Talk:Computer data storage/Archive 1

Question
what is Smartdisk. The link is red and as far as i know its a company. If you search google it doesn't seem to say anything different. Of course I didn't look at all 1,390,000 results but the first 30 are all the same. My question is: is at a type of storage or a company? If it is only a company should it be listed here? Bawolff 02:30, 8 Feb 2005 (UTC)
 * What you say looks right. I'll remove it. -R. S. Shaw 20:05, 1 May 2005 (UTC)


 * okay Bawolff 04:16, 4 May 2005 (UTC)

Incorrect redirect? (hey that rhymes!)
Sorry if something like this was mentioned already. I was redirected here from memory chips, but this article seems to be mainly more permanent memory, "storage," like hard disks, floppies, CDs, not RAM, Flash, etc. Is there a better page to redirect to? -- Twilight Realm 23:59, 26 September 2005 (UTC)


 * Computer memory probably. -- 205.175.225.5 01:44, 27 September 2005 (UTC)


 * Computer memory is just a redirect to Memory (computers). This is unfortunate since Computer memory is linked to by about 100 articles, while the actual article Memory (computers) is ref'd by about 15.
 * This storage article used to cover both primary and secondary storage (eg RAM and disk), but an editor recently split it into two articles and put up "Clean up" boxes on them. Things seem to be in a bit of a shambles since then, and the incorrect redirect may be one result of the split.  -R. S. Shaw 05:11, 27 September 2005 (UTC)


 * Maybe an admin could swap Computer memory and Memory (computers), which would simplify cleanup, considering the links. -- 205.175.225.5 21:27, 27 September 2005 (UTC)

Memory vs Storage
The first intro paragraph does not state clearly the difference between memory and storage in the sense that memory is volatile. I think the distiction is important even if both are explained in the same article under computer storage.--Kim Nevelsteen 22:34, 21 August 2005 (UTC)


 * Memory was not always volatile! Magnetic core memory used to be the most common main memory (and was also often used for the CPU registers) and is non-volatile. There is also work on MRAM and similar integrated circuit memory technologies that are non-volatile and operate at speeds of current DRAMs (for both read and write). MRAM may replace DRAM in a few years, making non-volatile main memory the standard again. -- 205.175.225.5 22:22, 21 September 2005 (UTC)


 * I think the difference is that memory is used in dynamic ways by a running progam, whereas storage is used for holding the overall input and output. There are exceptions of course (the obvious ones being virtual memory and RAM disk) but to me that seems the best way of explaining the basic difference. --LesleyW 23:43, 21 November 2005 (UTC)

Deeper informations
I think that in all types of computer memories, if it does not exist, we need informations of how it does function, what is its construction? --User:Vanished user 8ij3r8jwefi 16:52, 2 February 2006 (UTC)

As an organizing concept for computer data storage I have found the "Storage Pyramid" useful. The Y axis is access time and the X axis is capacity. I think a highly granular approach, starting with the CPU's registers, on-chip I&D caches, L2 cache, main memory (RAM), and then off to disk, flash, tape & optical makes a lot of sense. Then it is not about processing, but about data. I believe I have some cycles to help with this too. Let me know. www.storagemojo.com

Cleanup and merging
I have started cleaning up the article. Any comments regarding the direction I'm taking, and any help would be welcome. I also propose that the article should re-merged with memory (computers), since storage and memory are separated only in a very casual sense and it is very hard to discuss the related concepts separately, especially since there used to be no distinction, and a large part of the industry is busily working to remove any present distinction. I think the problem with the August 2005 version was bad organization, not the fact that storage and memory were discussed together. Aapo Laitinen 00:13, 20 November 2005 (UTC)


 * Without having looked through all the related pages, I think it needs to start with a very simple summary (possibly as a separate page) that goes through the "what do we mean by storage" question, and then go into the details of (or links to) memory types and various media types. I actually came here via the "data storage device" page, looking for consolidated information about video tape formats - such as a table with names, dimensions, data formats typically used - and there doesn't seem to be such a thing. There are lots of detailed articles about XYZ product or format, but not much (as far as I could see) that brings it all together. --LesleyW 23:38, 21 November 2005 (UTC)


