Most people have thousands upon thousands of megabytes of data on their computer. Indeed, the hard drives we use today to store all that information have grown by astronomical amounts since the early days. Have you ever wondered just how a hard drive actually works? Before we touch on how data is actually read and written to the disk, it’s important to cover the basic elements of a hard drive. We will be primarily covering today’s platter-based hard drives and briefly cover the new solid-state disks (SSDs) you may have heard about as well.
In today’s hard disks, there are seven basic elements to their composition: The enclosure, spindle, platter, motor, actuator arm, interface and logic board. The enclosure is the most simple of the seven pieces, being nothing more than the shell to which everything is installed or bolted. The enclosure houses the internal components (spindle, motor, actuator, platter) and carries the logic board on the underside. All this is completed with an interface on the back.
The Logic Board
Turning the hard drive over reveals the logic board. This is an extremely important piece that handles several important elements. First and foremost, the logic is the instruction manual for the computer attempting to access the drive. This can be called the detection routine, and serves to give the computer an idea of what the drive is, how big it is, what cable it’s connected to and how to access the drive in your OS of choice. With a dead logic, your PC may never detect that you even have a hard drive installed in the PC, much less its capacity or model!
In addition to the detection routine, the logic board holds the read/write cache, which is crucial to the performance of hard disks. If you tell your computer to open 1000 megabytes of information, the hard drive passes the information to you as quickly as it is able. While the hard drive is loading the first 16MB of the file, the next chunk of data is prepared to roll and is waiting in the cache; when you open the cached chunk of data, another is fed into the cache, and so on until all the information is opened. The reverse of this process occurs when writing information as well. This procedure ensures that data retrieval/archival is expedient, overcoming some of the limitations of a hard drive’s inherently-sluggish, mechanical design.
Lastly, the hard drive’s logic board translates the computer’s requests for archival/retrieval into the commands that make the hard drive do its magic and actually read/write information. Logic boards often have more powerful processing units and more memory than computers of the early 1990s!
For more information on logic boards, StorageReview has a fantastic primer on its very complicated function.
The Spindle Motor
Moving to the internal components, the most basic component therein is the spindle motor. Very precisely-controlled by the logic board, the spindle motor is what rotates the platters, or the disks that store data inside the hard drive. For your average desktop hard drive, the spindle hurtles the platters along at 7,200 or 10,000 RPM. In a notebook hard drive, the rotational velocity is usually 4200 or 5400 RPM. Higher rotational velocities can significantly decrease data read and write time, as the location of the data on the platters can be sought more quickly. The spindle motor is connected to the spindle, which holds the platters centered and stable for the high rotational velocities.
Hard Drive Platters
The platters themselves are probably the most important and complicated component. Today’s platters are thin disks (Thus the “Hard disk” name) of glass or aluminum, coated with an ultra-thin layer of a cobalt alloy, which is naturally magnetic. Data is written to sectors which are organized into concentric rings outwards from the spindle called tracks, and all of those are managed into clusters by the file system you’ve chosen. To actually write the data, the actuator arm aligns the magnetization of the platter in a pattern recognizable to the hard drive’s logic board.
Today’s hard drives contain multiple platters, but the real Holy Grail of hard drive technology is the areal density. The areal density is a term that describes how many sectors fit in a given track (Or in², as a standard unit). The benefits of an increased areal density are two-fold. First, a single platter can contain more information, thusly increasing the capacity of a hard drive. Secondly, and more importantly, when the areal density is higher, the actuator arm that is responsible for physically moving to read/write information can move a shorter distance, thereby increasing sustained speeds in long periods of read/write activity.
A new technology called perpendicular recording aligns the sectors on a platter perpendicular to the physical platter, allowing drive manufacturers to shorten the space between sectors, increasing areal density remarkably while improving drive reliability. Older hard drives align sectors horizontally with the platter, wasting precious space platter surface area. Today’s perpendicular storage drives are approximately 180Gbit/in², while technologies in development, such as IBM’s Millipede, technology is at 1Tb/in² but is commercially unavailable.
The actuator arm is the last internal component explicitly responsible for the read/write process, and its job is very simple. Moved to and fro by the activity of two very strong rare earth magnets, two different tips on the actuator can sense the magnetization patterns to read information, and alter magnetization to write information. The actuator’s function is controlled by the logic board, as are all the other internal functions.
Reading/Writing to Disk
In the platter diagram, the green stripe is a cluster that we touched on briefly. A cluster is a collection of sectors grouped together by the file system for more simplified, but not necessarily efficient access. By virtue of clustering, some of the potential storage space on a hard drive can be lost. For example, if a hard drive is formatted with a 4k cluster size, and you write a 2k file to disk, half of that cluster’s storage space is used up by data, and the other half is empty, wasted space. Yet without clustering, disk performance would be very sluggish due to the file system’s inability to access data quickly. Millions of clusters is better than billions of sectors in the long run!
Let’s quickly step through the process of a read/write!
- User requests information on the hard drive.
- Operating system accesses the MFT, or master file table (An index of files and locations), via the motherboard’s hard drive controller to find the file’s cluster.
- Operating system tells the hard drive’s logic board, via the hard drive controller, that it wants a file from a cluster.
- The logic board spins up the platters on the spindle.
- The actuator arm is moved into position.
- The logic board reads and amplifies the very weak, isolated magnetic fields that comprise your data.
- The logic board begins using the actuator to read information from the sectors in the requested cluster.
- Information is streamed into the hard drive cache
- The information is fed from the cache, to the hard drive controller, to you and your RAM!
The write process is almost the exact opposite, except instead of accessing the MFT to find a file’s location, it’s accessing the file table to find free clusters for write space.
Solid State Hard Disks (SSD)
Unlike conventional platters which are, when compared to flash memory, extremely slow due to their mechanical nature, solid-state disks use NAND flash memory to store information. SSDs feature very fast burst read/write speeds, which are periods of disk activity that are much faster than normal due to data being physically close and easy to recover. On the other hand, they are slower than today’s hard drives for writing sequential clusters of information, a process known as sequential writes. Though, due to the nature of a user’s interaction with their PC, a majority of the accessing done on a hard drive is data reading, which SSDs are twice (Or more) as fast at as conventional hard disks.
Also, due to a complete absence of moving parts, SSDs have greatly increased longevity and reliability. The downside to these drives, despite all their amazing benefits, are very high cost per megabyte of capacity, and their low capacity. It is unfortunate that both these traits are what will really stop a consumer from adopting a technology that is superior. There can be no doubt that SSD is the future, but that future is assuredly not now. However, in light of the low capacities and high prices, the technology still has enough merit to be featured in new notebooks arriving in 2007. Reduced power consumption, reduced heat and a long battery life spells great success for this technology in the mobile arena.
There you have it! The seven essential elements of hard drives, and their most likely successor. While SSDs certainly sound promising, don’t think that conventional hard drives are dead. Major manufacturers like Seagate and Western Digital are feverishly dumping millions of R&D into new hard drive technology to further increase the capacity of our retail hard disks. With Hitachi’s 1TB drive, the world’s first, just around the corner, you can rest assured that the war is far from over, and the good old platter isn’t going anywhere any time soon.