Editing Tape library

{{short description|Storage device containing a robot which automatically loads tapes into tape drives}}
In [[computer storage]], a '''tape library''' is a physical area that holds [[magnetic tape data storage|magnetic data tapes]].  In an earlier era, tape libraries were maintained by people known as [[tape librarian]]s and [[computer operator]]s and the proper operation of the library was crucial to the running of [[batch processing]] jobs.  Although tape libraries of this era were not automated, the use of [[tape management system]] software could assist in running them.

Subsequently, tape libraries became physically automated, and as such are sometimes called a '''tape silo''', '''tape robot''', or '''tape jukebox'''.  These are a storage devices that contain one or more [[tape drive]]s, a number of slots to hold [[Magnetic tape data storage#Cartridges and cassettes|tape cartridges]], a [[barcode]] reader to identify tape cartridges, and an automated method for loading tapes (a robot). Such solutions are mostly used for [[backup]]s and for [[digital archiving]]. Additionally, the area where tapes that are not currently in a silo are stored is also called a tape library. One of the earliest examples was the [[IBM 3850]] Mass Storage System (MSS), announced in 1974.  

In either era, tape libraries can contain millions of tapes.

== Manual era ==
[[File:NDOC magnetic tape library.jpg|thumb|A manual magnetic tape library, common in the 1960s and 1970s. Rolling carts are used by staff to transfer tapes between the racks in the library and the computer room where the tape drives reside.]]

=== Tapes and batch processing ===
In the [[mainframe computer]] era, especially the [[IBM mainframe]], the most common format in use was the [[9-track tape]].<ref name="popkin-pike"/>
Some large application systems could require scores of different tapes as part of their batch job runs.<ref name="mccracken"/>

In the data processing applications of the era, the [[master file]]s for such things as employee payroll information, supplies and stores inventory, or customer accounts were typically kept on tape.<ref name="mcquillen"/><ref name="stern-stern"/>  Batch jobs to update these master files would take the existing tape master file as input and write out a new tape master file as output.<ref name="ashley-fernandez">{{cite book | title=Job Control Language: A Self-Teaching Guide | first1=Ruth | last1=Ashley | first2=Judi N. | last2=Fernandez | publisher=John Wiley & Sons | location=New York | date=1978 | isbn=0-471-03205-0 | page=43 }}</ref> In addition, the set of update transactions themselves might constitute a second input tape.<ref name="mcquillen">{{cite book | title=System/360–370 Assembler Language (OS) | first=Kevin | last=McQuillen | publisher=Mike Murach & Associates | location=Fresno, California | date=1975 | lccn=74-29645 | page=302}}</ref> The master file output of one update job would then be the master file input to the next time the job is run, perhaps a day, a week, or a month later.<ref name="popkin-pike"/> The tapes representing a few past iterations of a master file would typically be retained, in case a problem with the latest version were to be discovered and the job had to be rerun.<ref name="popkin-pike"/>

=== Role of tape libraries and librarians ===
[[Image:Tape Retention Scratch Control triplicate form.jpg|thumb|upright=0.7|left|Tape Retention / Scratch Control form, in triplicate]]

Mainframe computer installations often had a separate room, the tape library, to house their racks and cabinets of tapes.<ref name="popkin-pike"/> The typical workflow for running a batch job was to go into the library, pull certain tapes off the racks there and load them onto a rolling cart, move the cart into the computer area, mount the tapes onto tape drives for a production run, take the tapes off the drives when the run was over, move the cart back to the library, and put the tapes back on the library racks.  Such tape libraries existed at most computer installations.<ref name="conway-gries"/>

Even a modestly sized computer installation could have hundreds of tapes,<ref name="stern-stern">{{cite book | title=Structured COBOL Programming | first1=Nancy | last1=Stern | first2=Robert A. | last2=Stern | publisher=John Wiley & Sons | location=New York | edition=3rd | date=1980 | isbn=0-471-04913-1 | pages=494, 496, 498&ndash;499 }}</ref>
and library sizes of several thousand reels of tapes were commonplace.<ref name="conway-gries"/> 
And they could be much larger:  by the mid-1970s, the [[U.S. Census Bureau]] and [[NASA]] each had tape libraries with around one million tape reels in them.<ref name="mccracken"/>
The person in charge of all this was typically called the [[tape librarian]].<ref name="popkin-pike"/><ref name="stern-stern"/> 

