If you would like to read the next article in this series please go to
- Memory and Storage - Part 2: New Memory Technologies
- Memory and Storage - Part 3: Bus Specifications
In today’s typical networks there are many different forms of data storage. Understanding these methods of data storage is essential to understanding your network. In this series of articles I will explain some of the more common forms of data storage, some exciting, up-and-coming technologies, and I will outline some of the more common memory addressing protocols.
Hard drives, the most common form of magnetic storage, store data on disk shaped platters. These platters are typically made of aluminum, glass, or ceramics, and are coated with a ferromagnetic material which is often a cobalt alloy. This ferromagnetic coating allows read/write heads to magnetize small regions of the platter which represent a digital bit.
Figure 1: Hard Drive courtesy of www.samsung.com
Platters, yes there is more than one on a hard drive, are separated by spacers on a single spindle. This spindle is controlled by a motor which can spin the platters. This motor’s speed is constant, and is the speed advertised as the speed of the hard drive.
Read/write heads, one per platter side, are attached to a single actuator arm. The actuator arm is controlled with a servo motor which can move the heads closer to, or further away from the spindle in unison.
There are two ways to write the data onto the platter: longitudinal and perpendicular. Longitudinal is the traditional way to write the data onto the platter. You can think of the bits like a bar magnet laying flat on the platter surface, end to end. You can easily imagine that these regions can take up quite a lot of space. This is why hard drive manufacturers have been working to reduce the size of these regions. Currently we are reaching the limits of how small engineers can make these regions. This is because of the superparamagnetic effect. Basically, this effect describes how random thermal effects can flip the polarity of a very small magnet. On a hard drive, if the polarity of one of the magnetized regions is flipped then this would mean that the data is changed from a ‘1’ to a ‘0’. This is bad.
Perpendicular recording technology can allow engineers to pack more data on the same area of the platter without having to worry about the superparamagnetic effect for similarly sized regions. If you picture the bar magnets again, it should be obvious why you can fit more of them in the same area when they are stood on their ends, i.e. perpendicular. It is a little less obvious why they do not have to worry about the superparamagnetic effect. Essentially this is because the direction of the magnetic fields has changed, and therefore they react with their neighbours differently. This reaction is important in determining whether the superparamagnetic effect is in play.
Similarly to hard drives, tape drives store bits by polarizing a small magnetic region. There are basically two types of tape drives: linear and helical.
Linear tape drives have linear tracks. On the tape there are several dozen tracks which extend from one end of the tape to the other. Each track consists of many magnetic small regions which can be used to represent a ‘1’ or a ‘0’.
Helical tapes have tracks that run diagonally up and down the tape. This means that the tracks will overlap each other. Normally this would be bad; however this type of tape drive uses two write heads, each using an opposite polarization which allows the read heads to distinguish between the tracks. This allows for a higher capacity on the tape.
One of the most common types of semiconductor memory is RAM, shown below in figure 2. There are two general types of RAM, static and dynamic. Static RAM or SRAM stores data in a collection of 6 transistors, commonly known as a flip-flop. Dynamic RAM or DRAM stores data in capacitors which require continuous refreshing, and is the reason why DRAM loses the data when power is shut off. The advantage of DRAM is that it only takes one transistor and one capacitor for each bit. This gives a very high memory capacity compared to a similarly sized SRAM chip. The advantage of SRAM is that the transistors do not require refreshing and react faster than capacitors.
Figure 2: Ram courtesy of www.kingston.com
Another form of semiconductor memory growing in popularity is Flash memory. There are two basic types of flash memory, NOR and NAND. NOR (Not OR) refers to the NOR logic gate while NAND (Not AND) refers to the NAND logic gate.
Both NAND and NOR logic gates consist of a collection of transistors, and do not contain capacitors. This means that they do not require refreshing and therefore hold data when power is no longer applied.
Figure 3: A USB flash drive courtesy of www.sandisk.com
Although both NAND and NOR Flash are similar in some respects they are also quite different. NAND flash being a sequential access technology is better suited to be used to store data. NOR flash is a random access technology which makes it better suited to store programs which use little memory. NOR flash is usually used in applications such as running a cell phone’s operating system. NAND flash is typically used in applications such as USB memory sticks. According to www.appleinsider.com, Apple’s new iPhone uses both NOR and NAND flash.
The most common type of optical storage is the CD. CDs are made from injection molded polycarbonate plastic which has microscopic bumps, sometimes called pits, arranged in a continuous spiral around the disk. It is these bumps that represent the data. Over this polycarbonate is a thin layer of reflective material, usually aluminum or gold and over that is an acrylic layer to protect the disk.
When a CD is being read, a laser is shown through the polycarbonate layer and reflected off of the reflective material. The reflected laser light is in turn detected by an optical sensor which converts the received laser signal into electricity. Depending on whether the laser was focused on a bump or not, the electrical signal will be different because the reflected laser light will be different. The difference in the electrical signals is how a computer can read data off of the CD. That is the case for regular CDs, but what about recording data onto CD-Rs, and CD-RWs?
A CD-R is similar to a CD in their construction except for two key aspects. First, there are no bumps. Second, between the polycarbonate and the reflective aluminum there is a layer of transparent dye. To save data onto a CD-R the writing laser is focused onto the desired part of the spiral (which doesn’t really exist until you create it by writing data) and heats up the dye. The chemical properties of the dye are such that when it is heated to sufficiently high temperatures its level of opacity changes. So the writing laser can move along the spiral and change the opacity of small regions, this difference in opacity is how you can create a ‘1’ or a ‘0’. Data is then read from the CD-R in the same manner as a CD. CD-Rs of course can only be written to once. This is because once you make the dye opaque you cannot then make it transparent. So then what about CD-RWs? CD-RWs use a different dye which starts out opaque and when heated turns transparent. This dye also has the amazing property of turning back opaque if heated to an even higher temperature. This allows you to erase the data previously written to the disk.
Figure 4: An image of the bumps on a DVD courtesy of www.optics.rochester.edu
DVDs work exactly the same as CDs. DVDs can store more data because there are essentially many thin CDs stacked on top of each other. That is, they are made of several layers of polycarbonate and reflective material. The lasers and optical sensors are also more advanced, in that the laser has the ability to pass through the different layers and the optical sensor can detect all of these different layers.
Those are some of the more common methods of data storage available. Keep an eye out for my next article where I will discuss some newer, more advanced technologies such as pulse-change and holographic memory.
If you would like to read the next article in this series please go to