A HDD or Hard Disk Drive is the device that holds all of the information on your computer, including the operating system (“O/S” - such as Windows, Linux or Apple).
How it works: Traditionally, a hard disk drive is comprised of one or more hard metal disks (“platters”) and an actuator “arm” (sometimes referred to as the read-and-write head) for each that is magnetically maneuvered across the spinning disk in order to access the digital information stored on the drive . For a 240 Mb Western Digital drive, for example, each HDD has three magnetic platters at 80Gb of storage each, with its own actuator arm, for a total of 240Gb storage. The nearly weightless suspension arm hovers a hair (i.e. a nanometer) above the magnetic disk which spins at either 5400 or 7,200 revolutions per minute, server drives as fast as 15,000 rpm. The outer edge of a typical disk, therefore, travels at 100 miles per hour. The arm holds a recording head the size of a pepper flake, which sits above the disk at a height measured in nanometers - less than the ridges of a fingerprint.
Each disk drive contains more than 200 parts, most of them designed for specific models. The data is retained on the disk even when the HDD is off (as opposed to cache or RAM, where the data is not held in memory once the power is turned off. See MEMORY). An orange (usually) flat ribbon cable connects the mechanical portion of the drive to the logic board, a circuit board that contains a tiny, low power processor controller chip located underneath the spinning disks, which routes the data read from the disks via the actuator arm and sends it to the computer’s motherboard through the IDE, SCSI or SATA cable connected to the drive through the connectors at the edge of the HDD (along with the appropriate power cable). (See the diagram below for a graphic look at all this.)
Solid State Drives work much the same way as traditional drives, except there are nothing but chips and absolutely no moving parts. The chips are usually NAND flash memory as well as a (considerably faster) controller chip to route the data to the PCs motherboard much the same way as the physical type HDDs. In addition, most SSDs use a few hundred megabytes of RAM as a buffer between the NAND and the PC. Click HERE for the full story about SSDs.
Hard drives organize their user stored data into regions of the magnetic media called blocks. These blocks are, in turn, located in regions of the disk called a sector. (Sometimes, unfortunately, these terms are used interchangeably.) These drive sectors are located along the concentric tracks on the surface of the disk. Within each sector is also located a servo and error correction information that allows the magnetic heads to stay on track so that they can read and write data, and for the electronic channel to recover the data, if necessary, without errors. The sectors also contain information that allows the file system to reassemble the data blocks into computer files. While a sector is the smallest unit that can be accessed on a HDD, a cluster is a slightly larger unit that is used to organize and identify the files on a disk. Most files take up several clusters of disk space. Each cluster therefore has a unique ID, which enables the drive to locate all the clusters on a disk, as they are rarely contiguous and are “fragmented” (hence the periodic necessity for drive “defragmentation”.)
See the diagram immediately below for a graphic representation of this structure. As you can see, data is stored on the surface of the platter in sectors and tracks. Tracks (sometimes called “cylinders”) are concentric circles (one is selected in red [“A”]), while sectors are shown as (purple [“B”]) pie-shaped wedges radiating out from the center to the edge of the disk. Newer drives store bits in overlapping tracks (known as “shingled magnetic recording”) like shingles on a home’s roof. There are a fixed number of sectors per track. Tracks are small: On a typical drive, about 60 to 100 tracks can fit within the width of a human hair. Each sector contains a fixed number of bytes, usually 512, which comprise a block, which is a single cell in that sector [“C”]. There are about 3 trillion bits per square inch on the platter of a common hard disk. Further, at some point either the operating system or the drive firmware may group together one or more sectors into clusters (shown in green [“D”]).
However, before a computer can write data to a hard drive or any other media, it must be formatted, both at the low-level and the high-level. The low level formatting lays out the tracks and sectors discussed above. This is usually done at the factory, because it is based on a calculation for each drive, using “heads X sectors X tracks”. Each block usually holds 512 bytes of data. But the HDD still can’t be used by the computer without high-level formatting, sometimes called partitioning. This process, usually done through the computer’s operating system, prepares the file storage structure (like FAT or NTFS or ReFS) into the drive’s sectors, so that data can then be written. Each type of formatting has its own rules and limitations. Until this is done, creating a partition table on the disk, the computer’s BIOS (or UEFI, which allows more than 4 partitions per drive) can’t know whether the HDD is to be used as a single disk or whether it will be split into one of more logical disks, as well as which part of the disk the boot sector may be located on.
After reading and writing many files to the HDD, some clusters may remain labeled as being used although they contain no data. These are called lost clusters and can be fixed using Windows Scan Disk or the Mac Utility programs; also defragmentation can help free up additional hdd space. [Now you know what the report is talking about when you run the Windows ScanDisk utility.]
In addition, when installed on a computer, Windows drives may be configured as either basic or dynamic disks. [The concept of a “disk” should not be confused with a “drive”. The drive is the hardware, while the disk is the partition on the drive, formatted with a valid file system such as NTFS.] A basic disk uses primary and extended partitions and logical drives to organize data. A dynamic disk, on the other hand, can contain a very large number of volumes (a/k/a partitions). Most Windows PCs are basic disks, as they are simple to manage and more complexity is unnecessary. Dynamic disks are used more for servers, where multiple disks within a computer or across several computers are used to manage data, such as for RAID applications, as they can be spanned, striped and mirrored to create data reliability. Dynamic disks can only be used on the computer on which they are created, while basic disks may be used on other computers than the one on which it was created.