 * Seems like recording medium, recording formats, videotape or video tape recorder might be better candidates for such content. Video tape formats are somewhat (but not entirely) tangential to computer storage, which is the intersection of recording media and computing. Basically, computer storage is pretty much the "what do we mean by storage when computers are mentioned in the same sentence" page. I too am interested in finding the proper "what do we mean by storage in general" and "here are all the information storage formats we known about" pages, so that they can get linked from this article. In this vein, which do you think is the more fundamental concept, medium or format? Aapo Laitinen 20:06, 22 November 2005 (UTC)


 * Tricky question! A description of CD media should cover (or link to) the usages for audio, data and VCD, but on the other hand, MP3 files can be (are!) stored on a variety of media so an article that's about MP3 and mentions various media makes more sense. I think the question has to be approached on a case-by-case basis. As for an overview of the whole general topic... I think it probably needs to be listed both ways: "here are all the physical formats (media) we know about" and "here are all the data formats we know about". Thanks for the article suggestions, BTW, I'll check them out. --LesleyW 21:37, 22 November 2005 (UTC)


 * I have no objections to a merge, but... (I was the one who split them). I would like to know how you address the differences between memory and I/O storage when certain algorithms have been developed to handle memory and not I/O. Memory is in essence the same as I/O devices if you consider that flash memory is storage, BUT there is still a fundimental difference between a computer's memory and the allocation and management thereof compared with slow I/O calls within the computer architecture. I would split the article with computer storage referring to HD, flash memory, tertiary storage (robotic arms), AND computer memory referring to the computer's memory, the cache memory for the processor, the 1 or 2 Gbs of RAM memory your computer has, the amount of working memory a device has to load programs and data. Until you address these differences then I don't think a merge is a good idea. --None-of-the-Above 09:22, 24 November 2005 (UTC)


 * Well, a merge may not be what is neccessary here, but there needs to be some house cleaning. There's a good amount of overlap, and I'm not sure if the computer memory article has anything in it which isn't talked about elsewhere --cpritchett42 16:44, 24 November 2005 (UTC)


 * I see your point, but I don't think the difference is fundamental. It sometimes makes sense to use internal sorting algorithms to sort external memory, for example. It wasn't until lately (relatively speaking) that memory allocation and management became a function of the operating system kernel instead of the library runtime i.e. the abstraction became mandatory. However, though not fundamental, I think the difference is important. But there is quite a lot of overlap, and I think the difference between, for example, internal and external algorithms is best explained contrasting their characteristics.
 * Looking at what links to computer memory and memory (computers), I would suggest this: Main storage gets moved to main memory, and memory (computers) redirects to computer memory which becomes a disambiguation page with links to main memory and computer storage. Essentially, computer storage would be the overview page about the general concept and terminology. An another article could begin like: XYZ is a form of computer storage and is often used as the main memory, or working memory, of a computer, and yet another article would mention: ABC had 64 kilobytes of main memory.
 * Alternatively, computer memory could be about the characteristics of computer memory with the specifics explained in primary storage, main storage and computer storage, or alternative, computer memory could be about the technology of primary storage, or alternatively, primary storage could be moved to computer memory (which is essentially what you suggested) or computer memory could redirect to primary storage. In any case, I would like to keep the duplication of information at minimum, since I have seen contributions to one article rarely migrate to other articles or do that very slowly. Aapo Laitinen 20:11, 24 November 2005 (UTC)


 * Guess I arrived here late. I've been looking for something to work on since I finished up CPU, and I think computer memory in general needs some work, so here goes...
 * I tend to agree that the differences in the algorithmic treatment of primary and I/O storage are largely late concerns of microcomputers, not really fundamental to a discussion of computer storage (definitely noteworthy enough to discuss, though). I'm all for a good bit of merging since most of these offshoot articles seem to exist more or less because it's difficult to agree on exactly what to call everything.  So far we have (at least) these related articles:
 * Computer storage
 * Memory (computers)
 * Main storage
 * Primary storage
 * Secondary storage
 * Tertiary storage
 * Data storage device
 * Physical memory (slightly less related)


