Talk:Magnetic-tape data storage

This  level-5 vital article is rated C-class on Wikipedia's content assessment scale. It is of interest to the following WikiProjects:

Technology This article is within the scope of WikiProject Technology, a collaborative effort to improve the coverage of technology on Wikipedia. If you would like to participate, please visit the project page, where you can join the discussion and see a list of open tasks.TechnologyWikipedia:WikiProject TechnologyTemplate:WikiProject TechnologyTechnology Computing Mid‑importance This article is within the scope of WikiProject Computing, a collaborative effort to improve the coverage of computers, computing, and information technology on Wikipedia. If you would like to participate, please visit the project page, where you can join the discussion and see a list of open tasks.ComputingWikipedia:WikiProject ComputingTemplate:WikiProject ComputingComputing Mid This article has been rated as Mid-importance on the project's importance scale.

"...tape is a sequential access medium while disk is a random access medium..."

Please don't call disk random access. It may have a much lower seek time than tape, but it is still very much a sequential storage format. —Preceding unsigned comment added by 81.105.12.50 (talk) 20:42, 4 April 2010 (UTC)[reply]

I would suggest the use of "Direct Access" over "Random Access".

Also, I think it is important to discuss the concept of "soft" formatted sequential versus "hard" formatted sequential access in this section. In soft formatted systems, over-writing an earlier block on the medium destroys all later blocks, or at least renders them unavailable. Hard formatted media, such as most (all?) disk drives and only a very few types of tapes, can be accessed sequentially and non-sequentially without disrupting access to earlier or later blocks on the medium. —Preceding unsigned comment added by 157.127.155.214 (talk) 00:38, 21 May 2010 (UTC)[reply]

Hi "K", I agree that sequential access is an important subject to cover for this topic. I moved the previous text to a subsection of 'Technical Details (and edited it) because the file system stuff is a little more technical than an average reader might be interested in. Let's put a general sentence or two in the opening section and put the more specific file layout details in the technical section. -- Austin Murphy 21:05, 30 March 2007 (UTC)[reply]

Gosh, I tought you've removed it completely. It slipped my attention that you've just moved the text, sorry. --Kubanczyk 12:50, 3 April 2007 (UTC)[reply]

A merge of content from the Data tape cartridge article was suggested, and I believe that it is a good idea - there really isn't anything in that article that is not already (or could not already be) covered in this one. However, I had trouble figuring out where, and if, the data from that article should be placed in this one. Here is the content from Data tape cartridge:

---

Information storage in year 1949 meant books, filing cabinets or punch cards. Tape cartridge was the stuff of science fiction. The limitations of punch cards were becoming obvious. The data storage problem as acres of filing cabinets filled with punched cards containing Social Security data echoed that a more compact means of storing data needed to be developed. The emergence of magnetic tape in 1952 by IBM was a big leap in terms of data storage. Throughout the 1950s and 60s, magnetic tape units offered successive improvement in data storage.

IBM produced the first automated data tape in 1974 bringing with it great innovations that in turn led the technology to a newer heights of data storage. One breakthrough was the introduction of 5.5 inch square data tape cartridge storing up to 200 MB data and executing at the rate of 3 MB/s requiring less than half of the floor space of an equivalent installation of its predecessor technology – magnetic tape units - from just a decade earlier.

Data tape cartridge was widely used in the early 70s and 80s for Data storage (a format of storing and backing up computer data on a storage device). The drawback is that tape drive must spend a considerable amount of time winding tape between reels to read any one particular piece of data. As a result, tape cartridges have very slow average seek times. However, it replaced many of the demerits of its predecessor technology (magnetic tape) with its high speed, non- bulkiness, low cost and stability. Now-a-days, we use the same technology of data tape cartridge in terms of 4 mm/8 mm data tapes, mini DV, VHS video etc.

References:

1. Storing the information age: 50 years of tape storage innovation: http://www-8.ibm.com/nz/media/downloads/IBMarticles/Storing_the_Information_Age.pdf 2. Wikipedia reference: http://en.wikipedia.org 3. Data tape category range by Sony.

---

I am unfamiliar with this subject. Is there any information contained above that should be placed in the Magnetic tape data storage article? I know that the article I've redirected had major tone issues, but is there anything factual that we should place into this article from it that isn't already present? Thanks for the input. *Vendetta* _{(whois talk edits)} 21:54, 5 April 2007 (UTC)[reply]

It would be good to have some comments (or a reference to, for instance, Digital_permanence) about shelf life of the media, possibly compared with other media (notably hard drives and flash). —The preceding unsigned comment was added by 206.168.13.209 (talk) 14:34, August 22, 2007 (UTC) The manufactures of Enterprise Tape Media warrant their tape for 30 years of self life. The question becomes will there be a drive to read it? The 3480 media began in 1984, and still can be read today.141.202.248.52 (talk) 17:50, 2 November 2010 (UTC)Keith141.202.248.52 (talk) 17:50, 2 November 2010 (UTC)[reply]

Honeywell VLDS (very large data store) c.1990 stored 5.2 GB on an S-VHS tape. Later spun off as Metrum.

"METRUM, established in 1954, is the former Test Instruments Division of Honeywell (NYSE:HON). In September 1990, this business, under the METRUM corporate name, was part of a corporate spin-off of several divisions of Honeywell into the newly formed Alliant Techsystems Inc. (NYSE:ATK). METRUM was acquired by Group Technologies in December 1992."

--Group Technologies' Metrum Subsidiary Sells Peripheral Products Business To Mountaingate

There is only one passing reference to Exabyte, which needs to be expanded. Exabyte did much to kill 9-track (certainly) as a backup medium since a 2GB cartridge could be left unattended instead of having operators changing tapes. DLT effectively killed the Exabyte due to superior performance and capacity.

Seismic acquisition has always been a big user of tape , and 21 track , one inch tape was used in large quantities before GCR recording of 9 track , half inch tape really took over. —Preceding unsigned comment added by 80.169.130.254 (talk) 10:49, 25 August 2010 (UTC)[reply]

It is rather odd and sad that Potter Instruments, with its many, many contributions to magnetic tape (and disk)recording technology is all but forgotten. The company folded in the seventies. They held hundreds of patents including 6250 GCR which was done for Burroughs. They were another of the "plug compatable" tape, disk and printer manufacturers during the sixties, early seventies. IBM used (under license) many of those patents. Woodym1 (talk) 00:29, 6 October 2008 (UTC)[reply]

If you can provide more details on the early history of GCR, please add it to the Group-coded recording article. See also: Talk:Group_coded_recording#GCR_and_Potter_Instruments

--Matthiaspaul (talk) 21:15, 3 April 2018 (UTC)[reply]

What about the physical mechanisms involved in recording and reading? —Preceding unsigned comment added by 84.120.155.124 (talk) 19:29, 6 February 2010 (UTC)[reply]

Same as for hard disk drive. See also Tape drive. --Kubanczyk (talk) 09:00, 22 February 2010 (UTC)[reply]

The section confuses the notions of random access (it takes the same amount of time to read any block on the device), sequential access (it takes a lot longer to read some blocks than others), device i/o (reads and writes to the physical device with no structure imposed by the operating system), and a filesystem (reads and writes to abstract device; units of storage can be accessed by name).Glrx (talk) 16:01, 27 April 2010 (UTC)[reply]

I don't know much on the subject, but I think that the "Viability" section should be scrapped and re-written by someone who does. I looked up this article while reading about the fact that the tape market has barely declined in the last 30-odd years. It seems that the people who claim that tape storage is on its way out aren't the people who has a reason to use it. From what I have read, there is no current viable alternative for long-term storage. Disks need power; tapes do not. Disks usually fail within 10 years or so; tapes do not.

