View source for Hard disk drive

{{Short description|Electro-mechanical data storage device}}
{{Redirect|Hard drive|3=Hard Drive (disambiguation)}}
{{Use mdy dates|date=September 2019}}
{{Use American English|date=January 2025}}

[[File:ST3320620AS.jpg|thumb|Front side of the [[Seagate Barracuda]] 320&nbsp;GB HDD]]
[[File:Seagate Hard Drive Back Side.jpg|thumb|Back side of the HDD showing its controller board]]
[[File:2019 Rozebrane dyski twarde.jpg|thumb|A 2.5" and 3.5" HDD with the front cover removed to show its internals]]
[[File:Harddrive-engineerguy.ogv|thumb|Overview how HDDs work]]
{{Memory types}}
[[File:HDD_Startup_and_Shutdown.webm|thumb|HDD with front cover to show its operation]]

A '''hard disk drive''' ('''HDD'''), '''hard disk''', '''hard drive''', or '''fixed disk'''{{Efn|Other terms used to describe various hard disk drives include ''[[disk drive]]'', ''disk file'', ''[[direct access storage device]]'' (DASD), ''[[CKD disk]]'', and ''Winchester disk drive'' (after the [[IBM 3340]]). The term "DASD" includes devices with media other than disks. The term "hard disk drive" can refer to devices with removable media.}} is an electro-mechanical [[data storage device]] that stores and retrieves [[digital data]] using [[magnetic storage]] with one or more rigid rapidly rotating [[hard disk drive platter|platters]] coated with magnetic material. The platters are paired with [[disk read-and-write head|magnetic heads]], usually arranged on a moving [[actuator]] arm, which read and write data to the platter surfaces.<ref name="ostep">{{cite web |title=Operating Systems: Three Easy Pieces, Chapter: Hard Disk Drives |url=http://pages.cs.wisc.edu/~remzi/OSTEP/file-disks.pdf |publisher=Arpaci-Dusseau Books |year=2014 |first1=Remzi H. |last1=Arpaci-Dusseau |first2=Andrea C. |last2=Arpaci-Dusseau |access-date=March 7, 2014 |archive-url=https://web.archive.org/web/20150216141507/http://pages.cs.wisc.edu/~remzi/OSTEP/file-disks.pdf |archive-date=February 16, 2015 |url-status=live }}</ref> Data is accessed in a [[random-access]] manner, meaning that individual [[Block (data storage)|blocks]] of data can be stored and retrieved in any order. HDDs are a type of [[non-volatile storage]], retaining stored data when powered off.<ref>{{cite book |last1=Patterson |first1=David |first2=John |last2=Hennessy |title=Computer Organization and Design: The Hardware/Software Interface |publisher=[[Elsevier]]|date=1971 |page=23 |url=https://books.google.com/books?id=1lD9LZRcIZ8C&pg=PA23|isbn=9780080502571 }}</ref><ref>{{cite web|last1=Domingo|first1=Joel|title=SSD vs. HDD: What's the Difference?|url=http://uk.pcmag.com/storage-devices-reviews/8061/feature/ssd-vs-hdd-whats-the-difference|publisher=[[PC Magazine]] UK|access-date=March 21, 2018|archive-url=https://web.archive.org/web/20180328191847/http://uk.pcmag.com/storage-devices-reviews/8061/feature/ssd-vs-hdd-whats-the-difference|archive-date=March 28, 2018|url-status=live}}</ref><ref>{{cite book|chapter=Implications of non-volatile memory as primary storage for database management systems|publisher=[[IEEE]]|doi=10.1109/SAMOS.2016.7818344|title=2016 International Conference on Embedded Computer Systems: Architectures, Modeling and Simulation (SAMOS)|pages=164–171|year=2016|last1=Mustafa|first1=Naveed Ul|last2=Armejach|first2=Adria|last3=Ozturk|first3=Ozcan|last4=Cristal|first4=Adrian|last5=Unsal|first5=Osman S.|isbn=978-1-5090-3076-7|hdl=11693/37609|s2cid=17794134}}</ref> Modern HDDs are typically in the form of a small [[disk enclosure|rectangular box]].

Hard disk drives were introduced by [[IBM]] in 1956,<ref name="IBM350" /> and were the dominant [[secondary storage]] device for [[History of general-purpose CPUs|general-purpose computers]] beginning in the early 1960s. HDDs maintained this position into the modern era of [[server (computing)|servers]] and [[personal computer]]s, though personal computing devices produced in large volume, like [[mobile phone]]s and [[Tablet computer|tablets]], rely on [[flash memory]] storage devices. More than 224 companies have [[List of defunct hard disk manufacturers|produced HDDs historically]], though after extensive industry consolidation, most units are manufactured by [[Seagate Technology|Seagate]], [[Toshiba]], and [[Western Digital]]. HDDs dominate the volume of storage produced ([[exabyte]]s per year) for servers. Though production is growing slowly (by exabytes shipped<ref>{{Cite web|url=https://www.anandtech.com/show/15117/demand-for-hdd-storage-booming-240-eb-83-million-drives-shipped-in-q3-2019|title=Demand for HDD Storage Booming: 240 EB Shipped in Q3 2019|first=Anton|last=Shilov|website=AnandTech |date=November 18, 2019 |url-status=live |archive-url=https://web.archive.org/web/20230827154848/https://www.anandtech.com/show/15117/demand-for-hdd-storage-booming-240-eb-83-million-drives-shipped-in-q3-2019 |archive-date= Aug 27, 2023 }}</ref>), sales revenues and unit shipments are declining, because [[solid-state drive]]s (SSDs) have higher data-transfer rates, higher areal storage density, somewhat better reliability,<ref name="BBRel2021"/><ref>{{cite web|url=http://www.intel.de/content/dam/doc/technology-brief/intel-it-validating-reliability-of-intel-solid-state-drives-brief.pdf|title=Validating the Reliability of Intel Solid-State Drives|date=July 2011|publisher=Intel|access-date=February 10, 2012|archive-url=https://web.archive.org/web/20161019001340/http://www.intel.de/content/dam/doc/technology-brief/intel-it-validating-reliability-of-intel-solid-state-drives-brief.pdf|archive-date=October 19, 2016|url-status=live}}</ref> and much lower latency and access times.<ref>{{cite web|url          = https://ieeemagnetics.org/files/IEEEMS-N-54-2-final.pdf|title        = 5th Non-Volatile Memories Workshop (NVMW 2014)|date         = March 2014|publisher    = IEEE|access-date   = February 21, 2023|first        = Eric|last         = Fullerton|archive-url  = https://web.archive.org/web/20180928051219/http://ieeemagnetics.org/files/IEEEMS-N-54-2-final.pdf|archive-date = September 28, 2018|url-status   = dead
}}</ref><ref>{{cite web|url          = https://thessdguy.com/for-the-lack-of-a-fab/#more-538|title        = For the Lack of a Fab...|last         = Handy|first        = James|date         = July 31, 2012|publisher    = Objective Analysis|access-date   = November 25, 2012|archive-url  = https://web.archive.org/web/20130101051214/http://thessdguy.com/for-the-lack-of-a-fab/#more-538|archive-date = January 1, 2013|url-status   = dead
}}</ref><ref name="arstechnica.com">Hutchinson, Lee. (June 25, 2012) [https://arstechnica.com/information-technology/2012/06/inside-the-ssd-revolution-mobile-devices-and-modern-oss/ How SSDs conquered mobile devices and modern OSes] {{Webarchive|url=https://web.archive.org/web/20170707230253/https://arstechnica.com/information-technology/2012/06/inside-the-ssd-revolution-mobile-devices-and-modern-oss/ |date=July 7, 2017 }}. Ars Technica. Retrieved January 7, 2013.</ref><ref name="Santo Domingo">{{cite news|url =https://www.pcmag.com/article2/0,2817,2404258,00.asp|title =SSD vs HDD: What's the Difference?|last =Santo Domingo|first =Joel|date =May 10, 2012|access-date =November 24, 2012|work =PC Magazine|archive-url =https://web.archive.org/web/20170319184712/http://www.pcmag.com/article2/0,2817,2404258,00.asp|archive-date =March 19, 2017|url-status =live}}</ref>

The revenues for SSDs, most of which use [[NAND flash memory]], slightly exceeded those for HDDs in 2018.<ref name= Barrons>{{cite web |url=https://www.barrons.com/articles/why-western-digital-can-gain-45-despite-declining-business-1526323466 |title=Why Western Digital Can Gain 45% Despite Declining HDD Business |last=Hough |first=Jack |date=May 14, 2018 |publisher=Barron's |access-date=May 15, 2018 |archive-url=https://web.archive.org/web/20180515011821/https://www.barrons.com/articles/why-western-digital-can-gain-45-despite-declining-business-1526323466 |archive-date=May 15, 2018 |url-status=live }}</ref> Flash storage products had more than twice the revenue of hard disk drives {{as of|2017|lc=on}}.<ref name= "Register UK 2017" >{{cite web |url= https://www.theregister.co.uk/2017/07/31/flash_drive_revenues_twice_hdd_revenues/ |title= NAND that's that... Flash chip industry worth twice disk drive biz |last= Mellor |first= Chris |website= [[The Register]] |date= July 31, 2017 |access-date= November 21, 2019 }}</ref> Though SSDs have four to nine times higher cost per bit,<ref name="jcmit 2019"/><ref name= "blocks August2019" >{{cite web |url= https://blocksandfiles.com/2019/08/28/nearline-disk-drives-ssd-attack/ |website=Blocks and Files |title= How long before SSDs replace nearline disk drives? |last= Mellor |first= Chris |date= August 28, 2019 |access-date= November 15, 2019 }}</ref> they are replacing HDDs in applications where speed, power consumption, small size, high capacity and durability are important.<ref name="arstechnica.com"/><ref name="Santo Domingo"/> {{As of|2017}}, the cost per bit of SSDs is falling, and the price premium over HDDs has narrowed.<ref name= "blocks August2019"/>

The primary characteristics of an HDD are its capacity and [[Hard disk drive performance characteristics|performance]]. Capacity is specified in [[unit prefix]]es corresponding to powers of {{Not a typo|1000}}: a 1-[[terabyte]] (TB) drive has a capacity of {{Not a typo|1,000}} [[gigabyte]]s, where 1 gigabyte = 1 000 megabytes = 1 000 000 kilobytes (1 million) = 1 000 000 000 [[byte]]s (1 billion). Typically, some of an HDD's capacity is unavailable to the user because it is used by the [[file system]] and the computer [[operating system]], and possibly inbuilt redundancy for error correction and recovery. There can be confusion regarding storage capacity since capacities are stated in decimal gigabytes (powers of 1000) by HDD manufacturers, whereas the most commonly used operating systems report capacities in powers of 1024, which results in a smaller number than advertised. Performance is specified as the time required to move the heads to a track or cylinder (average access time), the time it takes for the desired sector to move under the head (average [[Latency (engineering)|latency]], which is a function of the physical [[rotational speed]] in [[revolutions per minute]]), and finally, the speed at which the data is transmitted (data rate).

The two most common form factors for modern HDDs are 3.5-inch, for desktop computers, and 2.5-inch, primarily for laptops. HDDs are connected to systems by standard [[computer interface|interface]] cables such as [[SATA]] (Serial ATA), [[USB]], SAS ([[Serial Attached SCSI]]), or [[Parallel ATA|PATA]] (Parallel ATA) cables.

== History ==
{{Main|History of hard disk drives}}
{{Infobox computer hardware
| image = IBM 350 RAMAC.jpg
| caption     = A partially disassembled IBM 350 hard disk drive (RAMAC)
| invent-date = {{Start date and age|1954|12|24}}{{Efn|This is the original filing date of the application which led to US Patent 3,503,060, generally accepted as the definitive hard disk drive patent.<ref>Kean, David W., 1977, ''IBM San Jose: A quarter century of innovation''. San Jose, CA: International Business Machines Corporation. CHM accession number: 102687875.</ref>}}
| invent-name = [[IBM]] team led by [[Rey Johnson]] 
}}

{| class="wikitable floatright" style="max-width: 35em;"
|+ Improvement of HDD characteristics over time
|-
! Parameter !! Started&nbsp;with&nbsp;(1957) !! Improved&nbsp;to!! Improvement
|-
| Capacity<br />(formatted) || 3.75&nbsp;[[megabyte]]s<ref name="350Cap" /> 
|| 36&nbsp;[[terabyte]]s ({{as of|2025|lc=on}})<ref>{{Cite web |last=Athow |first=Desire |date=2025-01-21 |title=Seagate smashes largest HDD world record with 36TB hard drive and reveals a 60TB model is coming |url=https://www.techradar.com/pro/seagate-smashes-largest-hdd-world-record-with-36tb-hard-drive-and-reveals-a-60tb-model-is-coming |access-date=2025-03-18 |website=TechRadar |language=en}}</ref><ref name="x-mozaic"/><ref>[https://investors.seagate.com/news/news-details/2024/Seagates-Breakthrough-30TB-Hard-Drives-Ramp-Volume-Marking-an-Inflection-Point-in-the-Storage-Industry/default.aspx Seagate’s Breakthrough 30TB+ Hard Drives Ramp Volume, Marking an Inflection Point in the Storage Industry]</ref>
|| 9.6-million-to-one{{efn|36,000,000,000,000 ÷ 3,750,000}}
|-
| Physical volume || {{convert|68|cuft|lk=on}}{{Efn|Comparable in size to two large refrigerators.}}<ref name="IBM350" /> || {{convert|2.1|cuin|cm3|lk=on}}<ref name="AutoMK-59">{{cite press release | title = Toshiba Storage Solutions&nbsp;– MK3233GSG | url = https://www.toshiba.co.jp/about/press/2009_11/pr0501.htm | access-date = November 7, 2009 | archive-url = https://web.archive.org/web/20120724122541/http://www.toshiba.co.jp/about/press/2009_11/pr0501.htm | archive-date = July 24, 2012 | url-status = live }}</ref>{{Efn| The 1.8-inch form factor is obsolete; sizes smaller than 2.5 inches have been replaced by flash memory.}} || 56,000-to-one{{efn|68 × 12 × 12 × 12 ÷ 2.1}}
|-
| Weight || {{convert|2000|lb|kg|lk=on|disp=br()|abbr=out}}<ref name="IBM350" /> || {{convert|2.2|oz|g|lk=on|disp=br()|abbr=out}}<ref name="AutoMK-59"/> || 15,000-to-one{{efn|910,000 ÷ 62}}
|-
| Average [[access time]] || approx. 600&nbsp;[[millisecond]]s<ref name="IBM350" /> || 2.5&nbsp;ms to 10&nbsp;ms; RW RAM dependent || about<br />200-to-one{{efn|600 ÷ 2.5}}
|-
| Price || {{US$|9,200}} per megabyte (1961;<ref>Ballistic Research Laboratories "A THIRD SURVEY OF DOMESTIC ELECTRONIC DIGITAL COMPUTING SYSTEMS," March 1961, [http://www.ed-thelen.org/comp-hist/BRL61-ibm03.html#IBM-305-RAMAC section on IBM 305 RAMAC] {{Webarchive|url=https://web.archive.org/web/20150302120119/http://ed-thelen.org/comp-hist/BRL61-ibm03.html#IBM-305-RAMAC |date=March 2, 2015 }} (p. 314-331) states a $34,500 purchase price which calculates to $9,200/MB.</ref> {{US$|97,500}} in 2022)
|| US$14.4 per [[terabyte]] by end of 2022<ref>{{cite web
 |url=https://www.backblaze.com/blog/hard-drive-cost-per-gigabyte/
 |title=Hard Drive Cost Per Gigabyte |last=Klein |first=Andy |date=November 29, 2022
 |publisher=Backblaze |access-date=November 22, 2023}}</ref>
|| 6.8-billion-to-one{{efn|(97,500 ÷ 14.4] *  10^6.}}<!-- seemingly outdated -->
|-
| [[Data density]] || 2,000&nbsp;[[bit]]s per [[square inch]]<ref>{{cite web
 | url =https://www-03.ibm.com/ibm/history/exhibits/storage/storage_magnetic.html
 | title =Magnetic head development
 | website =IBM Archives
 | access-date =August 11, 2014
 | archive-url =https://web.archive.org/web/20150321114231/http://www-03.ibm.com/ibm/history/exhibits/storage/storage_magnetic.html
 | archive-date =March 21, 2015
 | url-status =dead
 }}</ref> || 1.4 [[terabit]]s per square inch in 2023<ref name="WDHC680">{{cite web
 |url=https://www.westerndigital.com/tools/documentRequestHandler?docPath=/content/dam/doc-library/en_us/assets/public/western-digital/product/data-center-drives/ultrastar-dc-hc600-series/data-sheet-ultrastar-dc-hc680.pdf
 |title=Ultrastar DC HC690 Data Sheet |date=October 2023 |publisher=Western Digital
 |access-date=November 22, 2023}}</ref> || 700-million-to-one{{efn|1,400,000,000,000 ÷ 2,000.}}
|-
| Average lifespan || c. 2000 hrs [[MTBF]]{{citation needed|date=November 2016}} || c. 2,500,000 hrs (~285 years) MTBF<ref>{{cite web |url=https://www.hgst.com/products/hard-drives/ultrastar-he12 |title=Ultrastar DC HC500 Series HDD |work=Hgst.com |access-date=February 20, 2019 |archive-url=https://web.archive.org/web/20180829180648/http://www.hgst.com/products/hard-drives/ultrastar-he12 |archive-date=August 29, 2018 |url-status=live }}</ref> || 1250-to-one{{efn|2,500,000 ÷ 2,000.}}
|}

=== 1950s–1960s ===
The first production IBM hard disk drive, the [[IBM 350|350 disk storage]], shipped in 1957 as a component of the IBM 305 RAMAC system. It was approximately the size of two large refrigerators and stored five million six-bit characters (3.75 [[megabyte]]s)<ref name="350Cap">{{cite web
|title=Comment: Time Capsule, 1956 Hard Disk
|url=http://www.oracle.com/technetwork/issue-archive/2014/14-jul/o44timecapsule-2219543.html
|quote=IBM 350 disk drive held 3.75&nbsp;MB
|series=Oracle Magazine
|date=July 2014
|access-date=September 19, 2014
|archive-url=https://web.archive.org/web/20140811202917/http://www.oracle.com/technetwork/issue-archive/2014/14-jul/o44timecapsule-2219543.html
|archive-date=August 11, 2014
|url-status=live
}}</ref> on a stack of 52 disks (100 surfaces used).<ref name="ibm-350-dsu">{{cite web|title=IBM Archives: IBM&nbsp;350 disk storage unit|url=https://www-03.ibm.com/ibm/history/exhibits/storage/storage_350.html|publisher=IBM|access-date=July 26, 2015|date=January 23, 2003|archive-url=https://web.archive.org/web/20150617040819/https://www-03.ibm.com/ibm/history/exhibits/storage/storage_350.html|archive-date=June 17, 2015|url-status=dead}}</ref> The 350 had a single arm with two read/write heads, one facing up and the other down, that moved both horizontally between a pair of adjacent platters and vertically from one pair of platters to a second set.<ref name=ibm-650>{{citation
 |   title = IBM 650 RAMAC Manual of Operations
 |      id = 22-6270-3
 | edition = 4th
 |    date = June 1, 1957
 |   quote = Three mechanically independent access arms are provided for each file unit, and each arm can be independently directed to any track in the file.
 | chapter = 355 DISK STORAGE
 |    page = 17
 }}
</ref><ref name=ibm-7070>{{citation
 |   title = IBM Reference Manual 7070 Data Processing System
 |      id = A22-7003-1
 | edition = 2nd
 |    date = January 1960
 |   quote = Each disk-storage unit has three mechanically independent access arms, all of which can be seeking at the same time. 
 | chapter = Disk Storage
 |     chapter-url = https://bitsavers.org/pdf/ibm/7070/A22-7003-01_7070_Reference_Jan60.pdf
 }}
</ref><ref name=ibm-1401>{{citation
 |   title = Reference Manual IBM 1401 Data Processing System
 |      id = A24-1403-5
 |    date = April 1962
 | edition = 6th
 |   quote = The disk storage unit can have two access arms. One is standard and the other is available as a special feature.
 | chapter = IBM RAMAC 1401 System
 |    page = 63
 |     chapter-url = https://bitsavers.org/pdf/ibm/1401/A24-1403-5_1401_Reference_Apr62.pdf
 }}
</ref>  Variants of the IBM 350 were the [[IBM 355]], [[IBM 7300]] and [[IBM 1405]].

