Memories can be divided into primary memory and secondary memory for data access and placement of a computer system. Memories can also be broken down by access principle, registration durability over time and after the head of data adaptability.
For processing unit data can be accessed in a single-stage, two-stage or multi-stage operation.
The access time depends on the number of steps.
The main memory is directly accessible from the process unit, i.e., in one step, while the secondary memory is only accessible via the primary memory, i.e., by two steps. Parts of the secondary memory the data must first be transferred to the main memory to be processed by the processing unit.
Accessibility Capacity Access Time
To access data in one step
The cache hundreds of KB 0.1-0.5 US
Primary memory a few MB 0.5-1 US
Expanded main memory many MB 1-2 US
Access two-stage electronically
Secondary memory up to several GB 10-30 US
Access in two or more steps
File Memory hundreds GB seconds
Tape Archive thousands GB minutes
If one considers the memories with regard to their access principle, people talk of random access memories (RAM, Random-Access Memory), cyclic memories and sequential memories.
RAMs characterized in that each part of the memory is accessed at the same time, the access time is constant, this is typical of include primary memories.
Cyclic memories characterized by one to read a portion of the memory at a periodical timing. These include, among others, hard drives and CDs.
Sequential memories characterized in that only read data in a certain sequence, in series. These include, for example, magnetic tape. The access time varies and depends on where you are on the media.
Volatile or volatile memories are separated from non-volatile or persistent. The difference is that the volatile memory is dependent on the constant power supply to keep the data.
Volatile memory is divided into läsflyktiga memories, time fleeting and voltage-dependent memories.
Examples of non-volatile memory is a magnetic tape that can be stored for years without losing data.
This classification is based on the ease of writing into memory (modifying data). The most common types are permanent and erasable. A permanent memory can not be modified without special equipment. WORM drives (Write-Once-Read-Many-Times), or embossed memories can be written once but not erased.
A typical permanent memory is read only memory (ROM, Read-Only-Memory) that have programs from the manufacturer already built applications. A memory that can be loaded by the customer with special firing devices called programmable read only memories, PROM (Programmable ROM). Optical memories of the type semi-permanent memories can be cleared by ultraviolet light.
Examples of erasable memories are read-write memory (RWM, read-only-memory), with it's reading and writing time is the same.
Finally, one can classify memories after the physical functioning, this can be mentioned among other things, semiconductor memory and core memories.
A semiconductor is a material with an electrical conductivity between metals and insulators.
Silicon and germanium are the most important elements of these properties.
In the 50's transistors used as an alternative to pipes, since several interconnected transistors act as relays, ie break and stop flowing.
During the 60s invented the integrated circuit which triggered an ever-increasing development of more and more sophisticated circles. This movement is the basis for the mass production of the complete set of microelectronics-based products that make up our daily lives.
Instead of soldering individual trasistorer larger units, these build directly on a small silicon plate.
Because in a single component implements a full circle is called the integrated circuits.
In the beginning was held not more than a dozen transistors on each Platte with some millimeter side. By 1970, the number up in 100 countries. 1984 passed the dream threshold of one million transistors on a plate.
The semiconductor industry has become increasingly sophisticated and specialized. Now customers can order customized circuits by manufacturers for different applications. Of course made even general circuits components for computer manufacturers. Today computers are built by using the commercially available circuits components combined into an architecture.
In the manufacture of integrated circuits can be built a wiring pattern on a silicon wafer with photographic technique.
The pattern is built up by adding the thin layers of different materials.
Etching steps are controlled by an optical filter made of a CAD / CAM system.
Stocks are only a few angstroms thick, which sometimes makes many circles to be thrown away because it will workmanship.
The circuit is enclosed in a capsule. Capsule legs make contact with the circuit by thin gold wires. A printed circuit board consists of a number of circuits assembled on one (circuit) card.
In order to pack the components further develops new techniques for engraving direct laser.
All layers memories uses a ferrite layer magnetized in small spots and their magnetization may designate one and zero. Reading is done by a read head senses the magnetized spots. A print head then works the other way around by giving rise to a magnetic flux and an entry in the ferrite layer.
Disc Memory is the most used secondary memory because it is inexpensive, has a high storage capacity and a short access time.
It consists of a number of circular magnetic disks are made to rotate around its own axis.
The data is organized so that a physical disk page is divided into a number of circular concentric tracks which in turn is divided into a number of sectors. Logically organized in data blocks. A block includes an even number of sectors. It is the smallest addressable unit on the disc. The file consists of a number of blocks which need not be successive.
There are various types of skivminnnen, e.g., floppy discs, which usually accommodates 1.44 MB, and Winchester memories having varying sizes, usually around 1 GB.
Floppy disks composed of a plastic disc coated with magnetizable material which is enclosed in a cover. In a disk drive is the magnetic head against the disc ferrite which wears on the disc.
Winchester Memories are encapsulated, fixed disks. Often called hard disks.
To access a specific location on the disc have to go through three steps:
1. Position the read / write heads to the right track.
2. Wait for the track passes.
3. Write / read data.
The first two steps are by far the slowest since they are electromechanical.
In modern disk drives floats read and write heads of the air particles as the rotating surface pulls away and therefore never touches the ferrite layer.
The magnetic tape
The tape is the dominant medium for the archive data.
It consists of a coil with a plastic tape.
The plastic band is coated with a ferrite layer.
The bands have standard lengths of 400, 600, 1200, 2400 and 3600 feet.
A tape drive is used to read and write on the tape.
On the tape is a character stored vertically as 7 or 9 bits. A number of characters are gathered in one block. Each block starts and ends with special character that designates the block beginning and block end. Block length indicates how many characters, words or records in a block. Between each block is a block interval which is an empty stretch of the belt.
Storage density is set out in the 'bytes per inch. The highest storage density corresponds to approximately 2460 characters per centimeter. The storage capacity of a tape can amount to several 100 MB. The transmission rate varies between 0.1-2 MB / s. When reading transferred a block at a time to a buffer. Large block length offers higher speed.
The advantages of tape is that they are inexpensive, easy to store, rugged and easily mounted. Not to bands will wear out is held ferrite one hundredth of a millimeter from the read / write head. Band is suitable for archiving and long-term storage.
1984 IBM launched a magnetic memory in the cartridge has a packing density of 38,000 BPI. The transmission rate of this band is 3 Mbyte / s.
The newer optical memories based on laser technology.
These are available in two sizes: optodisketten and optical disc, of which the latter is greater.
At the earliest optoskivorna could only write once.
This was due to a hole in the surface was marked.
When the hole was hit by a beam reflected not this.
On a optoskiva up to 4 GB of storage. This offers the possibility of using new memory when something needs to be changed on the disc. With newer technology is also rewriting possible.
A optodiskett holds about 500 MB or 1 GB. There are also permanent optical memories. One example is the CD-ROM. The advantage of
CD-ROM is that the magnetism does not have any effect.
Shift register is a memory with no moving parts. It consists of a bit registers of the relatively large length. A clock control their impulses shifting the entire registry. The shift register is constructed of the semiconductor. Memories built by a large number of shift registers can reach capacity 1 MB and have a relatively fast access time.
Other memory types
Delay line memory used as the main memory in the early 50s.
It was based on acoustic phenomena.
Drum memory is a magnetic layer memory where the magnetizable layer is located on a cylinder mantle surface. It was used as the main memory in the 50's.
Core memory consists of a number of ferrite matrices strung on wires 3-4 through which the cores could be read and written. The memory used from the 50th century until well into the 70th century.
Bubble Memory is a memory with small bubbles in ferrite. The bubbles can be created electronically and later moved, read and killed.