Also, I have a feeling that the $37,000 figure is a bit misleading. Whether it is true or not, it doesn't necessarily reflect typical pricing. According to the .pdf file linked at the end, tapes sell for less than $1/GB (according to Brett Roscoe, Dell Senior manager) and that doesn't count compression. Note: the article is dated at 2008 and is sponsored by Dell. —Preceding unsigned comment added by 79.97.108.130 (talk) 10:56, 11 June 2010 (UTC)[reply]

The viability section appears biased, but the underlying issue seems significant. There are characteristics that should come out in the comparisons: cost per TB, volume of a cartridge, access time, transfer rate, areal density, operation costs, reliability, media lifetime, and power consumption. How such large stores are used is also significant. The section must also respect WP:NOR. Glrx (talk) 00:58, 12 August 2010 (UTC)[reply]

With a tape cartridge capacities of 1 TB, data transfer rates of 160 MB per second and a cost of $150 per cartridge that’s $50/TB (assuming 3:1 compression) tape remains a viable data storage device. Cartridge capacities are projected to reach 50 terabytes by 2019 with data rates over 1 gigabyte per second. Media life for both the midrange and enterprise tape is 15 and 30 yrs. MTBF (Mean Time Between Failure) ranges from 250,000 – 400,000 hours at 100 percent duty cycle. While the $37,000 figure may be MSRP the street price would be closer to $15K-22K.141.202.248.52 (talk) 18:27, 2 November 2010 (UTC)Keith141.202.248.52 (talk) 18:27, 2 November 2010 (UTC)[reply]

As noted two paragraphs up, the question of tape usage (and for which secondary data use cases) is a non trivial topic that can easily encompass pages and pages of analysis (particularly if you appropriately and separately consider backup, disaster recovery, active archive and cold archive, which you must do if you want to be accurate about the costing results). A basic model (assuming active archive, single copy) is do-able using the elements noted above, but it will require supportable assumptions about labor and disk replacement cost, which are bound to generate debate because public (citable) agreement on these topics is, to my knowledge, non-existent. I would also correct the point on compression -- compression is NOT a valid differentiating point for tape; as you can also compress data on disk. Plus - if you want to open the discussion of compression, you must then talk about deduplication, which is of high value on disk for backup but has never been made practical for use on tape. This is one of the key drivers for the decline of tape for backup, but has been less of an issue for archive, given that there are fewer duplicate copies of data. Secondly, while the author above notes (correctly) that volume of a cartridge is key - I think to put this in perspective for the reader, you have to separate open tape from higher capacity proprietary drives (both for cost and density reasons). I also suggest that rather than leaving this at the cartridge level, a comparison should use a standard 'space' element to make the comparison - as an example, comparing the amount of data that can be reasonably stored by an LTO6 library in a standard 19" rack vs. today's disk technology (and taking this comparison forward to LTO7 vs. forecasted improvements in disk) would be very useful. — Preceding unsigned comment added by JanaeSL (talk • contribs) 00:26, 7 July 2015 (UTC)[reply]

It's interesting (from a trivial point of view, anyway) that a 1969 episode of the spy parody show "Get Smart" already uses compact cassettes (called "data cartridges" in the show) to control a computerized grand piano, whereas (as I understand it) the Kansas City standard for data compact cassettes was introduced in 1975 (and the Yamaha digital Disklavier in 1987)... -- megA (talk) 12:05, 23 May 2011 (UTC)[reply]

Tape can be Sequential or Random. There is really no mention of current supported tape that is currently in the field. Difference of LTO generations. DLT vs LTO vs SDLT vs DDS vs MLR/SLR vs AIT? How about Ultrium and how that pertains to the advancement of LTO. What do all those aforementioned acronyms stand for? What is the future of tape? — Preceding unsigned comment added by 207.8.230.2 (talk) 12:30, 12 September 2013 (UTC)[reply]

Can the T10000C tape drive perform both a read and a write operation at the same time on the same cassette? Sofia Koutsouveli (talk) 11:36, 22 March 2014 (UTC)[reply]

In these edits, [1] and [2], an unsourced addition was made to the article:

Helical scan recording writes short dense tracks in diagonal manner. This recording method is used by virtually all videotape systems since quad (transverse) was being phased out, and several data tape formats.