In 1961, IBM announced, and in 1962 shipped, the IBM&nbsp;1301 disk storage unit,<ref>{{cite web|url=https://www-03.ibm.com/ibm/history/exhibits/storage/storage_1301.html|title=IBM Archives: IBM 1301 disk storage unit|work=ibm.com|date=January 23, 2003|access-date=June 25, 2015|archive-url=https://web.archive.org/web/20141219183129/http://www-03.ibm.com/ibm/history/exhibits/storage/storage_1301.html|archive-date=December 19, 2014|url-status=dead}}</ref> which superseded
the IBM&nbsp;350 and similar drives. The 1301 consisted of one (for Model 1) or two (for model 2) modules, each containing 25 platters, each platter about {{convert|1/8|inch|mm|adj=on}} thick and {{convert|24|inch|mm}} in diameter.<ref>{{cite web|url=http://www.computermuseum.li/Testpage/DiskPlatter-1301.htm|title=DiskPlatter-1301|work=computermuseum.li|url-status=dead|archive-url=https://web.archive.org/web/20150328161124/http://www.computermuseum.li/Testpage/DiskPlatter-1301.htm|archive-date=March 28, 2015}}</ref> While the earlier IBM disk drives used only two read/write heads per arm, the 1301 used an array of 48{{efn|40 for user data, one for format tracks, 6 for alternate surfaces and one for maintenance.}} heads (comb), each array moving horizontally as a single unit, one head per surface used. [[Cylinder-head-sector|Cylinder-mode]] read/write operations were supported, and the heads flew about 250 micro-inches (about 6&nbsp;μm) above the platter surface. Motion of the head array depended upon a binary adder system of hydraulic actuators which assured repeatable positioning. The 1301 cabinet was about the size of three large refrigerators placed side by side, storing the equivalent of about 21 million eight-bit bytes per module. Access time was about a quarter of a second.

Also in 1962, IBM introduced the [[IBM 1311|model 1311]] disk drive, which was about the size of a washing machine and stored two million characters on a removable [[disk pack]]. Users could buy additional packs and interchange them as needed, much like reels of [[magnetic tape]]. Later models of removable pack drives, from IBM and others, became the norm in most computer installations and reached capacities of 300 megabytes by the early 1980s. Non-removable HDDs were called "fixed disk" drives.

In 1963, IBM introduced the 1302,<ref name="1301Ref">{{cite book
 |     title = IBM 1301, Models 1 and 2, Disk Storage and IBM 1302, Models 1 and 2, Disk Storage with IBM 7090, 7094 and 7094 II Data Processing Systems
 |        id = A22-6785
 |       url = https://bitsavers.org/pdf/ibm/7090/A22-6785_1301_1302_on_709x.pdf
 | publisher = IBM
 }}
</ref> with twice the track capacity and twice as many tracks per cylinder as the 1301. The 1302 had one (for Model 1) or two (for Model 2) modules, each containing a separate comb for the first 250 tracks and the last 250 tracks.

{{Anchor|fixed-head}}Some high-performance HDDs were manufactured with one head per track, ''e.g.'', Burroughs B-475 in 1964, [[History of IBM magnetic disk drives#IBM 2305|IBM&nbsp;2305]] in 1970, so that no time was lost physically moving the heads to a track and the only latency was the time for the desired block of data to rotate into position under the head.<ref>Microsoft Windows NT Workstation 4.0 Resource Guide 1995, Chapter 17 – Disk and File System Basics</ref> Known as fixed-head or head-per-track disk drives, they were very expensive and are no longer in production.<ref name="CHAUDHURI2008">{{Cite book |last=Chaudhuri |first=P. Pal |title=Computer Organization and Design |url=https://books.google.com/books?id=5LNwVRpfkRgC&q=%22fixed+head%22&pg=PA568 |date=April 15, 2008 |publisher=PHI Learning Pvt. Ltd. |isbn=978-81-203-3511-0 |page=568|edition=3rd }}</ref>

=== 1970s ===
In 1973, IBM introduced a new type of HDD code-named "[[IBM 3340|Winchester]]". Its primary distinguishing feature was that the disk heads were not withdrawn completely from the stack of disk platters when the drive was powered down. Instead, the heads were allowed to "land" on a special area of the disk surface upon spin-down, "taking off" again when the disk was later powered on. This greatly reduced the cost of the head actuator mechanism but precluded removing just the disks from the drive as was done with the disk packs of the day. Instead, the first models of "Winchester technology" drives featured a removable disk module, which included both the disk pack and the head assembly, leaving the actuator motor in the drive upon removal. Later "Winchester" drives abandoned the removable media concept and returned to non-removable platters.

In 1974, IBM introduced the swinging arm actuator, made feasible because the Winchester recording heads function well when skewed to the recorded tracks. The simple design of the IBM GV (Gulliver) drive,<ref>"Design of a Swinging Arm Actuator for a disk file" J. S. HEATH IBM J. RES. DEVELOP. July 1976</ref> invented at IBM's UK Hursley Labs, became IBM's most licensed electro-mechanical invention<ref>US 3,849,800 Magnetic disk apparatus. Cuzner, Dodman, Heath, & Rigbey</ref> of all time, the actuator and filtration system being adopted in the 1980s eventually for all HDDs, and still universal nearly 40 years and 10 billion arms later.

Like the first removable pack drive, the first "Winchester" drives used platters {{convert|14|in}} in diameter. In 1978, IBM introduced a swing arm drive, the IBM 0680 (Piccolo), with eight-inch platters, exploring the possibility that smaller platters might offer advantages. Other eight-inch drives followed, then {{convert|5+1/4|in|abbr=on}} drives, sized to replace the contemporary [[floppy disk drive]]s. The latter were primarily intended for the then fledgling personal computer (PC) market.

=== 1980s–1990s ===
Over time, as recording densities were greatly increased, further reductions in disk diameter to 3.5" and 2.5" were found to be optimum. Powerful rare-earth magnet materials became affordable during this period and were complementary to the swing arm actuator design to make possible the compact form factors of modern HDDs.

As the 1980s began, HDDs were a rare and very expensive additional feature in PCs, but by the late 1980s, their cost had been reduced to the point where they were standard on all but the cheapest computers.

Most HDDs in the early 1980s were sold to PC end users as an external, add-on subsystem. The subsystem was not sold under the drive manufacturer's name but under the subsystem manufacturer's name such as [[Corvus Systems]] and [[Tallgrass Technologies]], or under the PC system manufacturer's name such as the [[Apple ProFile]]. The [[IBM PC/XT]] in 1983 included an internal 10&nbsp;MB HDD, and soon thereafter, internal HDDs proliferated on personal computers.

External HDDs remained popular for much longer on the [[Apple Macintosh]]. Many Macintosh computers made between 1986 and 1998 featured a [[SCSI]] port on the back, making external expansion simple. Older compact Macintosh computers did not have user-accessible hard drive bays (indeed, the [[Macintosh 128K]], [[Macintosh 512K]], and [[Macintosh Plus]] did not feature a hard drive bay at all), so on those models, external SCSI disks were the only reasonable option for expanding upon any internal storage.

=== 21st century ===
HDD improvements have been driven by increasing [[Density (computer storage)|areal density]], listed in the table above. Applications expanded through the 2000s, from the [[mainframe computer]]s of the late 1950s to most [[mass storage]] applications including computers and consumer applications such as storage of entertainment content.

In the 2000s and 2010s, NAND began supplanting HDDs in applications requiring portability or high performance. NAND performance is improving faster than HDDs, and applications for HDDs are eroding. In 2018, the largest hard drive had a capacity of 15&nbsp;TB, while the largest capacity SSD had a capacity of 100&nbsp;TB.<ref name= "100TB 2018"/> In 2018, HDDs were forecast to reach 100&nbsp;TB capacities around 2025 (but so far in 2025 the reality is far off),<ref>{{cite web|last=Mott |first=Nathaniel |url=http://tomshardware.com/news/seagate-hamr-100tb-drives-2025,38035.html |title=Seagate Wants to Ship 100TB HDDs by 2025 |work=Tomshardware.com |date=November 7, 2018 |access-date=February 20, 2019}}</ref> but {{as of|2019|lc=on}}, the expected pace of improvement was pared back to 50&nbsp;TB by 2026.<ref name= "blocks WWrevenue August2019" >{{cite web |url= https://blocksandfiles.com/2019/09/23/seagate-assumes-ssds-wont-kill-disk-drives/ |title= How long before SSDs replace nearline disk drives? |last= Mellor |first= Chris |date= September 23, 2019 |quote= the total addressable market for disk drives will grow from $21.8bn in 2019 |access-date= November 15, 2019 }}</ref> Smaller form factors, 1.8-inches and below, were discontinued around 2010. The cost of solid-state storage (NAND), represented by [[Moore's law]], is improving faster than HDDs. NAND has a higher [[price elasticity of demand]] than HDDs, and this drives market growth.<ref name= Appleton >{{cite web |url= https://www.cnet.com/news/flash-goes-the-notebook/ |title= Flash goes the notebook |last= Kanellos |first= Michael |date= January 17, 2006 |website= CNET |access-date= May 15, 2018 |archive-url= https://web.archive.org/web/20180519203122/https://www.cnet.com/news/flash-goes-the-notebook/ |archive-date= May 19, 2018 |url-status= live }}</ref> During the late 2000s and 2010s, the [[Marketing#Product life cycle|product life cycle]] of HDDs entered a mature phase, and slowing sales may indicate the onset of the declining phase.<ref name= lifecycle >{{cite web |url= https://www.inc.com/encyclopedia/industry-life-cycle.html |title= Industry Life Cycle - Encyclopedia - Business Terms |work= Inc. |access-date= May 15, 2018 |archive-url= https://web.archive.org/web/20180708234048/https://www.inc.com/encyclopedia/industry-life-cycle.html |archive-date= July 8, 2018 |url-status= live }}</ref>

The [[2011&nbsp;Thailand floods]] damaged the manufacturing plants and impacted hard disk drive cost adversely between 2011 and 2013.<ref>{{cite web
 | url = https://blog.backblaze.com/2012/10/09/backblaze_drive_farming/
 | title = Farming hard drives: how Backblaze weathered the Thailand drive crisis
 | year = 2013
 | access-date = May 23, 2014
 | website = blaze.com
 | archive-url = https://web.archive.org/web/20140625060029/http://blog.backblaze.com/2012/10/09/backblaze_drive_farming/
 | archive-date = June 25, 2014
 | url-status = live
 }}</ref>

In 2019, [[Western Digital]] closed its last Malaysian HDD factory due to decreasing demand, to focus on SSD production.<ref>{{cite web |url=https://www.theregister.com/2018/07/17/western_digital_petaling_jaya_malaysia/ |title=Western Digital formats hard disk drive factory as demand spins down |work=The Register |last=Mellor |first=Chris |date=July 17, 2018 |access-date=July 21, 2021}}</ref> All three remaining HDD manufacturers have had decreasing demand for their HDDs since 2014.<ref>{{cite web |url=https://www.extremetech.com/computing/273849-western-digital-to-close-hdd-plant-increase-ssd-production |title=Western Digital to Close HDD Plant, Increase SSD Production |work=extremetech.com |last=Hruska |first=Joel |date=July 20, 2018 |access-date=July 21, 2021 }}</ref>

== Technology ==
[[File:magneticMedia.svg|thumb|Magnetic cross section & [[frequency modulation]] encoded binary data]]

=== Magnetic recording ===
{{See also|Magnetic storage}}

A modern HDD records data by magnetizing a thin film of [[ferromagnetic material]]{{Efn|Initially gamma iron oxide particles in an epoxy binder, the recording layer in a modern HDD typically is domains of a granular Cobalt-Chrome-Platinum-based alloy physically isolated by an oxide to enable [[perpendicular recording]].<ref>{{Cite arXiv |title=New Paradigms in Magnetic Recording |eprint = 1201.5543|last1 = Plumer|first1 = M. L.|last2 = van Ek|first2 = J.|last3 = Cain|first3 = W. C.|year = 2012|class = physics.pop-ph}}</ref>}} on both sides of a disk. Sequential changes in the direction of magnetization represent binary data [[bit]]s. The data is read from the disk by detecting the transitions in magnetization.  User data is encoded using an encoding scheme, such as [[run-length limited]] encoding,{{Efn|Historically a variety of run-length limited codes have been used in magnetic recording including for example, codes named [[Frequency modulation|FM]], [[Modified Frequency Modulation|MFM]] and [[Group Coded Recording|GCR]] which are no longer used in modern HDDs.}} which determines how the data is represented by the magnetic transitions.

A typical HDD design consists of a ''{{visible anchor|spindle|Spindle}}'' that holds flat circular disks, called [[hard disk drive platter|platters]], which hold the recorded data. The platters are made from a non-magnetic material, usually [[aluminum alloy]], [[glass]], or [[ceramic]]. They are coated with a shallow layer of magnetic material typically 10–20 [[nanometer|nm]] in depth, with an outer layer of carbon for protection.<ref name="headcrash" /><ref name="AutoMK-6" /><ref name="AutoMK-7"  /> For reference, a standard piece of copy paper is {{convert|0.07|-|0.18|mm|nm|sp=us|abbr=on}}<ref name="AutoMK-8" /> thick.

[[File:Toshiba MK1403MAV - broken glass platter-93375.jpg|thumb|left|Destroyed hard disk, glass platter visible]]
[[File:Hard drive-en.svg|thumb|left|Diagram labeling the major components of a computer HDD]]
[[File:Aufnahme einzelner Magnetisierungen gespeicherter Bits auf einem Festplatten-Platter..jpg|thumb|Recording of single magnetisations of bits on a 200&nbsp;MB HDD-platter (recording made visible using CMOS-MagView)<ref name="AutoMK-9" />]]
[[File:Perpendicular Recording Diagram.svg|thumb|Longitudinal recording (standard) & [[perpendicular recording]] diagram]]

The platters in contemporary HDDs are spun at speeds varying from {{val|4200|ul=rpm}} in energy-efficient portable devices, to 15,000&nbsp;rpm for high-performance servers.<ref name="AutoMK-10" /> The first HDDs spun at 1,200&nbsp;rpm<ref name="IBM350" /> and, for many years, 3,600&nbsp;rpm was the norm.<ref>{{cite web |url=https://www.karlstechnology.com/blog/hard-drive-spindle-speed/ |title=Hard Drive Spindle Speed |publisher=The PC Guide |last=Kozierok |first=Charles |access-date=May 26, 2019 |date=October 20, 2018 |archive-url=https://web.archive.org/web/20190526103244/https://www.karlstechnology.com/blog/hard-drive-spindle-speed/ |archive-date=May 26, 2019 |url-status=live }}</ref> {{As of|November 2019}}, the platters in most consumer-grade HDDs spin at 5,400 or 7,200&nbsp;rpm.

Information is written to and read from a platter as it rotates past devices called [[disk read-and-write head|read-and-write heads]] that are positioned to operate very close to the magnetic surface, with their [[flying height]] often in the range of tens of nanometers. The read-and-write head is used to detect and modify the magnetization of the material passing immediately under it.

In modern drives, there is one head for each magnetic platter surface on the spindle, mounted on a common arm. An actuator arm (or access arm) moves the heads on an arc (roughly radially) across the platters as they spin, allowing each head to access almost the entire surface of the platter as it spins. The arm is moved using a [[voice coil]] actuator or, in some older designs, a [[stepper motor]]. Early hard disk drives wrote data at some constant bits per second, resulting in all tracks having the same amount of data per track, but modern drives (since the 1990s) use [[zone bit recording]], increasing the write speed from inner to outer zone and thereby storing more data per track in the outer zones.

In modern drives, the small size of the magnetic regions creates the danger that their magnetic state might be lost because of [[superparamagnetism|thermal effects]]⁠ ⁠— thermally induced magnetic instability which is commonly known as the "[[superparamagnetic limit]]". To counter this, the platters are coated with two parallel magnetic layers, separated by a three-atom layer of the non-magnetic element [[ruthenium]], and the two layers are magnetized in opposite orientation, thus reinforcing each other.<ref name="AutoMK-15" /> Another technology used to overcome thermal effects to allow greater recording densities is [[perpendicular recording]] (PMR), first shipped in 2005,<ref name="AutoMK-16" /> and {{as of|2007|lc=on}}, used in certain HDDs.<ref name="AutoMK-17" /><ref name="AutoMK-18" /><ref name="AutoMK-19" /> Perpendicular recording may be accompanied by changes in the manufacturing of the read/write heads to increase the strength of the magnetic field created by the heads.<ref>{{cite web | url=https://patents.google.com/patent/US20080002290 | title=Damascene coil design for a perpendicular magnetic recording head }}</ref>

In 2004, a higher-density recording media was introduced, consisting of coupled soft and hard magnetic layers. So-called ''[[exchange spring media]]'' magnetic storage technology, also known as ''exchange coupled composite media'', allows good writability due to the write-assist nature of the soft layer. However, the thermal stability is determined only by the hardest layer and not influenced by the soft layer.<ref name=AutoMK-19a /><ref name=AutoMK-19b />

Flux control MAMR (FC-MAMR) allows a hard drive to have increased recording capacity without the need for new hard disk drive platter materials. MAMR hard drives have a microwave-generating spin torque generator (STO) on the read/write heads which allows physically smaller bits to be recorded to the platters, increasing areal density. Normally hard drive recording heads have a pole called a main pole that is used for writing to the platters, and adjacent to this pole is an air gap and a shield. The write coil of the head surrounds the pole. The STO device is placed in the air gap between the pole and the shield to increase the strength of the magnetic field created by the pole; FC-MAMR technically doesn't use microwaves but uses technology employed in MAMR. The STO has a Field Generation Layer (FGL) and a Spin Injection Layer (SIL), and the FGL produces a magnetic field using spin-polarised electrons originating in the SIL, which is a form of spin torque energy.<ref>{{cite web | url=https://blocksandfiles.com/2021/06/14/toshiba-disks-get-2-gen-leg-up-from-flux-control/ | title=Toshiba disks get 2-gen leg-up from flux control | date=June 14, 2021 }}</ref>

=== Components ===
[[File:Hard disk dismantled.jpg|thumb|left|An HDD with disks and motor hub removed, exposing copper-colored stator coils surrounding a bearing in the center of the spindle motor. The orange stripe along the side of the arm is a thin printed-circuit cable, the spindle bearing is in the center and the actuator is in the upper left.]]
[[File:Circuit board of a Samsung hard disk MP0402H.jpg|thumb|right|Circuit board of a 2.5-inch Samsung hard disk MP0402H]]

A typical HDD has two electric motors: a spindle motor that spins the disks and an actuator (motor) that positions the read/write head assembly across the spinning disks. The disk motor has an external rotor attached to the disks; the stator windings are fixed in place. Opposite the actuator at the end of the head support arm is the read-write head; thin printed-circuit cables connect the read-write heads to [[amplifier]] electronics mounted at the pivot of the actuator. The head support arm is very light, but also stiff; in modern drives, acceleration at the head reaches 550 [[g-force|''g'']].

[[File:Kopftraeger WD2500JS-00MHB0.jpg|thumb|Head stack with an actuator coil on the left and read/write heads on the right]]
[[File:HDD read-write head.jpg|thumb|Close-up of a single [[disk read-and-write head|read–write head]], showing the side facing the platter]]

The ''{{visible anchor|actuator|Actuator}}'' is a [[permanent magnet]] and [[moving coil]] motor that swings the heads to the desired position. A metal plate supports a squat [[neodymium magnet|neodymium–iron–boron]] (NIB) high-flux [[magnet]]. Beneath this plate is the moving coil, often referred to as the ''[[voice coil]]'' by analogy to the coil in [[loudspeaker]]s, which is attached to the actuator hub, and beneath that is a second NIB magnet, mounted on the bottom plate of the motor (some drives have only one magnet).

The voice coil itself is shaped rather like an arrowhead and is made of doubly coated copper [[magnet wire]]. The inner layer is insulation, and the outer is thermoplastic, which bonds the coil together after it is wound on a form, making it self-supporting. The portions of the coil along the two sides of the arrowhead (which point to the center of the actuator bearing) then interact with the [[magnetic field]] of the fixed magnet. Current flowing radially outward along one side of the arrowhead and radially inward on the other produces the [[magnetic field#Force on a charged particle|tangential force]]. If the magnetic field were uniform, each side would generate opposing forces that would cancel each other out. Therefore, the surface of the magnet is half north pole and half south pole, with the radial dividing line in the middle, causing the two sides of the coil to see opposite magnetic fields and produce forces that add instead of canceling. Currents along the top and bottom of the coil produce radial forces that do not rotate the head.

The HDD's electronics controls the movement of the actuator and the rotation of the disk and transfers data to or from a [[disk controller]]. Feedback of the drive electronics is accomplished by means of special segments of the disk dedicated to [[servomotor|servo]] feedback. These are either complete concentric circles (in the case of dedicated servo technology) or segments interspersed with real data (in the case of embedded servo, otherwise known as sector servo technology). The servo feedback optimizes the signal-to-noise ratio of the GMR sensors by adjusting the voice coil motor to rotate the arm. A more modern servo system also employs milli or micro actuators to more accurately position the read/write heads.<ref>A. Al-Mamun, G. Guo, C. Bi, Hard Disk Drive: Mechatronics and Control, 2006, Taylor & Francis.</ref> The spinning of the disks uses fluid-bearing spindle motors. Modern disk firmware is capable of scheduling reads and writes efficiently on the platter surfaces and remapping sectors of the media that have failed.

=== {{anchor|ERRORRATESHANDLING}}Error rates and handling ===
Modern drives make extensive use of [[error correction code]]s (ECCs), particularly [[Reed–Solomon error correction]]. These techniques store extra bits, determined by mathematical formulas, for each block of data; the extra bits allow many errors to be corrected invisibly. The extra bits themselves take up space on the HDD, but allow higher recording densities to be employed without causing uncorrectable errors, resulting in much larger storage capacity.<ref>{{cite web |url=https://www.karlstechnology.com/blog/hard-drive-error-correcting-code-ecc/ |title=Hard Drive Error Correcting Code (ECC) |publisher=The PC Guide |last=Kozierok |first=Charles |access-date=May 26, 2019 |date=November 25, 2018 |archive-url=https://web.archive.org/web/20190526072642/https://www.karlstechnology.com/blog/hard-drive-error-correcting-code-ecc/ |archive-date=May 26, 2019 |url-status=live }}</ref> For example, a typical 1&nbsp;[[Terabyte|TB]] hard disk with 512-byte sectors provides additional capacity of about 93&nbsp;[[Gibibyte|GB]] for the [[forward error correction|ECC]] data.<ref>{{cite web
 | url = https://www.idema.org/wp-content/uploads/downloads/2011/12/AF-in-Legacy-Infrastructures-SDC2011_IDEMA-AF.pdf
 | title = Advanced Format in Legacy Infrastructures: More Transparent than Disruptive
 | year = 2011
 | access-date = November 5, 2013
 | first = Curtis E.
 | last = Stevens
 | website = idema.org
 | archive-url = https://web.archive.org/web/20131105222506/http://www.idema.org/wp-content/uploads/downloads/2011/12/AF-in-Legacy-Infrastructures-SDC2011_IDEMA-AF.pdf
 | archive-date = November 5, 2013
 | url-status = dead
 }}</ref>

In the newest drives, {{as of|2009|lc=on}},<ref name="AutoMK-21" /> [[low-density parity-check code]]s (LDPC) were supplanting Reed–Solomon; LDPC codes enable performance close to the [[Shannon limit]] and thus provide the highest storage density available.<ref name="AutoMK-21" /><ref>[https://docplayer.net/3699743-2-5-inch-hard-disk-drive-with-high-recording-density-and-high-shock-resistance.html "2.5-inch Hard Disk Drive with High Recording Density and High Shock Resistance] {{Webarchive|url=https://web.archive.org/web/20190526075908/https://docplayer.net/3699743-2-5-inch-hard-disk-drive-with-high-recording-density-and-high-shock-resistance.html |date=May 26, 2019 }}, Toshiba, 2011</ref>

Typical hard disk drives attempt to "remap" the data in a [[Bad sector|physical sector that is failing]] to a spare physical sector provided by the drive's "spare sector pool" (also called "reserve pool"),<ref>{{cite web|author=MjM Data Recovery Ltd |url=http://datarecovery.mjm.co.uk/sectorremapping.html |archive-url=https://web.archive.org/web/20140201174433/http://datarecovery.mjm.co.uk/sectorremapping.html| title=MJM Data Recovery Ltd: Hard Disk Bad Sector Mapping Techniques |website=Datarecovery.mjm.co.uk |access-date=January 21, 2014 | archive-date=February 1, 2014}}</ref> while relying on the ECC to recover stored data while the number of errors in a bad sector is still low enough. The S.M.A.R.T ([[Self-Monitoring, Analysis and Reporting Technology]]) feature counts the total number of errors in the entire HDD fixed by ECC (although not on all hard drives as the related S.M.A.R.T attributes "Hardware ECC Recovered" and "Soft ECC Correction" are not consistently supported), and the total number of performed sector remappings, as the occurrence of many such errors may predict an [[HDD failure]].

The "No-ID Format", developed by IBM in the mid-1990s, contains information about which sectors are bad and where remapped sectors have been located.<ref>{{cite web |url=https://www.karlstechnology.com/blog/hard-drive-sector-format-and-structure/ |title=Hard Drive Sector Format and Structure |publisher=The PC Guide |last=Kozierok |first=Charles |access-date=May 26, 2019 |date=December 23, 2018 |archive-url=https://web.archive.org/web/20190526072622/https://www.karlstechnology.com/blog/hard-drive-sector-format-and-structure/ |archive-date=May 26, 2019 |url-status=live }}</ref>

Only a tiny fraction of the detected errors end up as not correctable. Examples of specified uncorrected bit read error rates include:
* 2013 specifications for enterprise SAS disk drives state the error rate to be one uncorrected bit read error in every 10<sup>16</sup> bits read,<ref name="SGAT2013">{{cite web
 | url = https://www.seagate.com/files/www-content/product-content/savvio-fam/enterprise-performance-15k-hdd/savvio-15k-4/en-us/enterprise-performance-15k-hdd-ds1797-1-1307us.pdf
 | title = Enterprise Performance 15K HDD: Data Sheet
 | year = 2013
 | access-date = October 24, 2013
 | publisher = Seagate
 | archive-url = https://web.archive.org/web/20131029192706/http://www.seagate.com/files/www-content/product-content/savvio-fam/enterprise-performance-15k-hdd/savvio-15k-4/en-us/enterprise-performance-15k-hdd-ds1797-1-1307us.pdf
 | archive-date = October 29, 2013
 | url-status = live
 }}</ref><ref name="WD2013">{{cite web
 | url = https://www.wdc.com/wdproducts/library/SpecSheet/ENG/2879-771463.pdf
 | title = WD Xe: Datacenter hard drives
 | year = 2013
 | access-date = October 24, 2013
 | publisher = Western Digital
 | archive-url = https://web.archive.org/web/20131029193439/http://www.wdc.com/wdproducts/library/SpecSheet/ENG/2879-771463.pdf
 | archive-date = October 29, 2013
 | url-status = live
 }}</ref> 
* 2018 specifications for consumer SATA hard drives state the error rate to be one uncorrected bit read error in every 10<sup>14</sup> bits.<ref name="SGAT2018">{{cite web
 | title = 3.5" BarraCuda data sheet
 | publisher = Seagate
 | date = June 2018
 | url = https://www.seagate.com/www-content/datasheets/pdfs/3-5-barracudaDS1900-11-1806US-en_US.pdf
 | access-date = July 28, 2018
 | archive-url = https://web.archive.org/web/20180728070449/https://www.seagate.com/www-content/datasheets/pdfs/3-5-barracudaDS1900-11-1806US-en_US.pdf
 | archive-date = July 28, 2018
 | url-status = live
 }}</ref><ref name="WD2018">{{cite web
 | title = WD Red Desktop/Mobile Series Spec Sheet
 | publisher = Western Digital
 | date = April 2018
 | url = https://www.wdc.com/content/dam/wdc/website/downloadable_assets/eng/spec_data_sheet/2879-800002.pdf
 | access-date = July 28, 2018
 | archive-url = https://web.archive.org/web/20180728070730/https://www.wdc.com/content/dam/wdc/website/downloadable_assets/eng/spec_data_sheet/2879-800002.pdf
 | archive-date = July 28, 2018
 | url-status = live
 }}</ref>
Within a given manufacturers model the uncorrected bit error rate is typically the same regardless of capacity of the drive.<ref name="SGAT2013" /><ref name="WD2013" /><ref name="SGAT2018" /><ref name="WD2018" />

The worst type of errors are [[silent data corruption]]s which are errors undetected by the disk firmware or the host operating system; some of these errors may be caused by hard disk drive malfunctions while others originate elsewhere in the connection between the drive and the host.<ref>{{Cite news |title= Keeping Bits Safe: How Hard Can It Be? |work= ACM Queue |date= October 1, 2010 |author= David S. H. Rosenthal |url= https://queue.acm.org/detail.cfm?id=1866298 |access-date= January 2, 2014 |author-link= David S. H. Rosenthal |archive-url= https://web.archive.org/web/20131217020947/http://queue.acm.org/detail.cfm?id=1866298 |archive-date= December 17, 2013 |url-status= live }}</ref>

{{Anchor|TDMR}}

=== Development ===
[[File:Full History Disk Areal Density Trend.png|thumb|Leading-edge hard disk drive [[Density (computer storage)|areal densities]] from 1956 through 2009 compared to [[Moore's law]]. By 2016, progress had slowed significantly below the extrapolated density trend.<ref name="Hayes 2016">{{cite web |url=http://bit-player.org/2016/wheres-my-petabyte-disk-drive |title= Where's My Petabyte Disk Drive? |first= Brian |last= Hayes|page= chart of historical data courtesy of Edward Grochowski  |date= March 27, 2016 |access-date= December 1, 2019 }}</ref>]]

The rate of areal density advancement was similar to [[Moore's law]] (doubling every two years) through 2010: 60% per year during 1988–1996, 100% during 1996–2003 and 30% during 2003–2010.<ref name="Byrne2015b">{{cite web |url=http://www.federalreserve.gov/econresdata/notes/feds-notes/2015/prices-for-data-storage-equipment-and-the-state-of-it-innovation-20150701.html#fn2 |title=Prices for Data Storage Equipment and the State of IT Innovation |publisher=The Federal Reserve Board FEDS Notes |first=David |last=Byrne |page=Table 2 |date=July 1, 2015 |access-date=July 5, 2015 |archive-url=https://web.archive.org/web/20150708124555/http://www.federalreserve.gov/econresdata/notes/feds-notes/2015/prices-for-data-storage-equipment-and-the-state-of-it-innovation-20150701.html#fn2 |archive-date=July 8, 2015 |url-status=live }}</ref> Speaking in 1997, [[Gordon Moore]] called the increase "flabbergasting",<ref name="Moore1997">{{cite news
 | url = https://www.pcmag.com/article2/0,2817,1172800,00.asp
 | title = Gallium Arsenide
 | work = PC Magazine
 | date = March 25, 1997
 | quote = Gordon Moore: ... the ability of the magnetic disk people to continue to increase the density is flabbergasting--that has moved at least as fast as the semiconductor complexity.
 | access-date = August 16, 2014
 | archive-url = https://web.archive.org/web/20140821103700/http://www.pcmag.com/article2/0,2817,1172800,00.asp
 | archive-date = August 21, 2014
 | url-status = live
 }}</ref> while observing later that growth cannot continue forever.<ref name="Moore2005">{{cite news |last=Dubash |first=Manek |url=http://www.techworld.com/news/operating-systems/moores-law-is-dead-says-gordon-moore-3576581/ |title=Moore's Law is dead, says Gordon Moore |work=techworld.com |date=April 13, 2005 |url-status=dead |archive-url=https://web.archive.org/web/20140706081110/http://news.techworld.com/operating-systems/3477/moores-law-is-dead-says-gordon-moore/ |archive-date=July 6, 2014 |access-date=March 18, 2022 |quote=It can't continue forever. The nature of exponentials is that you push them out and eventually disaster happens. }}</ref> Price improvement decelerated to −12% per year during 2010–2017,<ref name="McCallum">{{cite web |title=Disk Drive Prices (1955–2017) |url=http://www.jcmit.net/diskprice.htm |year=2017 |first=John C. |last=McCallum |access-date=July 15, 2017 |archive-url=https://web.archive.org/web/20170711202719/http://www.jcmit.net/diskprice.htm |archive-date=July 11, 2017 |url-status=live }}</ref> as the growth of areal density slowed. The rate of advancement for areal density slowed to 10% per year during 2010–2016,<ref name="IBM Fontana 2016">{{cite web |url=http://www.ibmsystemsmag.com/mainframe/storage/Support/cloud-trends-projections/?page=1 |title=A Look at Cloud Storage Component Technologies Trends and Future Projections |first1=Gary M. |last1=Decad |author2=Robert E. Fontana Jr. |website=ibmsystemsmag.com |date=July 6, 2017 |access-date=July 21, 2014 |page=Table 1 |archive-url=https://web.archive.org/web/20170729050933/http://ibmsystemsmag.com/mainframe/storage/support/cloud-trends-projections/?page=1 |archive-date=July 29, 2017 |url-status=dead }}</ref> and there was difficulty in migrating from perpendicular recording to newer technologies.<ref name="Mellor 2014-11-10">{{cite news |last=Mellor |first=Chris |url=https://www.theregister.co.uk/2014/11/10/kryders_law_of_ever_cheaper_storage_disproven/ |title=Kryder's law craps out: Race to UBER-CHEAP STORAGE is OVER |work=theregister.co.uk |location=UK |publisher=The Register |date=November 10, 2014 |access-date=November 12, 2014 |quote=The 2011 Thai floods almost doubled disk capacity cost/GB for a while. Rosenthal writes: 'The technical difficulties of migrating from PMR to HAMR, meant that already in 2010 the Kryder rate had slowed significantly and was not expected to return to its trend in the near future. The floods reinforced this.' |archive-url=https://web.archive.org/web/20141112004831/http://www.theregister.co.uk/2014/11/10/kryders_law_of_ever_cheaper_storage_disproven/ |archive-date=November 12, 2014 |url-status=live }}</ref>

As bit cell size decreases, more data can be put onto a single drive platter. In 2013, a production desktop 3&nbsp;TB HDD (with four platters) would have had an areal density of about 500&nbsp;Gbit/in<sup>2</sup> which would have amounted to a bit cell comprising about 18 magnetic grains (11 by 1.6 grains).<ref name="Anderson2013a" >{{cite web | url = https://www.dtc.umn.edu/resources/bd2013_anderson.pdf | title = HDD Opportunities & Challenges, Now to 2020 | year = 2013 | access-date = May 23, 2014 | quote = 'PMR CAGR slowing from historical 40+% down to ~8-12%' and 'HAMR CAGR = 20-40% for 2015–2020' | first = Dave | last = Anderson | publisher = Seagate | archive-url = https://web.archive.org/web/20140525232605/https://www.dtc.umn.edu/resources/bd2013_anderson.pdf | archive-date = May 25, 2014 | url-status = live }}</ref> Since the mid-2000s, areal density progress has been challenged by a [[Superparamagnetism#Effect on hard drives|superparamagnetic]] trilemma involving grain size, grain magnetic strength and ability of the head to write.<ref>{{cite journal
| title = New Paradigms in Magnetic Recording
| last = Plumer |display-authors=etal 
| first = Martin L.
| journal = Physics in Canada
| volume = 67
| issue = 1
| date = March 2011
| pages = 25–29
|arxiv = 1201.5543| bibcode = 2012arXiv1201.5543P }}</ref> In order to maintain acceptable signal-to-noise, smaller grains are required; smaller grains may self-reverse ([[electrothermal instability]]) unless their magnetic strength is increased, but known write head materials are unable to generate a strong enough magnetic field sufficient to write the medium in the increasingly smaller space taken by grains.