In this era, there were no automated tape delivery and mounting systems, and so this action had to be done by [[computer operator]]s.<ref name="conway-gries">{{cite book |last1=Conway |first1=Richard  |last2=Gries |first2=David |year=1973 |title=An Introduction to Programming: A Structured Approach using PL/1 and PL/C |publisher=Winthrop |location=Cambridge, Massachusetts | pages=333&ndash;334 }}</ref>
These people were the ones responsible for mounting tapes onto [[tape drive]]s as part of running a job.<ref name="popkin-pike"/>
Even careful computer operators could sometimes mount the wrong tape as input to a job or present the reels of a multi-tape dataset out of order.<ref name="mccracken">{{cite book| title=A Simplified Guide to Structured COBOL Programming| publisher=John Wiley & Sons|location=New York| year=1976 | isbn=0-471-58284-0  | author-first=Daniel D. | author-last=McCracken | pages=259, 264 }}</ref> Overwriting a tape that was meant to be preserved was another potential mistake.<ref name="stern-stern"/>

It was the tape librarian's responsibility to set up procedures for the handling of tapes to minimize the chances of errors taking place.<ref name="stern-stern"/>
As one book of the era wrote, "keeping track of the whereabouts of the tapes is a formidable and responsible job."<ref name="popkin-pike">{{cite book | title=Introduction to Data Processing | author-first=Gary S. | author-last=Popkin | author2-first=Arthur H. | author2-last=Pike | publisher=Houghton Mifflin Company | location=Boston | year=1977 | isbn=0-395-20628-6 | pages=149&ndash;151, 260&ndash;263 }}</ref>

=== Supporting software ===
{{seealso|Tape management system}}

Tape management systems of this era were software packages whose purpose was to help facilitate tape library operations and management. They kept track of data sets on tape, and produced reports indicating whether a data set should be retained on, or could be scratched from, a tape; they aided in the setup and running of scheduled production jobs, through such things as tape pull lists and pre-printed external gummed tape labels; and they kept track of the physical inventory of tape reels. The most popular of these packages was UCC-1 from [[University Computing Company]],<ref name="cw-jan1977">{{cite news | url=https://books.google.com/books?id=XEToBl0e2NIC&dq=%22ucc%22+%22honor+roll%22&pg=PA23 | title=Users Put 38 Packages on Honor Roll | first=Don | last=Leavitt | newspaper=Computerworld | date=January 17, 1977 | page=23 }}</ref> a product that was also known as the Tape Management System.<ref>{{cite news | url=https://books.google.com/books?id=JF4iQwmmCwYC&dq=%22university+computing%22+%22ucc-1%22&pg=PA35 | title=UCC-1 Tape Management Updated with Release 4.7 | newspaper=Computerworld | date=July 4, 1983 | page=35 }}</ref> It made several appearances on [[Datapro Research Corporation]]'s Software Honor Roll.<ref name="cw-jan1977"/> Another was Valu-Lib from [[Value Computing, Inc.]],<ref>{{cite news | url=https://books.google.com/books?id=esYk9fkCBWgC&dq=valu-lib+tape&pg=PA15
 | title='Valu-Lib' Can Run Tape Library, Can Interface With Scheduler | newspaper=Computerworld | date=May 16, 1973 | page=15 }}</ref><ref>{{cite news | url=https://books.google.com/books?id=aDCH6OPGmh8C&dq=valu-lib+tape&pg=PA32 | title='Valu Lib' Update Released For IBM 4300s, Series/36 | newspaper=Computerworld | date=December 19, 1983 | page=32 }}</ref>
and a third was TLMS II from [[Capex Corporation]].<ref>{{cite news | url=https://books.google.com/books?id=K1VUAAAAMAAJ&q=%22capex%22+%22tape+library%22 | title=uncertain | magazine=Infosystems | publisher=Hitchcock Publishing Company | year=1980 | volume=uncertain| page=90 | access-date=February 22, 2023 | archive-date=March 25, 2023 | archive-url=https://web.archive.org/web/20230325204700/https://books.google.com/books?id=K1VUAAAAMAAJ&q=%22capex%22+%22tape+library%22 | url-status=live }}</ref>

As use of the mainframe continued on into the following century, tape library management, both manual and automatic, was one element of the offerings of the [[Data Facility Storage Management Subsystem (MVS)]] from IBM.<ref>{{cite news | url=https://www.ibm.com/docs/en/zos/3.1.0?topic=libraries-introduction-tape-library-management | title=Introduction to tape library management | publisher=IBM | date=April 5, 2023 | access-date=November 1, 2023}}</ref>

== Automated era ==
[[File:StorageTek Powderhorn tape library.jpg|thumb|upright=0.7|Large [[StorageTek]] Powderhorn tape library, showing tape cartridges with barcodes packed on shelves in the front and a robot arm moving in the back]]
[[File:ADIC Scalar 100 tape library.jpg|thumb|upright=0.6|Small [[Advanced Digital Information Corporation|ADIC]] Scalar 100 tape library, showing a robot visible on the bottom with two [[IBM LTO2]] tape drives behind it]]