 * ... and probably a few more I missed. This article definitely has the most solid content on the subject; I'm of the opinion that most of the others can simply be redirected here (fortunatly few of them have substantial content that isn't already covered somewhere else).  I support the re-merge with memory (computers) and offer my assistance in doing so.  I think the main article should stay entitled "Computer storage" and be fairly inclusive.  Wikipedia really should have a single article that provides a good and thorough overview of computer storage methods and concerns.  Anyone still around to help out? -- uberpenguin 06:18, 20 February 2006 (UTC)

Data storage device has broader meaning than computer storage
My digital camera is not a computer, its data storage device should not be considered 'computer storage'. The name 'computer storage' is simply the wrong term for something like a flash card. Kevin_b_er 00:59, 14 July 2006 (UTC)


 * Yes, your digital camera is in all likelihood a computer, but that's a moot point. When used in the context of a computer, flash memory implements computer storage.  The scope of this article is (or, should be) more about storage functionality than the actual devices, which may have limited usage outside the microcomputer model.  For specifics on the actual devices we have many other articles, just as the solid state flash memory has its own article. -- uberpenguin

The name "computer storage" to cover all storage media is irrespective of the fact that analog data storage still is not obsolete and even digital storage is used in applications other than computing. The merge should be reversed. Oicumayberight 21:32, 8 September 2006 (UTC)


 * Nobody ever claimed that this article should cover analog storage media. The title "computer storage" nearly implies digital, and the few exceptions (analog computers) are generally not notable enough for the purposes of this article.  Again, read what I said above.  This article should not be about device specifics, but about storage roles within the modern (von Neumann) computer model.  Even in its rough state, note that the article mostly covers storage models, roles and characteristics.  There are (and should be) very little about functional device specifics in this article since there are individual articles for those.  In other words, in this article we talk about (for example) the significance of primary storage in the von Neumann model and mention a few technologies commonly used to implement primary storage.  We do NOT go into a description of how DRAM functions. -- mattb


 * Let me add something that should address what you're getting at. I do think that there is probably a better article to redirect data storage device to.  I think it was originally directed here because its content was largely duplicate of this article.  However, you're absolutely correct that "data" can be in many forms, including analog, and if you can find a more appropriate article to direct data storage device to, I'd be happy to support it.  I think electronic media might be a possibility, but the article contained there is pretty weak, methinks. -- mattb


 * I agree. It doesn't belong under computer storage. I agree that electronic media needs work. I've been working on it myself, but it could use some help from an engineer type. I'm not sure that electronic media should be the final page. Maybe data storage device is better kept separate with a little expansion to analog storage. Oicumayberight 07:13, 9 September 2006 (UTC)

I moved this page to Electronic Information Storage. If someone wants to split the page, having storage signal processing on one page and storage media on another, thats fine. If someone want's to contrast computers storage from other uses on a different page that's fine too. I just don't want to write articles about the alternate uses of common electronic information storage methods with duplicate information and have to worry about version control. My suggestion to the computer writers is to not try to force every technology into a von Neumann computer model, because not everything fits. Some of these technologies have a much longer history and possibly an alternate future than computing. Oicumayberight 12:11, 4 October 2006 (UTC)


 * I have to say I disagree heartily with the new page name. There is nothing that mandates that a computer be electronic, and the concepts of computer storage are abstract in nature.  Even though the examples of implementations of computer storage discussed in this article are mostly electronic, the concepts aren't totally implementation-bound, and I see no reason for this article to limit its scope.  As I pointed out above, this article's main scope is the role of storage in computers, and a little about implementation.  This article is not intended to be a catch-all for all forms of data storage.  What's more, the new title doesn't even suit the de facto content of the article, since several computer storage implementations are mentioned that are not electronic at all: punch cards, delay-line memory, holographic and molecular storage, optical disks.  I understand what you're trying to accomplish, but this isn't the correct place to create an article about all forms of information storage media. -- mattb

new page name and the merge
Taking into consideration that computers don't have to be electronic and that "computer storage" sounds like a store room for PCs, I recommend 3 things:

1) This page is renamed "Computer memory" and only discusses processes for reading and writing digital data RAM, ROM, PROM, device formatting, parallel, serial, etc. Only the storage devices that are unique to the process of computing should be discussed in detail on this page. Storage devices that can store analog data or data for non-computing devices (e.g. magnetic tape recorder) should be discussed on the appropriate page. Oicumayberight 00:39, 9 October 2006 (UTC)


 * "Computer memory" vs "computer storage" is fairly pedantic. The reason we stuck with storage is that "memory" has the slight connotation of RAM, and we wanted to be implementation-agnostic.  If you have another idea why "memory" would be more appropriate, do share.  I fear you're still missing the point, however.  This article should address how storage/memory works within the computer model.  Not implementations.  Not devices.  Not storage outside the computer (von Neumann) model.  If you want to do that, improve or create another page.  Also, RS-232 is a communications standard, not a type of storage. -- mattb


 * I wasn't suggesting that RS-232 is a type of storage. I said RAM, ROM, PROM, device formatting, parallel, serial, were all process-realated. RS-232 is a standard for a process. Computing is a process. Information is not memory until is used in a storage/retrieval process. This was my point in suggesting the pages be kept separate. Storage devices are not processes. Processes are not storage devices. Oicumayberight 23:43, 9 October 2006 (UTC)


 * Okay, then exactly why should information on an I/O standard be included on a page about computer storage? You're making less and less sense now. -- mattb

2) The "Data storage device" page stays separate and expands to include analog as well as digital. Since most electrical information is mechanical, magnetic or optical before it is converted to electrical, use those categories for everything that doesn't require electrical power to maintain the information. Batteries, capacitors, tubes and transistors are examples of the few exclusively electrical storage devices. You can talk about punch cards on that page. You could technically say that a printed book is an optical data storage device on that page if you want to be accurate. The page should make the distinction between the "accessing device" (e.g. player, reader, writer) and "holding device" (e.g. disk, microchip, reel, etc.) Oicumayberight 00:39, 9 October 2006 (UTC)


 * That's fine by me. I hate the title of that page, but that's a trivial concern.  I'm not exactly sure what you're getting at in your second sentence as it's rather a contradiction of terms to say "electrical information can be mechanical, magnetic, or optical"...  That's confusing and largely incorrect.  Electrical signals are electrical, period.  I think what you mean to say is that data (which is a far more general term) can be stored in a plethora of formats on a plethora of media with a plethora of techniques.  I'd be careful about batteries and transistors, though.  Batteries are certainly energy storage devices, but I think you'd be very hard pressed to argue that they are viable data storage media.  Transistors are even fuzzier; a transistor controls current flow.  While several transistors can be coupled with one another or other devices to create a data storage device, a single transistor doesn't really store "data" in a useful form. Anyway, aside from my nitpicks, it seems that you have the right idea for the content of the data storage device article. -- mattb


 * I'm working on it. You'll understand more of what I mean when I'm through. Oicumayberight 23:43, 9 October 2006 (UTC)

3) Stop trying to fit everything that is used by computers to fit exclusively within the context of the Von Neumann computer model. It seems to limit emerging technologies that are not computer related from getting exposure on the wikipedia. Devices like cell phones and digital cameras may integrate with computers, but are not considered computers, nor should they be. This will keep non-computer scientists like myself from having to rename pages that are monopolizing terminology. I wouldn't have renamed this page if there weren't so many redirects to it that had little or nothing to do with computing. Oicumayberight 00:39, 9 October 2006 (UTC)


 * Again, consider that this article is about the significance, organization, and hierarchy of storage within the (von Neumann) computer. It is NOT about general data storage technologies.  What's more, I fail to understand what you're trying to convey.  Taking your own examples, modern cell phones and digital cameras most certainly ARE computers, and usually von Neumann computers at that.  Consider that the relevance of digital storage media grew with and as a result of the relevance of the computer.  The concepts of digital storage and von Neumann machines are very much intertwined.  Anyway, if a lot of general information storage articles redirect here, that should be changed, but the content and scope of this article is more or less correct as it stands.  Could you give some examples of pages that redirect here which you believe should not?  -- mattb


 * Just because something contains a computer, doesn't make it a computer. Haven't you heard of ubicomp? Do you call a guided missile cruiser a computer? When your PDA is in your hand, do you call yourself a computer? Do you call your car a motor?


 * BTW, Storage medium redirected to here. There were some other pages that redirected to data storage device, which would have redirected here if the pages were merged. I can't remember the other pages, but there was more than a few. Oicumayberight 23:43, 9 October 2006 (UTC)


 * What's your point? Let's not get into a pedantic argument about "what qualifies as a computer" or detract from the discussion at hand with silly patronizing rhetoric.  You take issue with fitting digital storage into the von Neumann model, but I find that silly since you very rarely if ever see digital storage used outside the realm of computers.  Cellular phones have digital memory that is accessed by a computer inside them.  Digital cameras (usually) have a computer that access their storage space.  If a guided missile has digital memory in it, you can bet that its used by a computer.  As I said earlier, the historical relevance of digital memory is closely tied with the increased relevance of computers.
 * I understand and agree with the point that articles like storage medium probably shouldn't redirect here. That's fine.  However, I don't entirely agree with the assertion that digital memory is highly relevant outside the model of what we know as computers.  If you really feel that digital memory is significant outside the realm of computers, you really need to find several notable examples of digital memory usage without a computer attached.  Note that computers are in no way limited to a desktop PC.  I don't know if that was the point you were attempting to make, but claiming that a cell phone using digital memory is an example of a device that isn't a computer using digital memory is simply wrong. -- mattb

Tertiary Storage?
quote from article: In a home computer, storage will often take the form of:

Could someone perhaps expand upon the 'tertiary storage' concept (which isn't discussed in the article), or take it out if it's vandalism (can't tell which it is)? Thanks. AustinZ 01:43, 18 October 2005 (UTC)
 * A hard disk, which stores the user's files and programs. It keeps data even if the computer is turned off, and has a large capacity.
 * different sorts of drives
 * tertiary storage, using robotic arms


 * Tertiary Storage as I learned it as have seen it in print is basically the slowest I/O that exists. This is because the data is stored on some media which is completely offline until it is called upon. Then by some mechanism (human or robotic or whatnot) the media is physically taken from a 'shelf' connected to the computer and accessable. The classic example is mainframe computers that use giant reels to store data with a robotic arm that places each reel on the player as needed.--None-of-the-Above 09:17, 30 November 2005 (UTC)

What about a jukebox? Wouldn't that be a better example for the general public? 163.1.143.187 11:03, 8 May 2007 (UTC)

purpose of memory cards
What exactly is the point of adding memory cards? I mean, the hardrive stores memory right? I'm having a hard time understanding the article. I need an answer to this before I start reading the article again. THROUGH FIRE, JUSTICE IS SERVED! 22:16, 21 April 2007 (UTC)
 * Hard Drives are slow.... :-) Memory cards could be faster --Jlc46 17:14, 21 June 2007 (UTC)--

Trends
One thing that is on the flops article is a nice trends table that tracks how the cost/flop has improved over time. I would LOVE to see something similar here for hard drive memory, flash memory, and RAM memory. Any idea where such information could come, and if we could include it here? --Jlc46 17:14, 21 June 2007 (UTC)

merge/purge
Since this article is focused on the Main storage/Primary storage, Secondary storage, Tertiary storage concepts and those pages are mostly stubs, I think they should be merged into this page. Also, the major section in this article that lists technologies is cluttered and disproportionate. Much of it should be moved to pages that can be much more specific. Data storage device and the other technology specific pages would be good places to put this info. -- Austin Murphy 16:44, 17 January 2007 (UTC)
 * Agreed on the merge, tagged those appropriately. --Kubanczyk 17:15, 8 September 2007 (UTC)

ECC
There is no mention of the concept of 'ECC' or 'non-ECC' memory. When I hunt, I can end up at Forward error correction but that's not a useful article explaining memory.

What is ECC memory? What does it look like?