Magnetic storage technologies are being developed to address this trilemma, and compete with [[flash memory]]–based [[solid-state drive]]s (SSDs). In 2013, [[Seagate Technology|Seagate]] introduced [[shingled magnetic recording]] (SMR),<ref name="Seagate 2013-09-09">{{cite press release |url=http://www.seagate.com/about/newsroom/press-releases/shingled-magnetic-recording-milestone-pr-master/ |archive-url=https://web.archive.org/web/20141009135540/http://www.seagate.com/about/newsroom/press-releases/shingled-magnetic-recording-milestone-pr-master/ |title=Seagate Delivers On Technology Milestone: First to Ship Hard Drives Using Next-Generation Shingled Magnetic Recording |location=New York |publisher=[[Seagate Technology]] plc |date=September 9, 2013 |access-date=July 5, 2014 |archive-date=October 9, 2014 |quote=Shingled Magnetic Technology is the First Step to Reaching a 20 Terabyte Hard Drive by 2020}}</ref> intended as something of a "stopgap" technology between PMR and Seagate's intended successor [[heat-assisted magnetic recording]] (HAMR). SMR utilizes overlapping tracks for increased data density, at the cost of design complexity and lower data access speeds (particularly write speeds and [[random access]] 4k speeds).<ref>{{cite web
 | url = https://lwn.net/Articles/591782/
 | title = Support for shingled magnetic recording devices
 | date = March 26, 2014
 | access-date = January 7, 2015
 | first = Jake
 | last = Edge
 | publisher = [[LWN.net]]
 | archive-url = https://web.archive.org/web/20150202075938/http://lwn.net/Articles/591782/
 | archive-date = February 2, 2015
 | url-status = live
 }}</ref><ref>{{cite web
 | url = https://lwn.net/Articles/548116/
 | title = LSFMM: A storage technology update
 | date = April 23, 2013
 | access-date = January 7, 2015
 | first = Jonathan
 | last = Corbet
 | publisher = [[LWN.net]]
 | quote = A 'shingled magnetic recording' (SMR) drive is a rotating drive that packs its tracks so closely that one track cannot be overwritten without destroying the neighboring tracks as well. The result is that overwriting data requires rewriting the entire set of closely-spaced tracks; that is an expensive tradeoff, but the benefit—much higher storage density—is deemed to be worth the cost in some situations.
 | archive-url = https://web.archive.org/web/20150107075254/https://lwn.net/Articles/548116/
 | archive-date = January 7, 2015
 | url-status = live
 }}</ref>

By contrast, [[HGST]] (now part of [[Western Digital]]) focused on developing ways to seal [[helium]]-filled drives instead of the usual filtered air. Since [[turbulence]] and [[friction]] are reduced, higher areal densities can be achieved due to using a smaller track width, and the energy dissipated due to friction is lower as well, resulting in a lower power draw. Furthermore, more platters can be fit into the same enclosure space, although helium gas is notoriously difficult to prevent escaping.<ref>{{Cite web|date=2020|title=Brochure: HelioSeal Technology: Beyond Air. Helium Takes You Higher.|url=https://documents.westerndigital.com/content/dam/doc-library/en_us/assets/public/western-digital/collateral/brochure/brochure-helioseal-technology.pdf|website=Western Digital}}</ref> Thus, helium drives are completely sealed and do not have a breather port, unlike their air-filled counterparts.

Other recording technologies are either under research or have been commercially implemented to increase areal density, including Seagate's [[heat-assisted magnetic recording]] (HAMR). HAMR requires a different architecture with redesigned media and read/write heads, new lasers, and new near-field optical transducers.<ref name="Shilov_anandtech_b">{{cite news |url=http://www.anandtech.com/show/9866/hard-disk-drives-with-hamr-technology-set-to-arrive-in-2018 |title=Hard Disk Drives with HAMR Technology Set to Arrive in 2018 |first=Anton |last=Shilov |date=December 18, 2015 |access-date=January 2, 2016 |quote=Unfortunately, mass production of actual hard drives featuring HAMR has been delayed for a number of times already and now it turns out that the first HAMR-based HDDs are due in 2018. ... HAMR HDDs will feature a new architecture, require new media, completely redesigned read/write heads with a laser as well as a special near-field optical transducer (NFT) and a number of other components not used or mass produced today. |archive-url=https://web.archive.org/web/20160102200055/http://www.anandtech.com/show/9866/hard-disk-drives-with-hamr-technology-set-to-arrive-in-2018 |archive-date=January 2, 2016 |url-status=live }}</ref> HAMR is expected to ship commercially in late 2024,<!-- did it ship? --><ref>{{cite web |url= https://www.tomshardware.com/news/seagate-reveals-hamr-roadmap-32-tb-comes-first |title= Seagate Reveals HAMR HDD Roadmap: 32TB First, 40TB Follows |last= Shilov |first= Anton |date= June 8, 2023 |access-date= Oct 3, 2024 }}</ref> after technical issues delayed its introduction by more than a decade, from earlier projections as early as 2009.<ref name= "HAMR 2008 for 2009" >{{cite web |url= https://blog.dshr.org/2018/05/longer-talk-at-msst2018.html |title= Longer talk at MSST2018 |last= Rosenthal |first= David |date= May 16, 2018 |access-date= November 22, 2019 }}</ref><ref name= "HAMR 2014 for 2015" >{{cite web |url= https://www.kitguru.net/components/hard-drives/anton-shilov/tdk-hamr-technology-could-enable-15tb-hard-drives-already-in-2015/ |title= TDK: HAMR technology could enable 15TB HDDs already in 2015 |last= Shilov |first= Anton |date= October 15, 2014 |access-date= November 15, 2019 }}</ref><ref name= "HAMR 2013 for 2016" >{{cite web |url= http://www.tomsitpro.com/articles/wd-hamr-hdd-heat-assisted-magnetic-recording,1-1396.html |title= WD Demos Future HDD Storage Tech: 60TB Hard Drives |last= Oliver |first= Bill |work= Tom's IT Pro |quote= …Seagate expects to start selling HAMR drives in 2016. |date= November 18, 2013 |access-date= November 15, 2019 |archive-url= https://web.archive.org/web/20131121065015/http://www.tomsitpro.com/articles/wd-hamr-hdd-heat-assisted-magnetic-recording,1-1396.html |archive-date= November 21, 2013 }}</ref><ref name= "blocks HAMR 2019" >{{cite web |url= https://blocksandfiles.com/2019/08/28/nearline-disk-drives-ssd-attack/ |title= How long before SSDs replace nearline disk drives? |last= Mellor |first= Chris |quote= Seagate CTO Dr John Morris told analysts that Seagate has built 55,000 HAMR drives and aims to get disks ready for customer sampling by the end of 2020. |date= August 28, 2019 |access-date= November 15, 2019 }}</ref> HAMR's planned successor, [[bit-patterned recording]] (BPR),<ref name="AutoMK-22" /> has been removed from the roadmaps of Western Digital and Seagate.<ref name= "BPR roadmaps 2018" >{{cite web |url= https://blog.dshr.org/2018/05/longer-talk-at-msst2018.html |title= Longer talk at MSST2018 |last= Rosenthal |first= David |quote= The most recent Seagate roadmap pushes HAMR shipments into 2020, so they are now slipping faster than real-time. Western Digital has given up on HAMR and is promising that Microwave Assisted Magnetic Recording (MAMR) is only a year out. BPM has dropped off both companies' roadmaps. |date= May 16, 2018 |access-date= November 22, 2019 }}</ref> Western Digital's microwave-assisted magnetic recording (MAMR),<ref>{{cite journal |last=Mallary |display-authors=etal |first= Mike|date=July 2014 |title= Head and Media Challenges for 3 Tb/in<sup>2</sup> Microwave-Assisted Magnetic Recording|journal=IEEE Transactions on Magnetics |volume=50 |issue=7 |pages=1–8 |doi=10.1109/TMAG.2014.2305693|s2cid=22858444 |issn = 0018-9464 }}</ref><ref>{{cite journal|last1=Li|first1=Shaojing|last2=Livshitz|first2=Boris|last3=Bertram|first3=H. Neal|last4=Schabes|first4=Manfred|last5=Schrefl|first5=Thomas|last6=Fullerton|first6=Eric E.|last7=Lomakin|first7=Vitaliy|title=Microwave assisted magnetization reversal in composite media|journal=Applied Physics Letters|date=2009|volume=94|issue=20|page=202509|doi=10.1063/1.3133354|url=https://www.karlstechnology.com/hard-drives/JAP_2009_MAMR.pdf|bibcode=2009ApPhL..94t2509L|access-date=May 24, 2019|archive-url=https://web.archive.org/web/20190524225721/https://www.karlstechnology.com/hard-drives/JAP_2009_MAMR.pdf|archive-date=May 24, 2019|url-status=live}}</ref> also referred to as energy-assisted magnetic recording (EAMR), was sampled in 2020, with the first EAMR drive, the Ultrastar HC550, shipping in late 2020.<ref>{{Cite web|last=Shilov|first=Anton|title=Western Digital Reveals 18 TB DC HC550 'EAMR' Hard Drive|url=https://www.anandtech.com/show/14869/western-digital-announces-18-tb-eamr-hard-drive |date=September 18, 2019 |access-date=2021-10-11|website=AnandTech}}</ref><ref name="Blocks MAMR 2019">{{cite web |url= https://blocksandfiles.com/2019/09/03/western-digital-18tb-and-20tb-mamr-disk-drives/ |website=Blocks & Files |title= Western Digital debuts 18TB and 20TB MAMR disk drives |last= Mellor |first= Chris |quote= …microwave-assisted magnetic (MAMR) recording technology…sample shipments are due by the end of the year. |date= September 3, 2019 |access-date= November 23, 2019 }}</ref><ref>{{Cite web|last=Raevenlord |title=Western Digital Finally Launches Ultrastar DC HC550 18 TB Drives With EAMR for Enterprise|url=https://www.techpowerup.com/269562/western-digital-finally-launches-ultrastar-dc-hc550-18-tb-drives-with-eamr-for-enterprise|access-date=2021-10-11|website=TechPowerUp|date=July 8, 2020 |language=en}}</ref> [[Two-dimensional magnetic recording]] (TDMR)<ref name="Anderson2013a" /><ref name="Wood 2010">{{cite web |url=https://www.ewh.ieee.org/r6/scv/mag/MtgSum/Meeting2010_10_Presentation.pdf |title=Shingled Magnetic Recording and Two-Dimensional Magnetic Recording |last=Wood |first=Roger |work=[[IEEE]] |publisher=Hitachi GST |date=October 19, 2010 |access-date=August 4, 2014 |archive-url=https://web.archive.org/web/20140810161542/http://www.ewh.ieee.org/r6/scv/mag/MtgSum/Meeting2010_10_Presentation.pdf |archive-date=August 10, 2014 |url-status=live }}</ref> and "current perpendicular to plane" [[giant magnetoresistance]] (CPP/GMR) heads have appeared in research papers.<ref name="Coughlin" /><ref>{{Cite arXiv |title=All-Heusler giant-magnetoresistance junctions with matched energy bands and Fermi surfaces |eprint = 1301.6106|last1 = Bai|first1 = Zhaoqiang|last2 = Cai|first2 = Yongqing|last3 = Shen|first3 = Lei|last4 = Han|first4 = Guchang|last5 = Feng|first5 = Yuanping|year = 2013|class = cond-mat.mes-hall}}</ref><ref>{{cite web|url=http://www1.hgst.com/hdd/research/recording_head/pr/PerpendicularAnimation.html|title=Perpendicular Magnetic Recording Explained - Animation|date=December 21, 2001 |access-date=July 27, 2014|archive-url=https://web.archive.org/web/20181006062214/http://www1.hgst.com/hdd/research/recording_head/pr/PerpendicularAnimation.html|archive-date=October 6, 2018|url-status=live}}</ref>

Some drives have adopted dual independent actuator arms to increase read/write speeds and compete with SSDs.<ref>{{cite web |url=https://www.anandtech.com/show/13935/seagate-hdd-plans-2019 |title= State of the Union: Seagate's HAMR Hard Drives, Dual-Actuator Mach2, and 24 TB HDDs on Track |work=Anandtech.com |access-date=February 20, 2019 |archive-url=https://web.archive.org/web/20190220002907/https://www.anandtech.com/show/13935/seagate-hdd-plans-2019 |archive-date=February 20, 2019 |url-status=live }}</ref>
A 3D-actuated vacuum drive (3DHD) concept<ref name= "3DHD blog" >{{Cite web |url= https://blog.dshr.org/2019/09/promising-new-hard-disk-technology.html |title= Promising New Hard Disk Technology |access-date= December 1, 2019 }}</ref><!--3DHD is developed by L2 drive--> and 3D magnetic recording have been proposed.<ref>{{Cite web|url=https://www.extremetech.com/extreme/168619-3d-magnetic-storage-breakthrough-enables-100tb-hard-drives|title=3D magnetic storage breakthrough enables 100TB+ hard drives &#124; Extremetech|date=October 15, 2013 }}</ref>

Depending upon assumptions on feasibility and timing of these technologies, Seagate forecasts that areal density will grow 20% per year during 2020–2034.<ref name= "blocks WWrevenue August2019"/>

== Capacity ==
[[File:Seagate Barracuda IMG 7901.jpg|thumb|Two [[Seagate Barracuda]] drives from 2003 and 2009, respectively 160&nbsp;GB and 1&nbsp;TB. {{As of|2025}}, Seagate offers capacities up to 36&nbsp;TB.]]
[[File:MSATA SSD vs. 2.5" SATA drive.JPG|thumb|mSATA SSD on top of a 2.5-inch hard drive]]
The highest-capacity HDDs shipping commercially {{as of|2025|lc=on}} are 36&nbsp;TB.<ref name="x-mozaic">{{cite web | url=https://www.seagate.com/products/enterprise-drives/exos-x/x-mozaic/ | title=Exos Mozaic 3+ |website= Seagate US |access-date=2025-02-17}}</ref>{{fv|reason=Shows that 36 TB drives exist but doesn't establish this is the highest available capacity|date=January 2025}}  The capacity of a hard disk drive, as reported by an operating system to the end user, is smaller than the amount stated by the manufacturer for several reasons, e.g. the operating system using some space, use of some space for data redundancy, space use for file system structures. Confusion of [[decimal prefix]]es and [[binary prefix]]es can also lead to errors.

=== Calculation ===
Modern hard disk drives appear to their host controller as a contiguous set of logical blocks, and the gross drive capacity is calculated by multiplying the number of blocks by the block size. This information is available from the manufacturer's product specification, and from the drive itself through use of operating system functions that invoke low-level drive commands.<ref name="SAS" /><ref name="SATA" /> Older IBM and compatible drives, e.g. [[IBM 3390]] using the [[Count key data|CKD]] record format, have variable length records; such drive capacity calculations must take into account the characteristics of the records. Some newer DASD simulate CKD, and the same capacity formulae apply.

The gross capacity of older sector-oriented HDDs is calculated as the product of the number of [[cylinder (disk drive)|cylinders]] per recording zone, the number of bytes per sector (most commonly 512), and the count of [[Zone bit recording|zones]] of the drive.{{citation needed|date=April 2016}} Some modern SATA drives also report [[cylinder-head-sector]] (CHS) capacities, but these are not physical parameters because the reported values are constrained by historic operating system interfaces. The C/H/S scheme has been replaced by [[logical block addressing]] (LBA), a simple linear addressing scheme that locates blocks by an integer index, which starts at LBA 0 for the first block and increments thereafter.<ref>{{cite web | url = http://www.idema.org/wp-content/plugins/download-monitor/download.php?id=1223 | title = LBA Count for Disk Drives Standard (Document LBA1-03) | date = June 15, 2009 | access-date = February 14, 2016 | publisher = [[IDEMA]] | format = PDF | archive-url = https://web.archive.org/web/20160222082703/http://www.idema.org/wp-content/plugins/download-monitor/download.php?id=1223 | archive-date = February 22, 2016 | url-status = live }}</ref> When using the C/H/S method to describe modern large drives, the number of heads is often set to 64, although a typical modern hard disk drive has between one and four platters. In modern HDDs, spare capacity for [[Internal hard-drive defect management|defect management]] is not included in the published capacity; however, in many early HDDs, a certain number of sectors were reserved as spares, thereby reducing the capacity available to the operating system. Furthermore, many HDDs store their firmware in a reserved service zone, which is typically not accessible by the user, and is not included in the capacity calculation.

For [[RAID]] subsystems, [[data integrity]] and fault-tolerance requirements also reduce the realized capacity. For example, a RAID&nbsp;1 array has about half the total capacity as a result of data mirroring, while a RAID&nbsp;5 array with {{mvar|n}} drives loses {{mvar|1/n}} of capacity (which equals to the capacity of a single drive) due to storing parity information. RAID subsystems are multiple drives that appear to be one drive or more drives to the user, but provide fault tolerance. Most RAID vendors use [[checksum]]s to improve data integrity at the block level. Some vendors design systems using HDDs with sectors of 520 bytes to contain 512 bytes of user data and eight checksum bytes, or by using separate 512-byte sectors for the checksum data.<ref name="AutoMK-38" />

Some systems may use [[Host protected area|hidden]] [[disk partitioning|partitions]] for system recovery, reducing the capacity available to the end user without knowledge of [[List of disk partitioning software|special disk partitioning utilities]] like [[diskpart]] in [[Windows]].<ref>{{Cite journal |last1=Gupta |first1=Mayank R. |last2=Hoeschele |first2=Michael D. |last3=Rogers |first3=Marcus K. |date=2006 |title=Hidden Disk Areas: HPA and DCO |url=https://www.utica.edu/academic/institutes/ecii/publications/articles/EFE36584-D13F-2962-67BEB146864A2671.pdf |journal=International Journal of Digital Evidence |volume=5 |issue=1}}</ref>

=== Formatting ===
{{Main|Disk formatting}}
Data is stored on a hard drive in a series of logical blocks.  Each block is delimited by markers identifying its start and end, error detecting and correcting information, and space between blocks to allow for minor timing variations.  These blocks often contained 512 bytes of usable data, but other sizes have been used.  As drive density increased, an initiative known as [[Advanced Format]] extended the block size to 4096 bytes of usable data, with a resulting significant reduction in the amount of disk space used for block headers, error-checking data, and spacing.