See also the definitions in the Glossary for the boot sector and the MBR (“Master Boot Record”), typically the first sector in the first partition in the drive which starts the operating system. [Although newer 64-bit drives and operating systems like Windows 8 now use the UEFI instead of the boot sector MBR and, if you have an Apple computer, you use APM).]
For the past 30 years, the size of the sectors on the hard drive was limited to 512 bytes for most OSs, including Windows and the ATA interface. Recently, Linux and other open-source developers have led a standards effort to create a 4,096 byte sector standard often referred to as the 4K sector. This 8x increase in sector size would increase formatting and operating efficiency of the newer, larger drives, especially by allowing for a more robust ECC (Error Correcting Code), critical to creating high area drive densities, and is probably long overdue.
HDDs come in different capacities, speeds and connections (e.g. IDE, SCSI, SATA; disk spin speeds starting at 5400rpm, then 7200 rpm [most common] then 10,000, 12,500 or even 15,000rpm for server drives; capacities from 40Gbs to 10 terabytes, the most common these days about 500Gb; see bits & bytes). Possibly a more important measure of a drive’s speed, however, is its seek time. That is, the time which it actually takes a hard drive controller to locate a specific piece of stored data on the disk. It is measured in nanoseconds or one-thousandth of a second, a range of 7 - 9ns (nanoseconds, or one billionth of a second) being common for most hard drives. Alternate measurements of drive speed include track-to-track time (the time, usually in milliseconds (thousands of a second), it takes the read/write head to search or seek between adjacent tracks, usually 2 - 4ms) and full stroke (the amount of time required to seek the entire disk drive, also measured in milliseconds, usually below 10ms). The measurement begins when the computer’s operating system makes a request to the drive controller firmware to locate certain data. Then, the controller actuates the read/write head to the position where that data is stored on the disk. If the head must switch between tracks on the disk, the actuator must move the access arm (see diagram below), increasing the seek time, and which is also dependent on the starting point for the arm when it completed the previous seek command. Optical drives (CDs/DVDs, floppy disks, and large or mobile drives) have a slower seek time because of their larger head construction. Average time for CD/DVDs is 65ms, 75ms for DVD-RAMs. Of course, these speeds are stated as “average” times, as actual times may vary due to transfer time (“data rate,” or the actual amount of time that it takes for data to be read/written) and rotational delay (“latency,” the actual amount of time it takes for the hard disc to rotate to the position for the read/write head to access the data), and also because there is no industry standard for measurement.
Hard drives can be internal, i.e. connected directly to the motherboard via parallel, SATA, eSATA, SCSI or other cable connections. They can also be connected externally via USB, SATA, eSATA or other cables that can allow them to be physically removed after whatever backup or storage activities have been completed (it’s not a good idea to leave such drives always on, it can wear them out). And, as the capacities for external flash drives have increased exponentially, they are also a good substitute for external memory or even backup.
HARD DRIVE MAINTENANCE: Aside from keeping the drive in a well-ventilated and clean environment, and one without extreme noise (yes, it’s been documented that high noise environments can cause spinning drive platters to misread data or slow performance), there are some utilities that keep it in tip-top shape. Your hard drive’s “health” falls into two categories - physical and logical. Physical health are monitored and checked by the Self-Monitoring, Analysis and Reporting Technology (a/k/a “SMART”) subsystem built into almost every modern hard drive. SMART checks for possible defects in the hardware itself and attempts to avert any upcoming failure. SMART works automatically. Logical health of the drive (i.e. the O/S and files written on the drive) involves manual testing. The Windows O/S has built in software tools for this purpose. From XP on, chkdsk (see check disk) serves that purpose. To get to this utility, in Windows Explorer, right click the appropriate drive, then select Properties, click on the Tools tab and, under Error-checking, click the “Check Now” button. That’s the basic version. For more comprehensive testing, run chkdsk.exe from the command line and you will see various switches that you can select for advanced options. [I usually select “chkdsk c: /f” to find and repair disk errors.]
MORE LINKS: For a more detailed discussion about the relationship between electricity, magnetism and binary computers, click HERE. Also, read about Solid State Drives (SSDs) and Hybrid drives, which have both benefits and drawbacks (like never defragmenting them!). Read those definitions before buying and installing an SSD!! Click HERE for some interesting history about hard drives by PC World Magazine. Also, for those businesses that don’t want to protect their data by migrating it to the CLOUD, there are also lines of fire and flood resistant hard drives (offering, for example, 30 minutes of fire protection up to 1550 degrees Fahrenheit and 72 hours of immersion protection in up to 10 feet of water). See ioSafe drives for between $250 - $600. Also, Virtual Hard Drives. And see FAQ #29 and the discussion at the SSD glossary definition about how to “defragment” or “wipe” your drive if necessary.
Interesting features can be available: For example, if you have an SSD drive, Secure Drives makes a drive equipped with a built-in self-destruct feature, activated to automatically self-destruct when you send an SMS message to its built in cell phone, or several other options.
Finally, drives can get quite HOT. Particularly those on laptops. Make sure that they’re properly ventilated (and dust free), otherwise they may crash. That’s why many laptops must be put on cooling “pads” if they’re used all day - the heat will cause shut-downs.
See the LINKS page for a short history and deconstructed photos of an actual HDD like the one below: