2.2 MAGNETIC TAPE
One of the first secondary storage devices invented and still one of the most
popular in use is magnetic tape. While magnetic tape is made in many sizes and
used in a variety of devices, we shall focus our attention on the half-inch tape
that is commonly used in most standard tape drives associated with computer
systems. This form of tape is stored on reels up to 10.5(10 1/2) inches in diameter
which contain lengths up to 3600 geet. Tape is easily the least expensive
medium of data storage commonly used. It is also a strictly sequential access
device, that is, a device that allows records to be read or written only in the
physical order in which they are stored.
The physical medium used to record the data consists of a thin, plastic
tape, coated on one side by a thin layer of magnetic material. The magnetic
meterial, when exposed to a strong magnetic field, becomes magnetized in the
direction of the external field. It retains this magnetism until another external
field is used to change it. Data is recorded on the tape by passing the tape over
a head which contains electrical and magnetic circuits caoable of selectively
magnetic patterns into electrical impulses.
A small length of tape is represented in Figure 2.2. The head records data
on nine parallel tracks. Each track represents a series of binary digits or bits.
Standard tape densities are 800, 1600, and 6250 bits per inch(bpi). The density
used by a particular tape drive is determined by the electrinic circuits used, the
method of coding bits into magnetic patterns, and especially the design of the
head. Some drives are capable of operating at more than one density, although
the same density is always used throughout the length of any one tape.
-------------------------------------------------------------------------------------
Figure 2.2 Length of magnetic tape with nine tracks.
-------------------------------------------------------------------------------------
Data is usually communicated between the computer and the tape drive in
units of 8-bit bytes. A byte many represent one alphanumeric character, or may
be part of a larger unit of data, such as a 64-bit floating-point number. Tape
drives recird a byte on tape by writing the 8 bits, one on each of eight different
tracks, all in one vertical column, as seen in Figure 2.2. Thus the abbreviation
bpi can mean both bits per inch (per track) or bytes per inch (for the whole
tape).
The ninth track is used to record parity bits. The parity bit for one byte is
set to a one or zero so as to make the total number of 1-bits in that column odd,
This bit is used to check for errors when reading back the data. Parity bytes are
commonly added as well to provide longitudinal parity checks.
The tape must move over the head at a precisely controlled speed when-
ever the tape is being read or written. If the tape were to move constantly, as
for example an audio tape moves, the computer would have to be able to trans-
mit or receive data continuously. Since this is seldom feasible, or even
desirable, a method is needed to start and stop the tape. However, since the tape
cannot be read or written unless it is running at a fixed speed, there will be gaps
without data where the tape stopped and restarted during recording. These gaps
can also be used to stop and start the tape during reading. These gaps are called
interrecord gaps and are illustrated in Figure 2.3. The minimum gap size for
standard half inch tapes is 0.5 inch. Most drives write a slightly longer gap with
a nominal size of about 0.6 inch. However, any drive that is capable of reading
standard tapes must be able to start and stop in half inch or less. This requires
a very high acceleration. One of the more significant cost factors in the
manufacture of a tape drive is the mechanism for accelerating the tape to full
speed in 0.25 inch.
Some tape drives are designed to write interrecord gaps without stopping.
these are called streaming tape drives and used in special applications such
as in backing up disks. It is still necessary to insert the gaps since, when the
tape is read it may not be possible to process the data as quickly as it can be
read. Streaming tape drives typically have slower start/stop times and are not
able to stop or start in the space of the interrecord gap. Such drives will over-
shoot the gap and must then back up to a point before the gap to make a "flying
start" on the next block. This process requires considerably more time than the
start/stop times of a nonstreaming tape drive
-----------------------------------------------------------------------------------
Figure 2.3 Blocks and interrecord gaps.
-----------------------------------------------------------------------------------
The unit of data between interrecord gaps is called a block. Typically, one
block will contain a number of data records. The number of records per block is
called the blocking factor. Blocks are limited in size since t