The process of initializing these logical blocks on the physical disk platters is called ''low-level formatting'', which is usually performed at the factory and is not normally changed in the field.<ref>{{cite web|url=http://www.pcguide.com/ref/hdd/geom/formatLow-c.html|title=Low-Level Formatting|access-date=June 28, 2010|archive-url=https://web.archive.org/web/20170604064402/http://www.pcguide.com/ref/hdd/geom/formatLow-c.html|archive-date=June 4, 2017|url-status=dead}}</ref> ''High-level formatting'' writes data structures used by the operating system to organize data files on the disk. This includes writing [[Disk partitioning|partition]] and [[file system]] structures into selected logical blocks. For example, some of the disk space will be used to hold a directory of disk file names and a list of logical blocks associated with a particular file.

Examples of partition mapping scheme include [[master boot record]] (MBR) and [[GUID Partition Table]] (GPT). Examples of data structures stored on disk to retrieve files include the [[File Allocation Table]] (FAT) in the [[DOS]] file system and [[inode]]s in many [[UNIX]] file systems, as well as other operating system data structures (also known as [[metadata]]). As a consequence, not all the space on an HDD is available for user files, but this system overhead is usually small compared with user data.

=== Units ===
{{See also|Binary prefix#Disk drives|l1=Binary prefix § disk drives}}

{| class="wikitable floatcenter" style="width: 60%; margin-left: 1.5em;"
|+ Decimal and binary [[unit prefix]]es interpretation<ref name="SeagateStorageGuide" /><ref name="WD2" />
! colspan=2 rowspan=2 | Capacity advertised by manufacturers{{Efn|name="units-decimal"|Expressed using [[SI prefix#List of SI prefixes|decimal&nbsp;multiples]]}}
! colspan=2 rowspan=2 | Capacity expected by some consumers{{Efn|name="units-binary"|Expressed using [[binary prefix|binary&nbsp;multiples]]}}
! colspan=2 | Reported capacity
|-
! rowspan=2 | [[Windows]]{{Efn|name="units-binary"}}
! rowspan=2 | [[macOS]] ver 10.6+{{Efn|name="units-decimal"}}
|-
! With prefix
! Bytes
! Bytes
! Diff.
|-
| style="text-align:right;"| 100&nbsp;[[Gigabyte|GB]]
| style="text-align:right;"| 100,000,000,000
| style="text-align:right;"| 107,374,182,400
| style="text-align:right;"| 7.37%
| style="text-align:right;"| 93.1&nbsp;GB
| style="text-align:right;"| 100&nbsp;GB
|-
| style="text-align:right;"| 1&nbsp;[[Terabyte|TB]]
| style="text-align:right;"| 1,000,000,000,000
| style="text-align:right;"| 1,099,511,627,776
| style="text-align:right;"| 9.95%
| style="text-align:right;"| 931&nbsp;GB
| style="text-align:right;"| 1,000&nbsp;GB, 1,000,000&nbsp;MB
|}

In the early days of computing, the total capacity of HDDs was specified in seven to nine decimal digits frequently truncated with the idiom ''millions''.<ref name="AutoMK-26" /><ref name="1301Ref" />
By the 1970s, the total capacity of HDDs was given by manufacturers using [[SI]] decimal prefixes such as [[megabyte]]s (1&nbsp;MB&nbsp;= 1,000,000&nbsp;bytes), [[gigabyte]]s (1&nbsp;GB&nbsp;= 1,000,000,000&nbsp;bytes) and [[terabyte]]s (1&nbsp;TB&nbsp;= 1,000,000,000,000 bytes).<ref name="SeagateStorageGuide" /><ref name="AutoMK-27" /><ref name="AutoMK-28" /><ref name="AutoMK-29" /> However, capacities of [[main memory|memory]] are usually quoted using a [[binary prefixes|binary interpretation]] of the prefixes, i.e. using powers of 1024 instead of 1000.

Software reports hard disk drive or memory capacity in different forms using either decimal or binary prefixes. The [[Microsoft Windows]] family of operating systems uses the binary convention when reporting storage capacity, so an HDD offered by its manufacturer as a 1&nbsp;TB drive is reported by these operating systems as a 931&nbsp;GB HDD. [[Mac OS X]] 10.6 ("[[Mac OS X Snow Leopard|Snow Leopard]]") uses decimal convention when reporting HDD capacity.<ref name="Apple" /> The default behavior of the {{Mono|df}} [[command-line utility]] on Linux is to report the HDD capacity as a number of 1024-byte units.<ref>{{cite web
 | url = http://linux.die.net/man/1/df
 | title = df(1) – Linux man page
 | access-date = July 18, 2015
 | website = linux.die.net
 | archive-url = https://web.archive.org/web/20150718060713/http://linux.die.net/man/1/df
 | archive-date = July 18, 2015
 | url-status = live
 }}</ref>

The difference between the decimal and binary prefix interpretation caused some consumer confusion and led to class action suits [[Binary prefix#Legal disputes|against HDD manufacturers]]. The plaintiffs argued that the use of decimal prefixes effectively misled consumers, while the defendants denied any wrongdoing or liability, asserting that their marketing and advertising complied in all respects with the law and that no class member sustained any damages or injuries.<ref name="WDSettle" /><ref name="AutoMK-30" /><ref name="AutoMK-31" />  In 2020, a California court ruled that use of the decimal prefixes with a decimal meaning was not misleading.<ref>{{Cite web|url=https://www.courthousenews.com/wp-content/uploads/2020/01/flashdrives.pdf|title=Order granted motion to dismiss amended complaint without leave to amend, 22 January 2020}}</ref>

{{Clear}}

{{Anchor|FORM-FACTORS}}

== Form factors ==
{{main|List of disk drive form factors}}
[[File:SixHardDriveFormFactors.jpg|thumb|right|8-, 5.25-, 3.5-, 2.5-, 1.8- and 1-inch HDDs, together with a ruler to show the size of platters and read-write heads]]
[[File:5.25 inch MFM hard disk drive.JPG|thumb|A newer 2.5-inch (63.5&nbsp;mm) 6,495&nbsp;MB HDD compared to an older 5.25-inch full-height 110&nbsp;MB HDD]]

IBM's first hard disk drive, the [[IBM&nbsp;350]], used a stack of fifty 24-inch platters, stored 3.75 MB of data (approximately the size of one modern digital picture), and was of a size comparable to two large refrigerators. In 1962, [[IBM]] introduced its [[IBM 1311|model 1311]] disk, which used six 14-inch (nominal size) platters in a removable pack and was roughly the size of a washing machine. This became a standard platter size for many years, used also by other manufacturers.<ref name="AutoMK-39" /> The [[IBM 2314]] used platters of the same size in an eleven-high pack and introduced the "drive in a drawer" layout, sometimes called the "pizza oven", although the "drawer" was not the complete drive. Into the 1970s, HDDs were offered in standalone cabinets of varying dimensions containing from one to four HDDs.

Beginning in the late 1960s, drives were offered that fit entirely into a chassis that would mount in a [[19-inch rack]]. Digital's [[RK05]] and RL01 were early examples using single 14-inch platters in removable packs, the entire drive fitting in a 10.5-inch-high rack space (six rack units). In the mid-to-late 1980s, the similarly sized [[Fujitsu Eagle]], which used (coincidentally) 10.5-inch platters, was a popular product.

With increasing sales of microcomputers having built-in [[floppy disk|floppy-disk drives (FDDs)]], HDDs that would fit to the FDD mountings became desirable. Starting with the [[Shugart Associates|Shugart Associates SA1000]], HDD ''form factors'' initially followed those of 8-inch, 5¼-inch, and 3½-inch floppy disk drives. Although referred to by these nominal sizes, the actual sizes for those three drives respectively are 9.5", 5.75" and 4" wide. Because there were no smaller floppy disk drives, smaller HDD form factors such as 2½-inch drives (actually 2.75" wide) developed from product offerings or industry standards.

{{As of|2025}}, 2½-inch and 3½-inch hard disks are the most popular sizes.  By 2009, all manufacturers had discontinued the development of new products for the 1.3-inch, 1-inch and 0.85-inch form factors due to falling prices of [[flash memory]],<ref name="AutoMK-51" /><ref name="AutoMK-52" /> which has no moving parts. While nominal sizes are in inches, actual dimensions are specified in millimeters.

== Performance characteristics ==
{{Main|Hard disk drive performance characteristics}}
The factors that limit the [[access time|time to access the data]] on an HDD are mostly related to the mechanical nature of the rotating disks and moving heads, including:
* [[Hard disk drive performance characteristics#Seek time|Seek time]] is a measure of how long it takes the head assembly to travel to the track of the disk that contains data. 
* Rotational latency is incurred because the desired [[disk sector]] may not be directly under the head when data transfer is requested. Average rotational latency is shown in the table, based on the statistical relation that the average latency is one-half the rotational period.
* The [[bit rate]] or data transfer rate (once the head is in the right position) creates delay which is a function of the number of blocks transferred; typically relatively small, but can be quite long with the transfer of large contiguous files.

Delay may also occur if the drive disks are stopped to save energy.

[[Defragmentation]] is a procedure used to minimize delay in retrieving data by moving related items to physically proximate areas on the disk.<ref name="itworld-2001-04-18" /> Some computer operating systems perform defragmentation automatically. Although automatic defragmentation is intended to reduce access delays, performance will be temporarily reduced while the procedure is in progress.<ref name="AutoMK-67" />

Time to access data can be improved by increasing rotational speed (thus reducing latency) or by reducing the time spent seeking. Increasing areal density increases [[throughput]] by increasing data rate and by increasing the amount of data under a set of heads, thereby potentially reducing seek activity for a given amount of data. The time to access data has not kept up with throughput increases, which themselves have not kept up with growth in bit density and storage capacity.

=== Latency ===
{| class="wikitable" style="margin-left: 1.5em;"
|+Latency characteristics typical of HDDs
|-
! Rotational speed (rpm)
! Average rotational latency (ms){{Efn|Average rotational latency in milliseconds is computed as follows: 60&nbsp;× 1000&nbsp;÷ 2&nbsp;÷ R, where {{mvar|R}} is rotational speed revolutions per minute.}}
|-
| 15,000
| 2
|-
| 10,000
| 3
|-
| 7,200
| 4.16
|-
| 5,400
| 5.55
|-
| 4,800
| 6.25
|}

=== Data transfer rate ===
{{As of|2010}}, a typical 7,200-rpm desktop HDD has a sustained "disk-to-[[disk buffer|buffer]]" data transfer rate up to {{nowrap|1,030 Mbit/s}}.<ref name="AutoMK-71" /> This rate depends on the track location; the rate is higher for data on the outer tracks (where there are more data sectors per rotation) and lower toward the inner tracks (where there are fewer data sectors per rotation); and is generally somewhat higher for 10,000-rpm drives. A current, widely used standard for the "buffer-to-computer" interface is {{nowrap|3.0 Gbit/s}} SATA, which can send about 300 megabyte/s (10-bit encoding) from the buffer to the computer, and thus is still comfortably ahead of today's{{As of?|date=September 2023}} disk-to-buffer transfer rates. Data transfer rate (read/write) can be measured by writing a large file to disk using special file-generator tools, then reading back the file. Transfer rate can be influenced by [[file system fragmentation]] and the layout of the files.<ref name="itworld-2001-04-18" />

HDD data transfer rate depends upon the rotational speed of the platters and the data recording density. Because heat and vibration limit rotational speed, advancing density becomes the main method to improve sequential transfer rates. Higher speeds require a more powerful spindle motor, which creates more heat. While areal density advances by increasing both the number of tracks across the disk and the number of sectors per track,<ref>{{cite web |title=GLOSSARY of DRIVE and COMPUTER TERMS |url=http://ftp.seagate.com/techsuppt/misc/glossary.txt |publisher=Seagate |access-date=August 4, 2018}}</ref> only the latter increases the data transfer rate for a given rpm. Since data transfer rate performance tracks only one of the two components of areal density, its performance improves at a lower rate.<ref>{{cite journal |title=Bit Patterned Magnetic Recording: Theory, Media Fabrication, and Recording Performance |journal = IEEE Transactions on Magnetics|volume = 51|issue = 5|pages = 1–42|publisher=HGST, a Western Digital Company |arxiv = 1503.06664|doi = 10.1109/TMAG.2015.2397880|year = 2015|last1 = Albrecht|first1 = Thomas R.|last2 = Arora|first2 = Hitesh|last3 = Ayanoor-Vitikkate|first3 = Vipin|last4 = Beaujour|first4 = Jean-Marc|last5 = Bedau|first5 = Daniel|last6 = Berman|first6 = David|last7 = Bogdanov|first7 = Alexei L.|last8 = Chapuis|first8 = Yves-Andre|last9 = Cushen|first9 = Julia|last10 = Dobisz|first10 = Elizabeth E.|last11 = Doerk|first11 = Gregory|last12 = He Gao|last13 = Grobis|first13 = Michael|last14 = Gurney|first14 = Bruce|last15 = Hanson|first15 = Weldon|last16 = Hellwig|first16 = Olav|last17 = Hirano|first17 = Toshiki|last18 = Jubert|first18 = Pierre-Olivier|last19 = Kercher|first19 = Dan|last20 = Lille|first20 = Jeffrey|last21 = Zuwei Liu|last22 = Mate|first22 = C. Mathew|last23 = Obukhov|first23 = Yuri|last24 = Patel|first24 = Kanaiyalal C.|last25 = Rubin|first25 = Kurt|last26 = Ruiz|first26 = Ricardo|last27 = Schabes|first27 = Manfred|last28 = Lei Wan|last29 = Weller|first29 = Dieter|last30 = Tsai-Wei Wu|display-authors = 29|bibcode = 2015ITM....5197880A|s2cid = 33974771}}</ref>

=== Other considerations ===
Other performance considerations include quality-adjusted [[price index|price]], power consumption, audible noise, and both operating and non-operating shock resistance.

== Access and interfaces ==
{{Main|Hard disk drive interface}}
[[File:Seagate ST33232A hard disk inner view.jpg|thumb|right|Inner view of a 1998 [[Seagate Technology|Seagate]] HDD that used the [[Parallel ATA]] interface]]
[[File:2.5-inch SATA drive on top of a 3.5-inch SATA drive, close-up of data and power connectors.jpg|thumb|right|2.5-inch SATA drive on top of 3.5-inch SATA drive, showing close-up of (7-pin) data and (15-pin) power connectors]]

Current hard drives connect to a computer over one of several [[bus (computing)|bus]] types, including parallel [[Advanced Technology Attachment|ATA]], [[Serial ATA]], [[SCSI]], [[Serial Attached SCSI]] (SAS), and [[Fibre Channel]]. Some drives, especially external portable drives, use [[IEEE 1394 interface|IEEE&nbsp;1394]], or [[USB]]. All of these interfaces are digital; electronics on the drive process the analog signals from the read/write heads. Current drives present a consistent interface to the rest of the computer, independent of the data encoding scheme used internally, and independent of the physical number of disks and heads within the drive.

Typically, a [[Digital signal processor|DSP]] in the electronics inside the drive takes the raw analog voltages from the read head and uses [[PRML]] and [[Reed–Solomon error correction]]<ref name="AutoMK-79" /> to decode the data, then sends that data out the standard interface. That DSP also watches the error rate detected by [[error detection and correction]], and performs [[bad sector]] remapping, data collection for [[Self-Monitoring, Analysis, and Reporting Technology]], and other internal tasks.

Modern interfaces connect the drive to the host interface with a single data/control cable. Each drive also has an additional power cable, usually direct to the power supply unit. Older interfaces had separate cables for data signals and for drive control signals.
* [[Small Computer System Interface]] (SCSI), originally named SASI for Shugart Associates System Interface, was standard on servers, workstations, [[Amiga peripherals|Commodore Amiga]], [[Atari ST]] and [[Apple Macintosh]] computers through the mid-1990s, by which time most models had been transitioned to newer interfaces. The length limit of the data cable allows for external SCSI devices. The SCSI command set is still used in the more modern [[Serial Attached SCSI|SAS interface]].
* [[Integrated Drive Electronics]] (IDE), later standardized under the name [[AT Attachment]] (ATA, with the alias PATA ([[Parallel ATA]]) retroactively added upon introduction of SATA) moved the HDD controller from the interface card to the disk drive. This helped to standardize the host/controller interface, reduce the programming complexity in the host device driver, and reduced system cost and complexity. The 40-pin IDE/ATA connection transfers 16&nbsp;bits of data at a time on the data cable. The data cable was originally 40-conductor, but later higher speed requirements led to an [[AT Attachment#Parallel ATA interface|"ultra DMA" (UDMA)]] mode using an 80-conductor cable with additional wires to reduce [[crosstalk]] at high speed.
* EIDE was an unofficial update (by Western Digital) to the original IDE standard, with the key improvement being the use of [[direct memory access]] (DMA) to transfer data between the disk and the computer without the involvement of the [[CPU]], an improvement later adopted by the official ATA standards. By directly transferring data between memory and disk, DMA eliminates the need for the CPU to copy byte per byte, therefore allowing it to process other tasks while the data transfer occurs.
* [[Fibre Channel]] (FC) is a successor to parallel SCSI interface on enterprise market. It is a serial protocol. In disk drives usually the [[Fibre Channel Arbitrated Loop]] (FC-AL) connection topology is used. FC has much broader usage than mere disk interfaces, and it is the cornerstone of [[storage area network]]s (SANs). Recently other protocols for this field, like [[iSCSI]] and [[ATA over Ethernet]] have been developed as well. Confusingly, drives usually use ''copper'' twisted-pair cables for Fibre Channel, not fiber optics. The latter are traditionally reserved for larger devices, such as servers or [[disk array controller]]s.
* [[Serial Attached SCSI]] (SAS). The SAS is a new generation serial communication protocol for devices designed to allow for much higher speed data transfers and is compatible with SATA. SAS uses a mechanically compatible data and power connector to standard 3.5-inch SATA1/SATA2 HDDs, and many server-oriented SAS RAID controllers are also capable of addressing SATA HDDs. SAS uses serial communication instead of the parallel method found in traditional SCSI devices but still uses SCSI commands.
* [[Serial ATA]] (SATA). The SATA data cable has one data pair for differential transmission of data to the device, and one pair for differential receiving from the device, just like [[EIA-422]]. That requires that data be transmitted serially. A similar [[differential signaling]] system is used in [[RS485]], [[LocalTalk]], [[USB]], [[FireWire]], and differential [[SCSI]]. SATA I to III are designed to be compatible with, and use, a subset of SAS commands, and compatible interfaces. Therefore, a SATA hard drive can be connected to and controlled by a SAS hard drive controller (with some minor exceptions such as drives/controllers with limited compatibility). However, they cannot be connected the other way round&mdash;a SATA controller cannot be connected to a SAS drive.

== Integrity and failure ==
{{Main|Hard disk drive failure|Head crash|Data recovery}}
{{See also|Solid-state drive#SSD reliability and failure modes|RAID#Unrecoverable read errors during rebuild}}
{{More citations needed section|date=January 2025|talk=More citations needed in integrity section}}
[[File:Hard disk head.jpg|thumb|Close-up of an HDD head resting on a disk platter; its mirror reflection is visible on the platter surface. Unless the head is on a landing zone, the heads touching the platters while in operation can be catastrophic.]]
[[File:Hard disk head crash.jpg|thumb|HDD head crash]]
[[File:Broken HDD head (kaputt).jpg|thumb|Even worse crash]]

Due to the extremely close spacing between the heads and the disk surface, HDDs are vulnerable to being damaged by a [[head crash]]&nbsp;– a [[hard disk failure|failure of the disk]] in which the head scrapes across the platter surface, often grinding away the thin magnetic film and causing data loss. Head crashes can be caused by electronic failure, a sudden power failure, physical shock, contamination of the drive's internal enclosure, wear and tear, [[corrosion]], or poorly manufactured platters and heads.

The HDD's spindle system relies on [[air density]] inside the [[disk enclosure]] to support the heads at their proper [[flying height]] while the disk rotates. HDDs require a certain range of air densities to operate properly. The connection to the external environment and density occurs through a small hole in the enclosure (about 0.5&nbsp;mm in breadth), usually with a filter on the inside (the ''breather filter'').<ref name="AutoMK-80" /> If the air density is too low, then there is not enough lift for the flying head, so the head gets too close to the disk, and there is a risk of head crashes and data loss. Specially manufactured sealed and pressurized disks are needed for reliable high-altitude operation, above about {{convert|3000|m|ft|abbr=on}}.<ref name="AutoMK-81" /> Modern disks include temperature sensors and adjust their operation to the operating environment. Breather holes can be seen on most disk drives, excluding sealed drives, such as drives that use helium, where any exposure to outside air would cause a failure&nbsp;– they usually have a sticker next to them, warning the user not to cover the holes. The air inside the operating drive is constantly moving too, being swept in motion by friction with the spinning platters. This air passes through an internal recirculation filter to remove any leftover contaminants from manufacture, any particles or chemicals that may have somehow entered the enclosure, and any particles or outgassing generated internally in normal operation. Very high humidity present for extended periods of time can corrode the heads and platters. An exception to this are hermetically sealed, helium-filled HDDs that largely eliminate environmental issues that can arise due to humidity or atmospheric pressure changes. Such HDDs were introduced by HGST in their first successful high-volume implementation in 2013.

For [[giant magnetoresistive]] (GMR) heads in particular, a minor head crash from contamination (that does not remove the magnetic surface of the disk) still results in the head temporarily overheating, due to friction with the disk surface and can render the data unreadable for a short period until the head temperature stabilizes (so-called "thermal asperity", a problem which can partially be dealt with by proper electronic filtering of the read signal).

When the logic board of a hard disk fails, the drive can often be restored to functioning order and the data recovered by replacing the circuit board with one of an identical hard disk. In the case of read-write head faults, they can be replaced using specialized tools in a dust-free environment. If the disk platters are undamaged, they can be transferred into an identical enclosure and the data can be copied or cloned onto a new drive. In the event of disk-platter failures, disassembly and imaging of the disk platters may be required.<ref>{{cite web|url=http://electronics.howstuffworks.com/how-to-tech/how-to-recover-data-hard-drive4.htm|title=How To Recover Lost Data from Your Hard Drive|last=Grabianowski|first=Ed|publisher=HowStuffWorks|pages=5–6|access-date=October 24, 2012|date=May 29, 2009|archive-url=https://web.archive.org/web/20121105121608/http://electronics.howstuffworks.com/how-to-tech/how-to-recover-data-hard-drive4.htm|archive-date=November 5, 2012|url-status=live}}</ref> For logical damage to file systems, a variety of tools, including [[fsck]] on [[UNIX-like]] systems and [[CHKDSK]] on [[Windows]], can be used for [[data recovery]]. Recovery from logical damage can require [[file carving]].

A common expectation is that hard disk drives designed and marketed for server use will fail less frequently than consumer-grade drives usually used in desktop computers. However, two independent studies by [[Carnegie Mellon University]]<ref name="CMUDiskFailure">{{cite web |title=Everything You Know About Disks Is Wrong |url=https://storagemojo.com/2007/02/20/everything-you-know-about-disks-is-wrong/ |website=Storagemojo.com |date=February 22, 2007 |access-date=May 24, 2019 |archive-url=https://web.archive.org/web/20190524230125/https://storagemojo.com/2007/02/20/everything-you-know-about-disks-is-wrong/ |archive-date=May 24, 2019 |url-status=live }}</ref> and [[Google]]<ref name="GoogleDiskFailure">{{cite web |title=Failure Trends in a Large Disk Drive Population |first1=Eduardo |last1=Pinheiro |author2=Wolf-Dietrich Weber |author3=Luiz André Barroso |url=http://research.google.com/archive/disk_failures.pdf |publisher=Google Inc |date=February 2007 |access-date=December 26, 2011 |archive-url=https://web.archive.org/web/20100105013605/http://research.google.com/archive/disk_failures.pdf |archive-date=January 5, 2010 |url-status=live }}</ref> found that the "grade" of a drive does not relate to the drive's failure rate.

A 2011 summary of research, into SSD and magnetic disk failure patterns by [[Tom's Hardware]] summarized research findings as follows:<ref>[http://www.tomshardware.com/reviews/ssd-reliability-failure-rate,2923-9.html Investigation: Is Your SSD More Reliable Than A Hard Drive?] – [[Tom's Hardware]] long term SSD reliability review, 2011, "final words"</ref>
* [[Mean time between failures]] (MTBF) does not indicate reliability; the annualized failure rate is higher and usually more relevant. 
* HDDs do not tend to fail during early use, and temperature has only a minor effect; instead, failure rates steadily increase with age.
* S.M.A.R.T. warns of mechanical issues but not other issues affecting reliability, and is therefore not a reliable indicator of condition.<ref>{{cite web|last1=Anthony|first1=Sebastian|title=Using SMART to accurately predict when a hard drive is about to die|date=November 12, 2014 |url=http://www.extremetech.com/computing/194059-using-smart-to-accurately-predict-when-a-hard-drive-is-about-to-die|publisher=ExtremeTech|access-date=August 25, 2015|archive-url=https://web.archive.org/web/20150831050118/http://www.extremetech.com/computing/194059-using-smart-to-accurately-predict-when-a-hard-drive-is-about-to-die|archive-date=August 31, 2015|url-status=live}}</ref>
* Failure rates of drives sold as "enterprise" and "consumer" are "very much similar", although these drive types are customized for their different operating environments.<ref>{{cite web|title=Consumer hard drives as reliable as enterprise hardware|work=Alphr |date=December 4, 2013 |url=http://www.alphr.com/technology/23998/consumer-hard-drives-as-reliable-as-enterprise-hardware|access-date=August 25, 2015|archive-url=https://web.archive.org/web/20150911033552/http://www.alphr.com/technology/23998/consumer-hard-drives-as-reliable-as-enterprise-hardware|archive-date=September 11, 2015|url-status=live}}</ref><ref>{{cite web|last1=Beach|first1=Brian|title=Enterprise Drives: Fact or Fiction?|url=https://www.backblaze.com/blog/enterprise-drive-reliability/|publisher=Backblaze|access-date=August 25, 2015|date=December 4, 2013|archive-url=https://web.archive.org/web/20150818015807/https://www.backblaze.com/blog/enterprise-drive-reliability/|archive-date=August 18, 2015|url-status=live}}</ref>
* In drive arrays, one drive's failure significantly increases the short-term risk of a second drive failing.

{{As of|2019}}, Backblaze, a storage provider, reported an annualized failure rate of two percent per year for a storage farm with 110,000 off-the-shelf HDDs with the reliability varying widely between models and manufacturers.<ref name= "Backblaze 3Q2019" >{{cite web |title= Hard Drive Data and Stats |url= https://www.backblaze.com/b2/hard-drive-test-data.html |publisher=Backblaze |access-date= November 24, 2019 }}</ref> Backblaze subsequently reported that the failure rate for HDDs and SSD of equivalent age was similar.<ref name="BBRel2021">{{cite web
 |url=https://www.backblaze.com/blog/are-ssds-really-more-reliable-than-hard-drives/ 
 |title=Are SSDs Really More Reliable Than Hard Drives?  |last=Klein  |first=Andy
 |date=September 30, 2021  |website=Backblaze  |access-date=September 30, 2021
 |quote=Once we controlled for age and drive days, the two drive types were similar and the difference was certainly not enough by itself to justify the extra cost of purchasing a SSD versus a HDD.}}</ref>

To minimize cost and overcome failures of individual HDDs, storage systems providers rely on redundant HDD arrays. HDDs that fail are replaced on an ongoing basis.<ref name= "Backblaze 3Q2019"/><ref name= "HAMR 2008 for 2009"/>

== Market segments ==
===Consumer segment===
: [[File:Two Western Digital VelociRaptor 1 TB SATA 10,000 rpm 3.5-inch HDDs.jpg|thumb|right|Two high-end consumer SATA 2.5-inch 10,000&nbsp;rpm HDDs, factory-mounted in 3.5-inch adapter frames]]
; Desktop HDDs
: Desktop HDDs typically have one to five internal platters, rotate at 5,400 to 10,000&nbsp;rpm, and have a media transfer rate of {{nowrap|0.5 Gbit/s}} or higher (1&nbsp;GB = 10<sup>9</sup> bytes; {{nowrap|1 Gbit/s}} = {{nowrap|10<sup>9</sup> bit/s}}). Earlier (1980–1990s) drives tend to be slower in rotation speed. {{As of|January 2025|}}, the highest-capacity [[Desktop computer|desktop]] HDDs stored 36[[terabyte|TB]],<ref>{{Cite web|url=https://www.notebookcheck.net/Seagate-introduces-world-first-16TB-Exos-HDD-and-IronWolf-NAS-drives.423041.0.html|title=Seagate introduces world-first 16TB Exos HDD and IronWolf NAS drives|first=Deirdre O.|last=Donnell|website=Notebookcheck|date=June 4, 2019 }}</ref><ref>{{Cite web | url=https://www.seagate.com/em/en/internal-hard-drives/hdd/barracuda/ | title=BarraCuda en BarraCuda Pro interne harde schijven &#124; Seagate Nederland | access-date=November 9, 2019 | archive-url=https://web.archive.org/web/20190506051335/https://www.seagate.com/em/en/internal-hard-drives/hdd/barracuda/ | archive-date=May 6, 2019 | url-status=live }}</ref> with plans to release 50&nbsp;TB drives later in 2025.<ref>{{cite news |url=https://www.anandtech.com/show/13764/western-digital-2019-16tb-hdd-mamr-hamr |title=16 TB MAMR Hard Drives in 2019: Western Digital |access-date=May 24, 2019 |archive-url=https://web.archive.org/web/20190524072831/https://www.anandtech.com/show/13764/western-digital-2019-16tb-hdd-mamr-hamr |archive-date=May 24, 2019 |url-status=live }}</ref> 36&nbsp;TB HDDs were released in 2025.{{Citation needed|date=September 2023}}. {{As of|2016}}, the typical speed of a hard drive in an average desktop computer is 7,200&nbsp;rpm, whereas low-cost desktop computers may use 5,900&nbsp;rpm or 5,400&nbsp;rpm drives. For some time in the 2000s and early 2010s some desktop users and data centers also used 10,000&nbsp;rpm drives such as [[Western Digital Raptor]] but such drives have become much rarer {{as of|2016|lc=yes}} (since the [[Western Digital Raptor|WD VelociRaptor]] was discontinued) and are not commonly used now, having been replaced by NAND flash-based SSDs.

; Mobile (laptop) HDDs
: Smaller than their desktop and enterprise counterparts, they tend to be slower and have lower capacity, because typically has one internal platter and were 2.5" or 1.8" physical size instead of more common for desktops 3.5" form-factor. Mobile HDDs spin at 4,200&nbsp;rpm, 5,200&nbsp;rpm, 5,400&nbsp;rpm, or 7,200&nbsp;rpm, with 5,400&nbsp;rpm being the most common; 7,200&nbsp;rpm drives tend to be more expensive and have smaller capacities, while 4,200&nbsp;rpm models usually were in older laptops and portables but are now outdated. Because of smaller platter(s), mobile HDDs generally have lower capacity than their desktop counterparts.

; Consumer electronics HDDs
These drives typically spin at 5400&nbsp;rpm and include:
* {{anchor|VideoHDD}}'''Video hard drives''', sometimes called "'''surveillance hard drives'''", are embedded into [[digital video recorder]]s and provide a guaranteed streaming capacity, even in the face of read and write errors.<ref name="VHD">{{cite web  |url=https://learncctv.com/how-to-choose-the-best-hard-drives-for-dvrs-and-nvrs/
 |title=How to choose the best hard drive for DVRs and NVRs |date=August 29, 2019 |access-date=August 28, 2023}}</ref>
* Drives embedded into [[automotive vehicle]]s; they are typically built to resist larger amounts of shock and operate over a larger temperature range.
; {{Anchor|PORTABLE|EXTERNAL|REMOVABLE}}External and portable HDDs
: {{See also|USB mass storage device class|Disk enclosure}} 
[[File:WD External Hard Drives IMG 7899.jpg|thumb|Two 2.5" external USB hard drives]]
[[File:ST3400820AS.jpg|thumb|Seagate Hard Drive with a controller board to convert SATA to USB, FireWire, and eSATA]]
: Current external hard disk drives typically connect via [[USB-C]]; earlier models use USB-B (sometimes with using of a pair of ports for better bandwidth) or (rarely) [[eSATA]] connection. Variants using USB&nbsp;2.0 interface generally have slower data transfer rates when compared to internally mounted hard drives connected through SATA. [[Plug and play]] drive functionality offers system compatibility and features large storage options and portable design. {{As of|2015|3}}, available capacities for external hard disk drives ranged from 500&nbsp;GB to 10&nbsp;TB.<ref>{{cite web|title=Seagate Backup Plus External Hard Drive Review (8TB)|url=http://www.storagereview.com/seagate_backup_plus_external_hard_drive_review_8tb|url-status=live|archive-url=https://web.archive.org/web/20150725004556/http://www.storagereview.com/seagate_backup_plus_external_hard_drive_review_8tb|archive-date=July 25, 2015|access-date=July 20, 2015|work=storagereview.com|date=March 22, 2015}}</ref> External hard disk drives are usually available as assembled integrated products, but may be also assembled by combining an external [[Disk enclosure|enclosure]] (with USB or other interface) with a separately purchased drive. They are available in 2.5-inch and 3.5-inch sizes; 2.5-inch variants are typically called ''portable external drives'', while 3.5-inch variants are referred to as ''desktop external drives''. "Portable" drives are packaged in smaller and lighter enclosures than the "desktop" drives; additionally, "portable" drives use power provided by the USB connection, while "desktop" drives require external [[power brick]]s. Features such as [[Self encrypting drive|encryption]], [[Wi-Fi]] connectivity,<ref>{{cite web |url=https://www.storagereview.com/review/wd-my-passport-wireless-review |title=WD My Passport Wireless Review |website=storagereview.com |last=Smith |first=Lyle |date=September 3, 2014 |access-date=July 21, 2021 }}</ref> biometric security or multiple interfaces (for example, [[FireWire]]) are available at a higher cost.<ref name="AutoMK-98" /> There are pre-assembled external hard disk drives that, when taken out from their enclosures, cannot be used internally in a laptop or desktop computer due to embedded USB interface on their [[printed circuit board]]s, and lack of SATA (or [[Parallel ATA]]) interfaces.<ref>{{cite web|title=Western Digital My Passport, 2&nbsp;TB|url=http://content.hwigroup.net/images/products/xl/152766-7.jpg|url-status=dead|archive-url=https://web.archive.org/web/20131005000039/http://content.hwigroup.net/images/products/xl/152766-7.jpg|archive-date=October 5, 2013|access-date=January 11, 2014|website=hwigroup.net|quote=Example of a pre-assembled external hard disk drive without its enclosure that cannot be used internally on a laptop or desktop due to the embedded interface on its printed circuit board}}</ref><ref>{{cite web|last=Hsiung|first=Sebean|date=May 5, 2010|title=How to bypass USB controller and use as a SATA drive|url=http://www.datarecoverytools.co.uk/2010/05/05/how-to-connect-and-recover-usb-only-western-digital-drives-with-hd-doctor-suite/|url-status=live|archive-url=https://web.archive.org/web/20140915191641/http://www.datarecoverytools.co.uk/2010/05/05/how-to-connect-and-recover-usb-only-western-digital-drives-with-hd-doctor-suite/|archive-date=September 15, 2014|access-date=January 11, 2014|website=datarecoverytools.co.uk}}</ref>

