track to the next, it could take 10 times as long to move between the outer and inner track.

Platters are individual disks. Platters are stacked together to create the total capacity of the hard drive. One head is dedicated to each side of a platter. The head switch time specifies the time it takes to switch between two read/write heads, which is more efficient than track switching.

Cylinders are the identically positioned tracks on different platters. Cylinder switch time is similar to seek time, but specifies the time it takes for the heads to jump from cylinder to cylinder.

Rotational Latency is the time required for a platter to rotate to the required sector. Rotational latency is largely determined by the drive's rotational speed (rpm). Simply put, the faster the disk spins, the less time is spent waiting.

Data Access time specifies how long it takes a head to find a particular track and sector (or sectors) within the track. Access time is the sum of seek times, head switch time and rotational latency.

Read Caching allows data to be continually read until the disk buffer is full. Data is read to the buffer even after the requested data has been located. This pre-fetching (i.e. reading extra data) stores data in the disk buffer (i.e. cache) that might be required by the next read request. Since data is retrieved from cache much quicker than from disk, data response times can be greatly improved.

Write Caching buffers data to the drive in an attempt to keep the write transfer rate as fast as possible.

Data Transfer Rate, measured in megabytes per second (MBPS), is the speed at which data moves from disk to controller and then the speed at which the controller moves data to the CPU. Data throughput rate reflects both access time and transfer rate.

Self-Monitoring, Analysis and Reporting Technology (SMART) helps predict drive reliability by monitoring head flying height, data throughput performance, spin-up time and other key drive statistics.