Megnetic tape units:- The tape disk or tape drive has a tape supply reel and a tape take-up reel which works in unison to move a tape past a read/write head. It is constructed to keep the tape tight and to prevent tape extension or distortion even when the spools are started or stopped abruptly.
STORAGE STRUCTURE:-
1. Blocks:- The magnetic tape moves only when a record of data is to be read. During the stopping and restarting between records, a length of tape is by-passed and cannot be read. For this purpose, data on magnetic tape is organised in blocks, separated by interblock-gap (IRG) of around half inch lenths. Start/stop actions with the necessary times lost in reaching stable speed must be accommodated within the gaps.
2. Inter Record Gap (IRG):- Since processing takes place one record at a time, for long records an inter-record-gap is inserted after the last character in a record is written. For short records, the tape space and read/write time is wasted on IRG's. The software therefore groups several logical records into a single block of fixed or variable length and transfer itas a unit. Buffer space in main memory is used to store a block of records.
3. Labels:- Reflective markers are attached at convenient points and indicate the load point or start or end-of-reel point, thus defining the usbale area. Depending on the size of a program on data file, there may be just one file on a taoe or one file spread over several tapes.
Files are delimited by control lables, a header label identifying the file at the start and a trailor indicating the end, also storing a count of the number of blocks in a file. A special one character record stores the end-of-file marker at the end of data records. Between the start-of-tape marker and the first file record is a tape header labels, identifying the tape and its contents and supplying information on parameters and software to be used while running it. A special "scratched" tape mark in this area, indicates to the software that the tape may be erased and reused.
MAGNETIC DISK:-
Description:- Magnetic disks are formed by coating their round surfaces with magnetisable material. The thin sheet may be made of different kind of materials generally either rigid metals or flexible synthetic material similar to the one user for tapes. Both sides of the disk may be used for independent recording depending on the types of material and the structure. Disks may be portable or permanently mounted within the drive mechanism. A disk drive is required to rotate the disk with a read/write head to retrieve or record data. The disk is mounted on a vertical shaft connected to a motor which can spin it at a fixed speed.
Operation:- on a disk drive the head is fixed to an arm which can move in two dimensions, either radically or circularly. The heads contain magnetising coils which during a write or erase operation polarise or depolarise the media on the disks. During the read operation, the head senses the direction of polarisation on the disk. When the disk read/write operation is in progress the heads moves very close to the disk within the distance of a few microns. Disk surfaces must remain very clean because a small speak of dust or a finger print mark may be too large to be tolerated.
Disk packs:- Disk packs may be used singly on large computers. Rigid disks are stacked together in "disk packs". A typical pack may have six disks in it which have ten recording surfaces or the topmost and bottom most have six disks in it which have ten recording surfaces or the topmost and bottom most surfaces are susceptible to scratching and cannot be used. Larger pack could have many more disks while the smallest just have two. Disks in the packs are separated by spaces in which read/write head are positioned. A set of read/write heads on multiple arms move over the surfaces of the spining disks to access concentric tracks of data. Each head is very close to a disk surface but doesnot actually touch it
Two read/write heads are provided on each arm in the head assembly one for each of the recording surfaces on either side. All arms move together to position the heads corresponding areas on the surfaces, however only one head is "enabled" means selected by the circuitry at a given time so that data is passing only through the selected one. Some disk packs are permanently connected to computer called as fixed disk packs. Others can be removed and replaced as required and are exchangeable disk packs.
STORAGE STRUCTURE:-
1.Tracks:- The surface of the disk is divited into concentric circles known as tracks. The number of these tracks will vary accordingly to the diameter of the disk and the quality of the disk drive and can range from 40 to over 2000. Recording densities of over 400 tracks per inch have been achieved while disk diameters vary from 3.5 to 4 inches. The number of bits stored on a track depends upon the recording frequency used and is measured in terms of bit per inch of track or bpi.
2. Sectors:- Each track is identified by a number and is divided either by hardware or software into fixed size areas called as sectors or blocks. The unit of transfer of data at a time is in terms of one or more of these sectors.
3. Cylinders:- In a disk pack, the set of head describe a cylinder as the disk spins at a high speed. The cylinder consists of similar numbered tracks on each of the surfaces. Cylinders are numbered in the same way as tracks. All data on a particular cylinder is equally accessible to the computer without moving the read/write head assembly. Data is therefore organised into cylinders to reduce access time. When the machine has filled one track on one surface with data, it moves to the next surface rather than to the next track on the same surface.
4. Access Speed:- Before the specified sector from a track can be accessed, two actions must take place. First the read/write head must move axially to the track. This involves axial delay or seek time. Next the spinning disk must bring the sector under the head. This gives the rotational delay or latency delay. The maximum rotational delay would be the time taken for one complete revolution while the minimum could be zero. Lastly the access time or the time for actual data transfer must be added. The actual read time is negligible compared to the other two delays.
0 Comments