===Enterprise and business segment===
; Server and workstation HDDs
: [[File:IBM TotalStorage Exp400.jpg|thumb|[[hot-swapping|Hot-swappable]] HDD enclosure]]
: Typically used with multiple-user computers running [[enterprise software]]. Examples are: transaction processing databases, internet infrastructure (email, webserver, e-commerce), scientific computing software, and nearline storage management software. Enterprise drives commonly operate continuously ("24/7") in demanding environments while delivering the highest possible performance without sacrificing reliability. Maximum capacity is not the primary goal, and as a result the drives are often offered in capacities that are relatively low in relation to their cost.<ref name="AutoMK-99" />
: The fastest enterprise HDDs spin at 10,000 or 15,000&nbsp;rpm, and can achieve sequential media transfer speeds above {{nowrap|1.6 Gbit/s}}<ref name="erwpnw" /> and a sustained transfer rate up to {{nowrap|1 Gbit/s}}.<ref name="erwpnw" /> Drives running at 10,000 or 15,000&nbsp;rpm use smaller platters to mitigate increased power requirements (as they have less [[air drag]]) and therefore generally have lower capacity than the highest capacity desktop drives. Enterprise HDDs are commonly connected through [[Serial Attached SCSI]] (SAS) or [[Fibre Channel]] (FC). Some support multiple ports, so they can be connected to a redundant [[host bus adapter]].
: Enterprise HDDs can have sector sizes larger than 512 bytes (often 520, 524, 528 or 536 bytes). The additional per-sector space can be used by hardware RAID controllers or applications for storing [[Data Integrity Field]] (DIF) or Data Integrity Extensions (DIX) data, resulting in higher reliability and prevention of [[silent data corruption]].<ref>{{cite web
 | url = https://oss.oracle.com/~mkp/docs/lpc08-data-integrity.pdf
 | archive-url = https://web.archive.org/web/20150109235547/https://oss.oracle.com/~mkp/docs/lpc08-data-integrity.pdf
 | title = Linux Data Integrity
 | date = August 30, 2008 | access-date = January 23, 2015 | archive-date = January 9, 2015
 | first = Martin K. | last = Petersen | publisher = [[Oracle Corporation]]
 | quote = Most disk drives use 512-byte sectors. [...] Enterprise drives (Parallel SCSI/SAS/FC) support 520/528 byte 'fat' sectors.
}}</ref>

;[[#VideoHDD|Surveillance hard drives]]; 
: Video recording HDDs used in network video recorders.<ref name="VHD"/>

== Economy ==
=== Price evolution ===
HDD price per byte decreased at the rate of 40% per year during 1988–1996, 51% per year during 1996–2003 and 34% per year during 2003–2010.<ref name="jcmit">{{cite web
 | url = http://www.jcmit.com/diskprice.htm
 | archive-url = https://web.archive.org/web/20150714062141/http://www.jcmit.com/diskprice.htm
 | title = Disk Drive Prices (1955–2015)
 | date = May 16, 2015 | access-date = July 25, 2015 | archive-date = July 14, 2015
 | first = John C. | last = McCallum | website = jcmit.com
}}</ref><ref name="Byrne2015b" /> The price decrease slowed down to 13% per year during 2011–2014, as areal density increase slowed and the [[2011 Thailand floods]] damaged manufacturing facilities<ref name="Mellor 2014-11-10" /> and have held at 11% per year during 2010–2017.<ref>{{cite web |date=July 11, 2017 |title=Hard Drive Cost Per Gigabyte |url=https://www.backblaze.com/blog/hard-drive-cost-per-gigabyte/ |url-status=live |archive-url=https://web.archive.org/web/20190526075044/https://www.backblaze.com/blog/hard-drive-cost-per-gigabyte/ |archive-date=May 26, 2019 |access-date=May 26, 2019 |publisher=Backblaze}}</ref>

The [[Federal Reserve Board]] has published a quality-adjusted [[price index]] for large-scale enterprise storage systems including three or more enterprise HDDs and associated controllers, racks and cables. Prices for these large-scale storage systems decreased at the rate of 30% per year during 2004–2009 and 22% per year during 2009–2014.<ref name="Byrne2015b" />

=== Manufacturers and sales ===
{{See also|History of hard disk drives|List of defunct hard disk manufacturers}}
[[File:Diagram of Hard Disk Drive Manufacturer Consolidation.svg|thumb|Diagram of HDD manufacturer consolidation]]

More than 200 companies have manufactured HDDs over time, but consolidations have concentrated production to just three manufacturers today: [[Western Digital]], [[Seagate Technology|Seagate]], and [[Toshiba]].  Production is mainly in the Pacific rim.

HDD unit shipments peaked at 651 million units in 2010 and have been declining since then to 166 million units in 2022.<ref>{{cite web
 |url=https://www.statista.com/statistics/398951/global-shipment-figures-for-hard-disk-drives/
 |title=Hard disk drive (HDD) unit shipments worldwide from 1976 to 2022 |publisher=Statista  |access-date=July 28, 2023}}</ref> Seagate at 43% of units had the largest market share.<ref>{{cite web
 |url=https://www.statista.com/statistics/1257240/hard-disk-drive-hdd-supplier-market-share/
 |title=Hard disk drive (HDD) supplier market shipment share worldwide in 2022 |last=Alsop |first=Thomas |date=February 17, 2023 |publisher=Statista |access-date=July 28, 2023}}</ref>
<!--
Bad number:

HDD revenue in 2022 is estimated at $222 million.<ref>{{cite web |url=https://www.marketwatch.com/press-release/2023-hard-disk-drives-market-sales-and-forecast-till-2030-2023-06-11
 |title=2023 Hard Disk Drives Market Sales and Forecast till 2030|date=June 11, 2023 |publisher=Industry Research |access-date=July 28, 2023}}</ref>
-->

== Competition from SSDs ==
[[File:SEAGATE 7200.9 160GB HDD+2230 256GB NVME SSD+SANDISK X400 256GB SSD.jpg|thumb|HDD and SSD]]
HDDs are being superseded by [[solid-state drive]]s (SSDs) in markets where the higher speed (up to 7 [[gigabyte]]s per second for [[M.2]] (NGFF) [[NVMe]] drives<ref>{{Cite web|url=https://www.corsair.com/ww/en/Categories/Products/Storage/M-2-SSDs/Force-Series%E2%84%A2-Gen-4-PCIe-NVMe-M-2-SSD/p/CSSD-F2000GBMP600|title=Force Series Gen.4 PCIe MP600 2TB NVMe M.2 SSD|website=www.corsair.com|access-date=2020-03-06}}</ref> and 2.5 [[gigabyte]]s per second for [[PCIe]] expansion card drives)<ref>{{cite web|url=https://www.storagereview.com/intel_optane_ssd_900p_series_review |title=Intel Optane SSD 900P Series Review |work=StorageReview.com |date=March 16, 2018 |access-date=February 20, 2019|archive-url=https://web.archive.org/web/20181231092754/https://www.storagereview.com/intel_optane_ssd_900p_series_review |archive-date=December 31, 2018 |url-status=live }}</ref>), ruggedness, and lower power of SSDs are more important than price, since the bit cost of SSDs is four to nine times higher than HDDs.<ref name= "blocks August2019"/><ref name="jcmit 2019" >{{cite web|url=https://jcmit.net/disk2015.htm|title=Disk Drive Storage Price Decreasing with Time (1955-2019) |date = November 2019 |access-date = November 25, 2019 |first= John C. |last= McCallum |website = jcmit.com }}</ref> {{As of|2016}}, HDDs are reported to have a failure rate of 2–9% per year, while SSDs have fewer failures: 1–3% per year.<ref name= "Google 2016"/> However, SSDs have more un-correctable data errors than HDDs.<ref name= "Google 2016" >{{cite web |url = https://www.usenix.org/system/files/conference/fast16/fast16-papers-schroeder.pdf |title = Flash Reliability in Production: The Expected and the Unexpected |first1= Bianca |last1= Schroeder |author1-link= Bianca Schroeder |first2= Raghav  |last2= Lagisetty  |first3= Arif |last3= Merchant |date = February 22, 2016 |access-date = November 25, 2019 }}</ref>

SSDs are available in larger capacities (up to 100&nbsp;TB)<ref name= "100TB 2018" >{{cite web|last=Alcorn |first=Paul |url=https://www.tomshardware.com/news/100tb-ssd-nimbus-sata-flash,36687.html |title=Need A 100TB SSD? Nimbus Data Has You Covered With The ExaDrive DC100 |work=Tomshardware.com |date=March 19, 2018 |access-date=February 20, 2019}}</ref> than the largest HDD, as well as higher storage densities (100&nbsp;TB and 30&nbsp;TB SSDs are housed in 2.5 inch HDD cases with the same height as a 3.5-inch HDD),<ref>{{cite web|url=https://bgr.com/2018/02/20/samsung-30tb-ssd-release-date-specs-price/amp|title=You won't be able to afford Samsung's record-setting 30TB SSD|work=Bgr.com|date=February 20, 2018|access-date=February 20, 2019|archive-url=https://web.archive.org/web/20190410173238/https://bgr.com/2018/02/20/samsung-30tb-ssd-release-date-specs-price/amp|archive-date=April 10, 2019|url-status=live}}</ref><ref>{{cite web|author=Circuit Breaker|url=https://www.theverge.com/platform/amp/circuitbreaker/2018/2/20/17031256/worlds-largest-ssd-drive-samsung-30-terabyte-pm1643|title=Samsung unveils world's largest SSD with whopping 30TB of storage|work=The Verge|date=February 20, 2018|access-date=February 20, 2019|archive-url=https://web.archive.org/web/20190127210421/https://www.theverge.com/platform/amp/circuitbreaker/2018/2/20/17031256/worlds-largest-ssd-drive-samsung-30-terabyte-pm1643|archive-date=January 27, 2019|url-status=live}}</ref><ref>{{cite web|url=https://nimbusdata.com/products/exadrive-platform/advantages|title=Advantages|work=Nimbus Data|date=July 22, 2016 |access-date=February 20, 2019|archive-url=https://web.archive.org/web/20181231093052/https://nimbusdata.com/products/exadrive-platform/advantages|archive-date=December 31, 2018|url-status=live}}</ref><ref>{{cite web|url=https://nimbusdata.com/products/exadrive-platform/scalable-ssds|title=Scalable SSDs|work=Nimbus Data|date=July 22, 2016 |access-date=February 20, 2019|archive-url=https://web.archive.org/web/20181231143741/https://nimbusdata.com/products/exadrive-platform/scalable-ssds|archive-date=December 31, 2018|url-status=live}}</ref><ref>{{cite web|url=https://www.computerworld.com/article/3101165/data-storage/samsungs-massive-15tb-ssd-can-be-yours-for-about-10k.amp.html|title=Samsung's massive 15TB SSD can be yours - for about $10K|work=Computerworld|date=July 27, 2016|access-date=February 20, 2019|archive-url=https://web.archive.org/web/20181231092730/https://www.computerworld.com/article/3101165/data-storage/samsungs-massive-15tb-ssd-can-be-yours-for-about-10k.amp.html|archive-date=December 31, 2018|url-status=live}}</ref> although such large SSDs are very expensive.

A laboratory demonstration of a 1.33&nbsp;Tb 3D NAND chip with 96 layers (NAND commonly used in [[solid-state drive]]s (SSDs)) had 5.5&nbsp;Tbit/in<sup>2</sup> {{as of|2019|lc=yes}}),<ref name= "EETimes_2019Feb" >{{cite web|url=https://www.eetimes.com/document.asp?doc_id=1334344#|title=Toshiba Claims Highest-Capacity NAND|last=McGrath|first=Dylan|date=February 20, 2019|access-date=November 24, 2019}}</ref> while the maximum areal density for HDDs is 1.5&nbsp;Tbit/in<sup>2</sup>. The areal density of flash memory is doubling every two years, similar to [[Moore's law]] (40% per year) and faster than the 10–20% per year for HDDs. In 2025, the maximum capacity was 36&nbsp;terabytes for a HDD,<ref name="Mearian_2015Dec">{{cite web|url=https://www.alphr.com/storage/1010285/seagate-reveals-world-s-largest-and-most-ludicrous-16tb-hdd|title=Seagate reveals world's largest, and most ludicrous 16TB HDD|last=Bedford|first=Tom|date=December 4, 2018|work=Alphr|access-date=December 24, 2018|archive-url=https://web.archive.org/web/20181224074254/https://www.alphr.com/storage/1010285/seagate-reveals-world-s-largest-and-most-ludicrous-16tb-hdd|archive-date=December 24, 2018|url-status=live}}</ref> and 100&nbsp;terabytes for an SSD.<ref name="Coughlin_2016Feb">{{cite web|url=https://www.anandtech.com/show/12541/unlimited-5-year-endurance-100-tb-ssd|title=Unlimited 5 Year Endurance: The 100TB SSD from Nimbus Data|last1=Shilov|first1=Anton|date=March 19, 2018|website=AnandTech|access-date=December 24, 2018|archive-url=https://web.archive.org/web/20181224073911/https://www.anandtech.com/show/12541/unlimited-5-year-endurance-100-tb-ssd|archive-date=December 24, 2018|url-status=live}}</ref> HDDs were used in 70% of the desktop and notebook computers produced in 2016, and SSDs were used in 30%.<ref name="Coughlin_2016June">{{cite magazine|title= 3D NAND Enables Larger Consumer SSDs|url=https://www.forbes.com/sites/tomcoughlin/2016/06/07/3d-nand-enables-larger-consumer-ssds/#60d25bea579b|magazine=forbes.com|first1=Tom|last1=Coughlin|date=June 7, 2016|access-date=July 4, 2016|archive-url=https://web.archive.org/web/20160616092553/http://www.forbes.com/sites/tomcoughlin/2016/06/07/3d-nand-enables-larger-consumer-ssds/#60d25bea579b|archive-date=June 16, 2016|url-status=live}}</ref> In 2025 HDDs are not found in laptops, if not very rarely, and most desktops come with an SSD only, though some still are configured with an SSD and HDD, or rarely with an HDD only.{{citation needed|date=April 2025}}

The market for silicon-based flash memory (NAND) chips, used in SSDs and other applications, is growing faster than for HDDs. Worldwide NAND revenue grew 16% per year from $22 billion to $57 billion during 2011–2017, while production grew 45% per year from 19 exabytes to 175 exabytes.<ref name="IBM Fontana 2018" >{{cite web |url=https://storageconference.us/2018/Presentations/Fontana.pdf |title= A Ten Year (2008-2017) Storage Landscape LTO Tape Media, HDD, NAND |first1= Gary M. |last1= Decad |author2= Robert E. Fontana Jr. |date= May 15, 2018 |access-date= November 23, 2019 }}</ref>

== See also ==
{{Portal|Electronics}}

{{Div col|colwidth=18em}}
* [[Automatic acoustic management]]
* [[Cleanroom]]
* [[Click of death]]
* [[Comparison of disk encryption software]]
* [[Data erasure]]
* [[Drive mapping]]
* [[Error recovery control]]
* [[Hard disk drive performance characteristics]]
* [[Hybrid drive]]
* [[Microdrive]]
* [[Network drive]] (file server, [[shared resource]])
* [[Object storage]]
* [[Write precompensation]]
{{div col end}}