I reverted this the first time, because A) it is not sourced (although, the entire paragraph is unsourced), B) it introduces the word "quad" and truncates "transverse scan" in a way that could confuse readers, C) it is vague about "phased out by whom" and D) it is misplaced because it breaks up the intended conjunction of both videotape systems and data tape formats.

I've laid out my reasoning here. Pinging @2600:100E:B141:A97C:A116:A718:D04A:34A7 and 2600:100E:B108:A2C4:72B6:126F:1096:BF93:. If they would care to justify their version, without the insulting language, please, I'll be happy to discuss it here. — jmcgnh^{(talk) (contribs)} 22:52, 2 August 2017 (UTC)[reply]

Isn't the difference between transverse and helical just the angle of the tracks? It's not as if the tape is stopped while each head of a quad machine swipes across it. Jeh (talk) 23:06, 2 August 2017 (UTC)[reply]

@Jeh: I suppose you could view it that way, but the helical scan mechanism involves wrapping the tape around most of the circumference of the spinning cylindrical disk which makes it a fairly complex evolutionary step beyond the transverse scan. Except for it being unsourced, I don't have objections to the way the section was written. — jmcgnh^{(talk) (contribs)} 01:57, 3 August 2017 (UTC)[reply]

In 1969, a Tri-Data's Cartifile stored data on minicomputers. See http://www.bitsavers.org/pdf/tridata/Tri-Data_4096_Brochure_Feb69.pdf Glrx (talk) 19:53, 16 February 2018 (UTC)[reply]

The article mentions the Exabyte format just once. From the mid-80s it became ubiquitous on mainframes and PCs, lasting until the early 2000s. Tapes were 8mm video and were $5 apiece. It was replaced by DAT tapes on PCs and by AIT tapes for those with money. The biggest drawback of tapes is the sequential access, and low read speeds. Hard drives can also be placed in cartridges and used in storage silos. They are much faster, cheaper, and have random access. Disks can be read and re-written during a single pass to counter magnetic aging. Tapes must be fully read before a re-write.00:52, 14 May 2018 (UTC) — Preceding unsigned comment added by 14.203.207.166 (talk)

I am moving the following uncited material here until it can be properly supported with inline citations of reliable, secondary sources, per WP:V, WP:NOR, WP:CS, WP:NOR, WP:IRS, WP:PSTS, et al. This diff shows where it was in the article. Nightscream (talk) 18:24, 3 September 2022 (UTC)[reply]

@Nightscream: Interesting that you did not cite WP:UNSOURCED which clearly states that when removing material you should " state your concern that it may not be possible to find a published reliable source, and the material therefore may not be verifiable." It looks to me that most if not all of the material can be verified and some may be verified in linked articles. So unless you can express such a concern then I propose we reinstall the material and let the editors overtime find references. Tom94022 (talk) 05:34, 4 September 2022 (UTC)[reply]

Some of this material is going to be tough to verify and may be original research. It looks like someone with a very deep expertise of the topic added a bunch of stuff that may be tough to source except to technical manuals and out of print obscure books. Some of it also needs to be rewritten to be less detailed and more coherent for a general audience. The references, for example, to obscure facts about mainframes (diagnostics or boot code for larger systems such as the Burroughs B1700) are probably from someone's mind and while it's a shame to lose that knowledge, it may not be verifiable easily. The people who wrote this should definitely write a proper book or article in an original outlet about their experiences with computers and tape drives, as someone definitely is interested in that, but it may not belong in an encyclopedia about it. Andre🚐 22:23, 5 September 2022 (UTC)[reply]

@Andrevan: Some of the material may be tough to verify but the example you chose (diagnostics or boot code for larger systems such as the Burroughs B1700) is easy to verify:

IMO most of the facts in this article are not at all obscure and the problem with the bulk deletion is that it is not obvious where to put things back, as for example, in the case of this now sourced Burroughs item. Again WP:UNSOURCED requires a basis for removal and personally I don't think one's lack of subject matter knowledge is a sufficient concern. Tom94022 (talk) 06:06, 6 September 2022 (UTC)[reply]

Nice find on the field engineering manual, I did not think that would be so easy to find. I think you have a valid concern. This removal is too extensive and wholesale. Andre🚐 06:17, 6 September 2022 (UTC)[reply]

Tom, it's interesting that you cite only the one portion of BURDEN that supports your comment here, but not the rest of it, like the part that says, "The burden to demonstrate verifiability lies with the editor who adds or restores material" and "Any material lacking an inline citation to a reliable source that directly supports the material may be removed and should not be restored without an inline citation to a reliable source." I'm not an expert in this area, but I notice that some of the material had been fact-tagged for five and a half years. I then tagged the rest of the uncited info and moved it after well over a month. If no one bothered to cite it in all that time, then it's reasonable to conclude that it's not verifiable, or that no one is willing to do so. You now say that you found sources for some of it. Great! Why not add it? You care enough about that content to complain on a talk page, but not to roll up your sleeves and do the work sourcing it? Why? Nightscream (talk) 13:09, 6 September 2022 (UTC)[reply]

@Nightscream: I didn't cite the remainder of BURDEN because it clearly does not apply to your action of removing material - the fact that some editor in the past did not meet the burden is not relevant to our discussion. I can ask you the same sort of question, why didn't you comply with the clear requirement to state your concerns?

As an expert in this subject matter, I and I suspect other informed editors in the same situation, do not routinely respond to Source tags because for the most part they appear to cover bluesky material or in my informed opinion have a high likelyhood to have an RS. I have no concern about a stale tag - the fact that that tag has been there for hundreds or thousands of about 227 edits and tens of thousands of page views suggests that all subject matter experts did not see the need for a source - to them it was either obvious or readily sourceable. I spend a fair bit of time on Wikipedia and see no need to spend more time responding to such tags.

And I'll ask you the same question: you care enough about content marked as unsourced, even though you have no particular knowledge of the content, to remove it, but are not willing to roll up your sleeves and do the work sourcing it? Why? In particular, why are you depriving the readers of material that is likely valid when you have no idea whether it is or is not? Tom94022 (talk) 19:09, 6 September 2022 (UTC)[reply]

I have. On radio, for example, after doing an uncited info/tp move, I added two citations to an uncited passage. On cassette tape, after my tp move, I replaced an uncited paragraph with a larger version w/ 4 cites of three sources. On the phonograph article, I expanded a section with material supported by 19 cites of 12 sources.

Are you suggesting that it's my responsibility to source all of it? And that the people who add it bear none of this responsibility? Nightscream (talk) 21:18, 6 September 2022 (UTC)[reply]

@Nightscream: I think you might ascribe good faith to the editors who added the material now in question; that is, at the time the wrote it they didn't think sourcing was required. But that has nothing to do with your practice today of massive removal of material with a "talk page move."

Yes you are responsible! You better have a good reason to remove any material and be prepared to explain on a talk page. Before you remove any material for sourcing, you are responsible for at least making a good faith effort to source any content tagged {{cn}} and either source it or remove it and describe your effort on the talk page. That's what any editor has to do who wants to remove any one {{cn}} tag. Just because the tag is stale does not give an editor the right to remove content. Your "talk page move" is just disruptive editing by removing a large perentage of an article without any attempt at analysis and although it may have improved some sourcing in one article it has so far not produced any substantive sourcing in the articles I follow and I for one am uninteresting in correcting this sort of vandalism with other than a revert (which is how most vandalism is dealt with). Tom94022 (talk) 07:41, 8 September 2022 (UTC)[reply]

Question: Do you believe that before moving the material to the tp, I have to make a good faith effort to source all of it? And in every article in which I find this level of uncited info?

