Secondary Storage

Secondary Storage

computer’s secondary storage includes devices used to store and retrieve system software, application software, and data on magnetic or optical media, such as magnetic tape, hard or floppy disks, and CD-ROMs.

Magnetic Tape

Most modern magnetic tape systems use reels that are similar to a VCR tape. A tape drive is used to re- cord bits of data onto magnetic tape by winding the tape from one reel to the other and passing it across a read/write head. The tape drive reads and writes blocks of data at a time. Each block is separated by an interblock gap, which instructs the tape drive to stop reading or writing the data until another block is requested. Figure 2-34 shows records blocked together on a magnetic tape.

A byte of data (representing a character, digit, or special symbol) is recorded on the tape across its width. (One byte equals eight binary digits, or bits. A bit is the smallest possible unit of electronic information, with a value of either 0 or 1.) A logical sequence of characters makes a field, and several fields make a record.

Although seldom used for data processing these days, magnetic tape still offers some important advantages as a secondary storage medium. For example, large amounts of data can be stored on magnetic tape at a relatively low cost, and magnetic tape is reusable. The primary disadvantage is that tapes record data sequentially, making data retrieval slower than direct access storage media. Modern tape systems alleviate this problem by using a form of indexing, in which a separate lookup table provides the physical tape location for a given data block or by marking blocks with a tape mark that can be detected while winding the tape at high speed.

Historically, tape has offered cost advantages over disk storage to make it a viable solution for data backup. Rapid improvement in disk storage density, however, combined with sluggish innovation in tape storage technologies, is eroding the market share of tape storage devices.

Magnetic Disks

The data stored on magnetic disks (hard disks or floppy disks) are considered nonvolatile. The data will reside in a certain location on the magnetic surface until they are replaced with different data or erased. Data can be recorded to magnetic disks using either of the access methods described earlier.

To get the disk ready to receive data, its surface must be formatted. An operating system utility pro- gram formats the disk by dividing it into circular tracks and wedge-shaped sectors, which cut across the tracks. The number of bytes that can be stored at a particular track and sector determines the disk’s density.

Disks are known as direct access storage devices because a piece of data can be accessed directly on the disk. Database management systems and application software work with the operating system to determine the location of the required data.

A disk has a rotating magnetic surface and a read/write head. The read/write head is on an access arm that moves back and forth over the magnetic surface. The time that elapses from the request made of

the operating system for a piece of data to when it is read into the computer is called access time. The access time of a particular hard disk is a function of several factors: (1) the seek time—how fast the read/ write head moves into position over a particular track, (2) the switching time—the time needed to activate the read/write head, (3) the rotational delay time—the time it takes to rotate the disk area under the read/ write head, and (4) the data transfer time—the time it takes for the data to be transferred from the disk track to primary storage. Most microcomputer hard disks have an access time of 5 to 60 milliseconds.

The file allocation table is an area on the disk that keeps track of the name of each file, the number of bytes in the file, the date and time it was created, the type of file, and its location (address) on the disk. A file may be stored in only one place on the disk, or it may be spread across several locations. In the latter case, the read/write head of the disk must skip among various addresses to read the entire file into the primary memory.

We have used the term address several times to represent a disk storage location. Let’s now examine the elements of a disk address.

DISK ADDRESS

As we have seen, the surface of a disk is divided into magnetized tracks that form concentric circles of data. The floppy disk for a microcomputer may have 40 or 80 tracks on a surface, while the surface of a mainframe disk could contain several hundred tracks. These tracks are logically divided into smaller blocks or record locations where data records reside. Each location is unique and has an address—a numeric value. Depending on the disk’s size and density, hundreds or thousands of records may be stored on a single track. Figure 2-35 shows data storage on a disk. For illustration purposes, the physical size of the records is greatly exaggerated.

The concept of an address applies to all types of magnetic disks, including individual floppy disks and hard disks used in microcomputers and the larger mainframe disk packs. A difference lies in the way the disks are physically arranged. Mainframe disks are often stacked on top of one another in a disk-pack arrangement that resembles a stack of phonograph records. Figure 2-36 illustrates this technique. The disks are mounted to a central spindle that rotates at over 3,500 revolutions per minute. Each disk surface is provided with a separate read/write head that is used for storing and retrieving data.

DATA STORAGE ON A DISK PACK

Every disk in a disk pack has two surfaces with the same number of tracks on each surface. Figure 2-36 shows that Track 100 exists on the top and bottom surfaces of each disk in the disk pack. Therefore, this

disk pack containing 11 disks has 22 occurrences of Track 100. To protect the data from exposure to damage, the very top and bottom surfaces of the disk pack are not used, yielding 20 data storage surfaces for Track 100.

When viewed collectively, the same track on each surface in the disk pack is called a cylinder. There- fore, in our example, Cylinder 100 contains 20 tracks of data. However, the cylinders on a microcomputer’s floppy disk or hard disk contain only two tracks because these disks have only two surfaces.

LOCATING A RECORD BASED ON ITS ADDRESS

A disk address consists of three components: the cylinder number, the surface number, and the record (or block) number. To find a record, the system must know the numeric value for each of these components. For example, if a record’s address is Cylinder 105, Surface 15, and Record Block 157, the record in question could be directly accessed as follows: First, the disk-pack control device moves the read/write heads into position above Track 105 on each surface (Cylinder 105). Next, it activates the read/write head for Sur- face 15. Finally, as Record Block 157 passes under the active read/write head, it is either read or written.

The key task in direct access storage and retrieval is ascertaining the record’s address. This may be determined from tables or calculations based on its primary key. Several direct access techniques are examined in Chapter 9.

Optical Disks

Optical disks are growing in popularity. The advantage of optical disks is that they can store very large amounts of data. A compact disc, one type of optical disk, is as portable as a floppy disk but can store more than 600 MB of data. There are several types of optical disk storage systems, including CD-ROM, WORM, and erasable optical disks.

A CD-ROM (compact disc read-only memory) is a secondary storage device that contains data or programs imprinted by the manufacturer. However, the user cannot write to (alter) the data on the CD because it is a read-only device. The WORM (write-once, read-many) disk is a secondary storage device that allows the user to write to the disk one time. An erasable optical disk allows the user to store and modify data on the disk many times.