=== Design ===

Physically automated tape library devices can store immense amounts of data, ranging from 20 [[terabyte]]s<ref>{{Cite web|url=http://h18006.www1.hp.com/products/storageworks/msl2024/index.html|title=HP StorageWorks MSL2024 Tape Library - overview|date=March 18, 2006|access-date=June 19, 2018|archive-date=March 18, 2006|archive-url=https://web.archive.org/web/20060318182700/http://h18006.www1.hp.com/products/storageworks/msl2024/index.html|url-status=bot: unknown}}</ref> up to 2.1 [[exabyte]]s of data<ref>[[Oracle Corporation|Oracle]] {{cite web|url=http://www.oracle.com/us/products/servers-storage/storage/tape-storage/sl8500-modular-library-system/overview/index.html|title=StorageTek SL8500 Modular Library System}}</ref> as of 2016.  Such capacity is multiple thousand times that of a typical [[hard drive]] and well in excess of what is capable with [[network attached storage]]. Typical entry-level solutions cost around $10,000 USD,<ref>[http://h71016.www7.hp.com/dstore/ctoBases.asp?jumpid=re_NSS_msl2024&ProductLineId=450&FamilyId=2297 HP Small & Medium Business Online Store: HP StorageWorks MSL2024 Tape Libraries]</ref> while high-end solutions can start at as much as $200,000 USD<ref>[http://www.oracle.com/us/products/servers-storage/storage/tape-storage/sl8500-modular-library-system/overview/index.html]. Cites cost as "From $195,830. (US)"</ref> and cost well in excess of $1 million for a fully expanded and configured library.

For large data-storage, they are a cost-effective solution, with cost per gigabyte as low as 2 cents USD.<ref>{{cite web|url=https://www.forbes.com/sites/tomcoughlin/2016/07/24/the-costs-of-storage/#29cedc803239|title=The Costs Of Storage|work=Forbes}}</ref> The tradeoff for their larger capacity is their slower access time, which usually involves mechanical manipulation of tapes.  Access to data in a library takes from several seconds to several minutes.

Because of their slow sequential access and huge capacity, tape libraries are primarily used for [[backup]]s and as the final stage of [[digital archiving]]. A typical application of the latter would be an organization's extensive transaction record for legal or auditing purposes. Another example is [[hierarchical storage management]] (HSM), in which tape library is used to hold rarely used files from [[file system]]s.

=== Software support ===

There are several large-scale library-management packages available commercially. Open-source implementations include [[Advanced Maryland Automatic Network Disk Archiver|AMANDA]], [[Bacula]], and the minimal mtx program.

=== Barcode labels ===

Tape libraries commonly have the capability of optically scanning [[barcode]] labels which are attached to each tape, allowing them to automatically maintain an inventory of which tapes are where within the library. Preprinted barcode labels are commercially available or custom labels may be generated using commercial or free software. The barcode label is frequently part of the [[tape label]], information recorded at the beginning of the medium to uniquely identify the tape.

=== Autoloaders ===
[[File:Powervault 124T autoloader.jpg|thumb|left|Dell PowerVault 124T Autoloader]]

Smaller tape libraries with only one drive are known as autoloaders.<ref>{{cite web|url=http://www.snia.org/education/dictionary/t|quote=tape autoloader...[Storage System] A tape device that provides automated access to multiple tape cartridges, typically via a single tape drive.|title=SNIA Dictionary|publisher=Storage Network Industry Association|accessdate=2010-01-30}}</ref> The term ''autoloader'' is also sometimes used synonymously with ''stacker'',<ref>{{cite web|url=http://www.sunhelp.org/faq/backup.html#korea10|quote=What is a stacker (autoloader) vs a jukebox?|title=Ten common backup/restore related questions|publisher=Sun Microsystems, Inc|accessdate=2010-01-30}}</ref> a device in which the media are loaded necessarily in a sequential manner.<ref>{{cite web|url=http://www.snia.org/education/dictionary/m|quote=media stacker...[Data Recovery] A robotic media handler in which media must be moved sequentially by the robot.|title=SNIA Dictionary|publisher=Storage Network Industry Association|accessdate=2010-01-30}}</ref>

Other types of autoloaders may operate with [[optical disc]]s (such as [[compact disc]]s or [[DVD]]s) or [[floppy disk]]s{{citation needed|date=October 2011}}.

==See also==
{{commons category|Tape libraries}}
*[[Optical jukebox]]

== References ==
{{reflist}}

[[Category:Tape-based computer storage|Library]]
[[Category:Handling robots| ]]