Typically a disk drive consists of a disk pack consisting of several platters stacked like phonograph records. Information is recorded magnetically usually on both sides. All the platters move as one unit. Each platter is divided into concentric rings called tracks, and each track is divided into sectors. Data transferred to and from a sector is called a block.
The number of sectors per track and number of bytes per block is determined by how the disk is formatted. Some disk units allow the formatting to be done under software control.
Data are transferred from a platter side through a read/write head. These can move in and out to read different tracks; this motion is called seeking. The read/write heads of different disks are usually physically tied together so they all access the same track at the same time. The set of all track positions for a given head position is called a cylinder.
Three types of delay, or latency, affect how long it takes to read or write disk data. First, the seek latency is the amount of time needed to seek to the desired cylinder, and depends on how far the head has to seek. Rotational latency is the amount of time needed for the desired sector to arrive at the read/write head. .änd is about 10 milliseconds. Last, the transfer latency is the amount of time needed for the sector to be completely scanned.
As a concrete example, consider a Seagate 1.7 Gbyte disk drive. It has 512 bytes per sector, 63 sectors per track, 3305 cylinders, 2 disk platters, and 4 R/W heads, average seek time of 12.0 msec, track to track seek of 2.0 msec, spindle speed of 4500 RPM ( 13 msec per rotation).
An another example, consider the Seagate 4.3 GB drive. It has 512 bytes per sector, 63 sectors per track, 8894 cylinders, 4 disks, and 8 R/W heads, average seek of 12.0 msec, track to track seek of 2 msec, and spindle speed of 45000 RPM.
Thus, the difference seems to be in number of disks and cylinders.
One or more disk drives are connected to a disk controller,
which positions the disk arms,
controls transfer of data,
and senses any errors that occur.
A disk controller talks to the cpu through a host interface,
which is connected to the system bus.
The host interface passes requests from the cpu to the
controller to start an I/O operation,
and interrupts the CPU when the controller signals completion.
Like the SLU unit,
it is associated with device registers.
The main difference between the two is that the disk hardware can
transfer large amounts of data to and from memory,
interrupting only when the transfer is complete.