Vandalism is defined on Wikipedia as behavior deliberately intended to obstruct or defeat the project's purpose, which is to create a free encyclopedia. Attempting to bring articles in line with policies by fact-tagging articles, adding citations to some of the uncited material, and moving uncited material to talk pages a month after the tagging, is not that, by definition. Nightscream (talk) 14:08, 8 September 2022 (UTC)[reply]

Yes! In fact, every editor should make such an effort to remove a single {{cn}} tag regardless of its date. As a consequence of their effort they can either reference the material or remove it with a statement as to reason for removal which should in either the change line or the talk page or both. So the question becomes is a stale date sufficient reason? I think not for even a single tag much less your "talk page moves" of bulk content because you don't like the number and/or dates of the tags in an article. Tom94022 (talk) 16:44, 8 September 2022 (UTC)[reply]

Again you choose a definition to suit your purposes. The appropriate definition of vandalism is "Removing all or significant parts of a page's content without any legitimate reason" (emphasis added). While your intentions may have been good, in practice "talk page moves" obstruct the project's purpose by removing large amounts of likely valid content, discouraging editing and creating an environment approaching edit-warring - your reason is illitimate! Your continuing this failed practice makes it vandalism as defined. Tom94022 (talk) 16:44, 8 September 2022 (UTC)[reply]

Please do not break up my messages.

I did not "choose" a definition, it's the one that is given at WP:VANDAL and the only one I knew about! That's not choosing, it's citing! I even asked for the source the definition you gave? Is that the behavior of someone cherry picking to suit purposes???

What you cited (and thank you for that) is one type of vandalism, which is not the same thing as how the word is defined. That type indicates that it is done without a legitimate reason. My talk page moves, as I've stated above and elsewhere, and not done without a legitimate reason. The fact that you disagree with them as an approach does not mean it lacks a legitimate rationale on the part of the person performing them, which is why we're currently having a discussion on the matter, and those doing blanket reverts have been told to stop.

Please make a distinction between disagreeing with someone and speculating (or worse, declaring by fiat) that they have no reason for doing what they do. The first is legit. The second is argumentum ad hominem.

I did not asking you about removing cn tags. I asked you if I have to source all the uncited material in all the articles I find. The idea that I should (if that's what your'e saying) is absurd on its face. Think about what floodgates this would open. People like Chetvorno would be given carte blanche to create entire articles composed of material lacking inline cites, as he admitted here, whose accuracy/origin/verifiability would be unknown at a glance for the casual reader, and you're saying that people like me would be responsble for cleaning up their mess. That's ridiculous. Nightscream (talk) 19:01, 8 September 2022 (UTC)[reply]

The description of the IBM 727 and 729 tape drives is wrong, in a significant way. It claims that these use low inertia capstans, but they don't. It certainly is true that most tape drives do, with a servo motor driving the capstan to obtain tape motion and fast start/stop. The 727 and 729 use a more primitive mechanism with two capstans, rotating continuously in opposite directions, and a pair of solenoid-operated light weight idler wheels that push the tape against one of the two capstans depending on which direction of motion is needed. This can be clearly seen in the documentation, in particular in the other 727 manual on Bitsavers. Paul Koning (talk) 01:28, 8 March 2023 (UTC)[reply]

@Paul Koning: you are literally correct but I think the article is trying to distinguish between vacuum column drives and the later streaming drives. So maybe the solution is to just delete the links to those first IBM drives. Tom94022 (talk) 17:11, 9 March 2023 (UTC)[reply]

That distinction is an important one in the device history, which suggests to me it would be good to find what is the actual earliest start/stop capstan tape drive. It may be contemporary or even earlier than the 727, I don't know. What about the Univac tape drives?

BTW, when talking about streaming tapes the DECtape (and LINCtape) should be mentioned; those may well be the earliest example even though they aren't 1/2 inch 7/9 track devices. Paul Koning (talk) 18:45, 1 September 2023 (UTC)[reply]