== Notes ==
{{Notelist|40em}}

== References ==
{{reflist|30em|refs=
<ref name="IBM350">{{cite web | url=http://www-03.ibm.com/ibm/history/exhibits/storage/storage_350.html | title=IBM Archives: IBM 350 disk storage unit | access-date=October 19, 2012 | date=January 23, 2003 | archive-url=https://web.archive.org/web/20080531032031/http://www-03.ibm.com/ibm/history/exhibits/storage/storage_350.html | archive-date=May 31, 2008 | url-status=dead }}</ref>
<ref name=headcrash>{{cite web | url = http://www.escotal.com/harddrive.html | title = Hard Drives | website = escotal.com | access-date = July 16, 2011 | archive-url = https://web.archive.org/web/20110903080111/http://www.escotal.com/harddrive.html | archive-date = September 3, 2011 | url-status = live }}</ref>
<ref name="AutoMK-6">{{cite web | url = http://www.data-master.com/HeadCrash-explain-hard-disk-drive-fail_Q18.html | title = What is a "head-crash" & how can it result in permanent loss of my hard drive data? | website = data-master.com | access-date = July 16, 2011 | archive-url = https://web.archive.org/web/20110708233813/http://www.data-master.com/HeadCrash-explain-hard-disk-drive-fail_Q18.html | archive-date = July 8, 2011 | url-status = live }}</ref>
<ref name="AutoMK-7">{{cite web | url = http://www.hard-drive-help.com/technology.html | title = Hard Drive Help | website = hardrivehelp.com | access-date = July 16, 2011 | url-status = dead | archive-url = https://web.archive.org/web/20110903011543/http://www.hard-drive-help.com/technology.html | archive-date = September 3, 2011 | df = mdy-all }}</ref>
<ref name="AutoMK-8">{{cite web | url=https://hypertextbook.com/facts/2001/JuliaSherlis.shtml
| title=Thickness of a piece of paper
| first=Juliya
| last=Sherlis
| year=2001
| website=The Physics Factbook
| editor-last=Elert
| editor-first=Glenn |access-date=July 9, 2011 |archive-url=https://web.archive.org/web/20170608160141/http://hypertextbook.com/facts/2001/JuliaSherlis.shtml |archive-date=June 8, 2017 |url-status=live }}</ref>
<ref name="AutoMK-9">[http://www.matesy.de/index.php?option=com_content&view=article&id=84&Itemid=80&lang=en CMOS-MagView] {{Webarchive|url=https://web.archive.org/web/20120113080808/http://www.matesy.de/index.php?option=com_content&view=article&id=84&Itemid=80&lang=en |date=January 13, 2012 }} is an instrument that visualizes magnetic field structures and strengths.</ref>
<ref name="AutoMK-10">{{cite web | url = http://www.hitachigst.com/tech/techlib.nsf/techdocs/33c6a76df9338b0a86256d34007054c0/%24file/why7200mobilehdds.pdf | archive-url=https://web.archive.org/web/20120419091304/http://www.hitachigst.com/tech/techlib.nsf/techdocs/33c6a76df9338b0a86256d34007054c0/$file/why7200mobilehdds.pdf | title = Why 7,200&nbsp;RPM Mobile Hard Disk Drives? | first = Walker C. | last = Blount | date = November 2007 | access-date = July 17, 2011 | archive-date=April 19, 2012}}</ref>
<ref name="AutoMK-15">{{cite web | url = http://www.americanscientist.org/issues/pub/terabyte-territory | title = Terabyte Territory | publisher = American Scientist | first = Brian | last = Hayes | volume = 90 | issue = No 3 (May–June 2002) | page = 212 | access-date = September 20, 2014 | archive-url = https://web.archive.org/web/20140708221725/http://www.americanscientist.org/issues/pub/terabyte-territory | archive-date = July 8, 2014 | url-status = live }}</ref>
<ref name="AutoMK-16">{{cite web | url = http://www.toshiba.co.jp/about/press/2004_12/pr1401.htm | title = Press Releases December&nbsp;14, 2004 | publisher = Toshiba | access-date = March 13, 2009 | archive-url = https://web.archive.org/web/20090414064812/http://www.toshiba.co.jp/about/press/2004_12/pr1401.htm | archive-date = April 14, 2009 | url-status = live }}</ref>
<ref name="AutoMK-17">{{cite web | url = http://www.seagate.com/www/en-us/products/laptops/momentus/ | title = Seagate Momentus 2½" HDDs per webpage January 2008 | website = Seagate.com | date = October 24, 2008 | access-date = March 13, 2009 | archive-url = https://web.archive.org/web/20090311133828/http://www.seagate.com/www/en-us/products/laptops/momentus | archive-date = March 11, 2009 | url-status = live }}</ref>
<ref name="AutoMK-18">{{cite web | url = http://www.seagate.com/www/en-us/products/desktops/barracuda_hard_drives/ | title = Seagate Barracuda 3½" HDDs per webpage January 2008 | website = Seagate.com | access-date = March 13, 2009 | archive-url = https://web.archive.org/web/20090314041031/http://www.seagate.com/www/en-us/products/desktops/barracuda_hard_drives/ | archive-date = March 14, 2009 | url-status = live }}</ref>
<ref name="AutoMK-19">{{cite web | url = http://www.wdc.com/en/company/investor/q108remarks.asp | title = Western Digital Scorpio 2½" and Greenpower 3½" HDDs per quarterly conference, July 2007 | website = Wdc.com | access-date = March 13, 2009 | url-status = dead | archive-url = https://web.archive.org/web/20090316015932/http://www.wdc.com/en/company/investor/q108remarks.asp | archive-date = March 16, 2009 | df = mdy-all }}</ref>
<ref name="AutoMK-19a">{{cite journal | url = https://www.researchgate.net/publication/222074314 | title = Exchange spring recording media for areal densities up to 10Tbit/in2 |journal= J. Magn. Mag. Mat. |author=D. Suess |display-authors=etal |year=2004 }}</ref>
<ref name="AutoMK-19b">{{cite journal |title=Composite media for perpendicular magnetic recording |journal=IEEE Trans. Mag. Mat. |volume=41 |issue=2 |pages=537–542 |author=R. Victora |display-authors=etal |year=2005 |doi=10.1109/TMAG.2004.838075 |bibcode=2005ITM....41..537V |s2cid=29531529 }}</ref>
<ref name="AutoMK-21">[https://web.archive.org/web/20181010031401/https://www.hgst.com/sites/default/files/resources/IDRC_WP_final.pdf "Iterative Detection Read Channel Technology in Hard Disk Drives"], Hitachi</ref>
<ref name="Coughlin">{{cite web | url = https://ewh.ieee.org/r6/scv/mag/MtgSum/Meeting2012_06_Presentation.pdf | title = Years of Destiny: HDD Capital Spending and Technology Developments from 2012–2016 | author1 = Coughlin, Thomas | author2 = Grochowski, Edward | date = June 19, 2012 | publisher = IEEE Santa Clara Valley Magnetics Society | access-date = October 9, 2012 | archive-url = https://web.archive.org/web/20130302180119/http://ewh.ieee.org/r6/scv/mag/MtgSum/Meeting2012_06_Presentation.pdf | archive-date = March 2, 2013 | url-status = live }}</ref>
<ref name="AutoMK-22">{{cite magazine | url = https://www.pcmag.com/article2/0,2817,2368023,00.asp | title = Will Toshiba's Bit-Patterned Drives Change the HDD Landscape? | magazine = PC Magazine | date = August 19, 2010 | access-date = August 21, 2010 | archive-url = https://web.archive.org/web/20100822220804/http://www.pcmag.com/article2/0,2817,2368023,00.asp | archive-date = August 22, 2010 | url-status = live }}</ref>
<ref name="SAS">{{citation | title = Information technology&nbsp;– Serial Attached SCSI&nbsp;– 2 (SAS-2), INCITS 457 Draft 2, May 8, 2009, chapter 4.1 Direct-access block device type model overview | quote = The LBAs on a logical unit shall begin with zero and shall be contiguous up to the last logical block on the logical unit.}}</ref>
<ref name="SATA">ISO/IEC 791D:1994, AT Attachment Interface for Disk Drives (ATA-1), section 7.1.2</ref>
<ref name="AutoMK-38">{{cite web | url = http://blogs.netapp.com/efficiency/2009/08/measuring-storage-efficiency-part-ii-taxes.html | archive-url=https://web.archive.org/web/20110720131225/http://blogs.netapp.com/efficiency/2009/08/measuring-storage-efficiency-part-ii-taxes.html| title = How to Measure Storage Efficiency&nbsp;– Part II&nbsp;– Taxes | publisher = Blogs.netapp.com | date=August 14, 2009 |access-date=April 26, 2012 | archive-date=July 20, 2011}}</ref>
<ref name="SeagateStorageGuide">{{cite web | url = https://www.seagate.com/files/www-content/partners/my%20spp%20dashboard/learn/en-us/docs/amer-stx-storage-solutions-guide-Oct-2012.pdf | archive-url=https://web.archive.org/web/20130620212124/http://www.seagate.com/files/www-content/partners/my%20spp%20dashboard/learn/en-us/docs/amer-stx-storage-solutions-guide-Oct-2012.pdf | title = Storage Solutions Guide | publisher = Seagate | date = October 2012 | access-date = June 8, 2013 | archive-date=June 20, 2013}}</ref>
<ref name="WD2">{{cite web |url=http://sdd.toshiba.com/techdocs/MKxx33GSG_MK1235GSL_r1.pdf |title=MKxx33GSG MK1235GSL r1 |publisher=Toshiba |access-date=January 7, 2013 |url-status=dead |archive-url=https://web.archive.org/web/20091122075109/http://sdd.toshiba.com/techdocs/MKxx33GSG_MK1235GSL_r1.pdf |archive-date=November 22, 2009  }}</ref>
<ref name="AutoMK-26">{{cite web | url = http://www-03.ibm.com/ibm/history/exhibits/650/650_pr2.html | title = 650 RAMAC announcement | date = January 23, 2003 | access-date = May 23, 2011 | archive-url = https://web.archive.org/web/20110605130810/http://www-03.ibm.com/ibm/history/exhibits/650/650_pr2.html | archive-date = June 5, 2011 | url-status = dead }}</ref>
<ref name="AutoMK-27">Mulvany, R.B., "Engineering Design of a Disk Storage Facility with Data Modules". IBM JRD, November 1974</ref>
<ref name="AutoMK-28">Introduction to IBM Direct Access Storage Devices, M. Bohl, IBM publication SR20-4738. 1981.</ref>
<ref name="AutoMK-29">[http://www.bitsavers.org/pdf/cdc/discs/brochures/ProductLine_Oct74.pdf CDC Product Line Card] {{Webarchive|url=https://web.archive.org/web/20110605112046/http://www.bitsavers.org/pdf/cdc/discs/brochures/ProductLine_Oct74.pdf |date=June 5, 2011 }}, October 1974.</ref>
<ref name="Apple">{{cite web | url = https://support.apple.com/en-us/HT201402 | title = How OS X and iOS report storage capacity | author = Apple Support Team | publisher = Apple, Inc. | access-date = March 15, 2015 | archive-url = https://web.archive.org/web/20150402170043/https://support.apple.com/en-us/HT201402 | archive-date = April 2, 2015 | url-status = live }}</ref>
<ref name="WDSettle">{{cite news | url = https://www.foxnews.com/story/western-digital-settles-hard-drive-capacity-lawsuit | title = Western Digital Settles Hard-Drive Capacity Lawsuit, Associated Press June 28, 2006 | publisher = Fox News | date = March 22, 2001 | access-date = May 24, 2019 | archive-url = https://web.archive.org/web/20190524225524/https://www.foxnews.com/story/western-digital-settles-hard-drive-capacity-lawsuit | archive-date = May 24, 2019 | url-status = live }}</ref>
<ref name="AutoMK-30">{{cite web | first = Phil | last = Cogar | url = http://www.bit-tech.net/news/bits/2007/10/26/seagate_lawsuit_concludes_settlement_announced/1 | title = Seagate lawsuit concludes, settlement announced | publisher = Bit-tech.net | date = October 26, 2007 | access-date = April 26, 2012 | archive-url = https://web.archive.org/web/20120320020954/http://www.bit-tech.net/news/bits/2007/10/26/seagate_lawsuit_concludes_settlement_announced/1 | archive-date = March 20, 2012 | url-status = live }}</ref>
<ref name="AutoMK-31">{{cite web | url = http://www.xtremesystems.org/forums/showthread.php?t=93512 | title = Western Digital&nbsp;– Notice of Class Action Settlement email | publisher=Xtremesystems.org | access-date = April 26, 2012 }}</ref>
<ref name="AutoMK-39">Emerson W. Pugh, Lyle R. Johnson, John H. Palmer ''IBM's 360 and early 370 systems'' MIT Press, 1991 {{ISBN|0-262-16123-0}}, page 266.</ref>
<ref name="AutoMK-51">[http://www.eetasia.com/ART_8800474064_499486_NT_3335be30.HTM Flash price fall shakes HDD market], EETimes Asia, August 1, 2007. {{Webarchive|url=https://web.archive.org/web/20080201232357/http://www.eetasia.com/ART_8800474064_499486_NT_3335be30.HTM |date=February 1, 2008 }}</ref>
<ref name="AutoMK-52">[http://www.samsung.com/global/business/hdd/newsView.do?b2b_bbs_msg_id=143 In 2008 Samsung] {{Webarchive|url=https://web.archive.org/web/20110616035929/http://www.samsung.com/global/business/hdd/newsView.do?b2b_bbs_msg_id=143 |date=June 16, 2011 }} introduced the 1.3-inch SpinPoint A1 HDD but by March 2009 the family was listed as [http://www.samsung.com/global/business/hdd/products/Product_EOLProducts.html End Of Life Products] and new 1.3-inch models were not available in this size. {{Webarchive|url=https://web.archive.org/web/20090211054619/http://www.samsung.com/global/business/hdd/products/Product_EOLProducts.html |date=February 11, 2009 }}</ref>
<!--ref name="AutoMK-66">{{cite web | url = http://www.wdc.com/en/company/glossaryofterms/wdglossarycontent.asp | title = Western Digital definition of ''Average Access Time'' | publisher = Wdc.com | date = July 1, 2006 | access-date = April 26, 2012 | url-status = dead | archive-url = https://web.archive.org/web/20120427074538/http://www.wdc.com/en/company/glossaryofterms/wdglossarycontent.asp | archive-date = April 27, 2012 | df = mdy-all }}</ref-->
<ref name="itworld-2001-04-18">{{cite news |title=How to defrag |first=Dave |last=Kearns |date=April 18, 2001 |newspaper=ITWorld |url=http://www.itworld.com/NWW01041100636262 |access-date=November 26, 2010 |archive-url=https://web.archive.org/web/20100220000835/http://www.itworld.com/NWW01041100636262 |archive-date=February 20, 2010 |url-status=live }}</ref>
<ref name="AutoMK-67">{{cite news | title = Turning Off Disk Defragmenter May Solve a Sluggish PC | first = Rick | last = Broida | newspaper = PCWorld | date = April 10, 2009 | url = http://www.pcworld.com/article/162955/turning_off_disk_defragmenter_may_solve_a_sluggish_pc.html | access-date = November 26, 2010 | archive-url = https://web.archive.org/web/20101108094534/http://www.pcworld.com/article/162955/turning_off_disk_defragmenter_may_solve_a_sluggish_pc.html | archive-date = November 8, 2010 | url-status = live }}</ref>
<ref name="AutoMK-71">{{cite web|url=http://www.seagate.com/www/en-us/support/before_you_buy/speed_considerations|archive-url=https://web.archive.org/web/20110210155634/http://www.seagate.com/www/en-us/support/before_you_buy/speed_considerations|title=Speed Considerations|publisher=Seagate|access-date=January 22, 2011|archive-date=February 10, 2011}}</ref>
<ref name="AutoMK-79">{{cite web|url=http://gaussianwaves.blogspot.com/2008/06/reed-solomon-codes-introduction.html|archive-url=https://web.archive.org/web/20110708040145/http://gaussianwaves.blogspot.com/2008/06/reed-solomon-codes-introduction.html|archive-date=July 8, 2011|title=Reed Solomon Codes&nbsp;– Introduction}}</ref>
<ref name="AutoMK-80">{{cite web |url=https://www.karlstechnology.com/blog/hard-drive-air-filters/ |title=Micro House PC Hardware Library Volume I: Hard Drives |first=Scott |last=Mueler |publisher=Macmillan Computer Publishing |access-date=May 24, 2019 |date=February 24, 2019 |archive-url=https://web.archive.org/web/20190524232939/https://www.karlstechnology.com/blog/hard-drive-air-filters/ |archive-date=May 24, 2019 |url-status=live }}</ref>
<ref name="AutoMK-81">{{cite web | url = http://apex.aero/Portals/0/harddisk.pdf | title = Ruggedized Disk Drives for Commercial Airborne Computer Systems |first1=Robert G. |last1=Kaseta | url-status = dead | archive-url = https://web.archive.org/web/20120504043804/http://apex.aero/Portals/0/harddisk.pdf | archive-date = May 4, 2012 | df = mdy-all }}</ref>
<ref name="AutoMK-98">{{cite web | url = https://biometricsecurityproducts.org/biometric-safe/back-up-your-important-data-to-external-hard-disk-drive.html |archive-url=https://web.archive.org/web/20120525025411/http://biometricsecurityproducts.org/biometric-safe/back-up-your-important-data-to-external-hard-disk-drive.html| title = Back Up Your Important Data to External Hard disk drive &#124; Biometric Safe &#124; Info and Products Reviews about Biometric Security Device&nbsp;– | publisher = Biometricsecurityproducts.org | date = July 26, 2011 | access-date = April 26, 2012|archive-date=May 25, 2012}}</ref>
<ref name="AutoMK-99">{{cite web |url=https://www.intel.com/content/dam/support/us/en/documents/server-products/Enterprise_vs_Desktop_HDDs_2.0.pdf |title=Enterprise-class versus Desktop class Hard Drives |publisher=Intel |access-date=September 25, 2013 |archive-url=https://web.archive.org/web/20160803170316/http://www.intel.com/content/dam/support/us/en/documents/server-products/Enterprise_vs_Desktop_HDDs_2.0.pdf |archive-date=August 3, 2016 |url-status=live }}</ref>
<ref name="erwpnw">{{Cite web |url=http://www.seagate.com/docs/pdf/marketing/po_cheetah_15k_5.pdf |title=Seagate Cheetah 15K.5 Data Sheet |access-date=December 19, 2013 |archive-url=https://web.archive.org/web/20131228051734/http://www.seagate.com/docs/pdf/marketing/po_cheetah_15k_5.pdf |archive-date=December 28, 2013 |url-status=live }}</ref>
}}

== Further reading ==
* {{cite journal |last1=Kheong Chn |first1=Sann |title=An Introduction to the HDD, modelling, detection and decoding for magnetic recording channels |date=2005 |journal=The Eleventh Advanced International Conference on Telecommunications |url=https://www.iaria.org/conferences2015/filesAICT15/AnIntroductionToHDDModellingDetectionAndDecoding.pdf |access-date=10 January 2020}}
* {{cite book|last=Messmer|first=Hans-Peter|year=2001|title=The Indispensable PC Hardware Book|edition=4th|publisher=Addison-Wesley|isbn=978-0-201-59616-8}}
* {{cite book|last=Mueller|first=Scott|author-link=Scott Mueller|year=2011|title=Upgrading and Repairing PCs|edition=20th|publisher=Que|isbn=978-0-7897-4710-5}}

== External links ==
{{Sister project links|commons=Hard disk|b=Minimizing hard disk drive failure and data loss|wikt=hard disk drive|q=no|s=no|v=Hard disk drive}}
{{External links|section|date=July 2020}}
* [https://smarthdd.com/help.htm Hard Disk Drives Encyclopedia]
* [https://www.youtube.com/watch?v=Wiy_eHdj8kg Video showing an opened HD working]
* [https://web.archive.org/web/20121014191631/http://faculty.plattsburgh.edu/jan.plaza/teaching/papers/seektime.html Average seek time of a computer disk] (archived)
* [https://www.pcworld.com/article/127105/article.html Timeline: 50 Years of Hard Drives]. {{Webarchive|url=https://web.archive.org/web/20131006003652/http://www.pcworld.com/article/127105/article.html |date=October 6, 2013 }}.
* [https://hddscan.com/doc/HDD_Tracks_and_Zones.html HDD from inside: Tracks and Zones. How hard it can be?]
* [https://spritesmods.com/?art=hddhack Hard disk hacking]{{snd}} firmware modifications, in eight parts, going as far as booting a Linux kernel on an ordinary HDD controller board
* [https://web.archive.org/web/20150226053423/http://www.recover.co.il/SA-cover/SA-cover.pdf Hiding Data in Hard Drive's Service Areas] (PDF), February 14, 2013, by Ariel Berkman (archived)
* [https://www.wdc.com/wdproducts/library/other/2579-001079.pdf Rotary Acceleration Feed Forward (RAFF) Information Sheet] (PDF), [[Western Digital]], January 2013
* [http://www.seagate.com/files/docs/pdf/en-GB/whitepaper/tp608-powerchoice-tech-provides-gb.pdf PowerChoice Technology for Hard Disk Drive Power Savings and Flexibility] (PDF), [[Seagate Technology]], March 2010
* [http://www.hgst.com/science-of-storage/emerging-technologies/shingled-magnetic-recording Shingled Magnetic Recording (SMR)], [[HGST]], Inc., 2015
* [http://www.hgst.com/science-of-storage/emerging-technologies/helium-technology The Road to Helium], HGST, Inc., 2015
* [https://www.nature.com/articles/ncomms13406 Research paper about perspective usage of magnetic photoconductors in magneto-optical data storage.]

{{Data storage}}
{{Solid-state drive}}
{{Hard disk drive manufacturers}}
{{IBM}}
{{Basic computer components}}
{{Disk images}}
{{Commodore disk drives}}
{{RAID}}
{{Authority control}}

[[Category:20th-century inventions]]
[[Category:American inventions]]
[[Category:Articles containing video clips]]
[[Category:Computer data storage]]
[[Category:Computer storage devices]]
[[Category:Hard disk drives| ]]
[[Category:Rotating disc computer storage media]]