-- Sy / (talk) 11:28, 28 September 2007 (UTC)
 * It looks pretty much like normal memory though usually with a couple of extra chips. While rare errors in memory do happen, in desktops which are price driven we mostly just leave it to chance that such errors won't hit anything too important but in higher end systems it is desirable to have systems in place to detect and where possible correct such errors. Plugwash 14:40, 28 September 2007 (UTC)

New name suggestion
I still think computer storage sounds like a store room for computer. I suggest renaming this page "computer data storage". I directed that page name to this page, but I think the redirect should be reversed. Oicumayberight 09:42, 8 December 2006 (UTC)


 * I agree that Computer Storage is a clumsy name. It is too jargony and not a proper use of English.
 * Enterprise storage is even more jargony.
 * Computer Memory is accurate, but too many tech people are stuck on the idea that memory = RAM and storage = disk.
 * Data Storage is a good, encyclopedic topic that technically includes more than the focus of this article. I'm not sure if wikipedia is ready for a separate Data Storage page yet to cover those additional topics.  For now, I think it should redirect here.
 * Computer data storage is accurate and not as jargony. I think this is the best title.
 * -- Austin Murphy 16:44, 17 January 2007 (UTC)
 * Anyway, if a move is needed I would suggest a proper one, not a cut and paste. See Moving a page for the reason. I temporarily marked Computer data storage for speedy deletion, this is a preparation for the move. I will re-apply reverted changes here. --Kubanczyk 06:05, 9 October 2007 (UTC)

Unneeded description
"Content-addressable storage can be implemented using software (computer program) or hardware (computer device), with hardware being faster but more expensive option."

Are "(computer program)" and "(computer device)" needed? "software" and "hardware" are linked for a reason.

I am a Wikipedia noob so feel free to correct me.

--Pokeman 02:33, 13 October 2007 (UTC)


 * Yes, those extra phrases only detract from the article (IMO). Feel free to remove them. -R. S. Shaw 19:59, 13 October 2007 (UTC)

proposed merge of Mass storage
Mass storage is vague term. How much storage do you need to have before it's mass storage? Is a 40 MB HDD from 1988 still a mass storage device? I suggest dumping the existing mass storage page and replacing it with a redirect to this page. There isn't much there to salvage... Also mass storage is not strictly limited to secondary storage. The typical example of mass storage is a hard disk, but tapes and DVDs are also a form of mass storage that don't usually fit into the the definition of secondary storage. -- Austin Murphy 18:37, 11 October 2007 (UTC)


 * Mass storage = Secondary + tertiary + off-line storage. So, 40 MB HDD is a mass storage, even today. But it's only my intuition. And I don't feel like looking for sources right now. About deleting - I see four or five useful sentences to merge. --Kubanczyk 18:55, 11 October 2007 (UTC)

Please see IEEE Mass Storage Systems Technical Committee. While the definition of Mass Storage may have changed over the years it is clearly one of many subsets of Computer Data Storage and therefore should remain as a separate article. Perhaps the article needs rewriting but not eliminating Tom94022 17:45, 23 October 2007 (UTC)


 * Nobody is proposing a deletion, it is about merging to this section, see e.g. the history of Primary storage redirect. Wikipedia is not a list of notable definitions, see WP:NOT. Wikipedia is a list of long (!) and interesting articles. If you are going to extend Mass storage to make a good article, please proceed. In such case there will be no reason to merge it. --Kubanczyk 21:43, 23 October 2007 (UTC)


 * Since the Mass storage article is neither long, nor interesting, nor relevant or even correct it is a candidate for elimination; however, there should be a stub left for some future author to write such an article about Mass storage as the term has been used for 20+ years in the IEEE society of that name. Tom94022 22:33, 23 October 2007 (UTC)


 * Do you think leaving a stub is really neccessary for future expansion? I would disagree. For example: click Primary storage, scroll to the top section, click link "Redirected from Primary Storage", so you will be on redirect page, able to normally edit it as a separate article. This is a common practice for redirects. There is a number of reasons why this is better than leaving a stub. --Kubanczyk 09:35, 24 October 2007 (UTC)


 * I did send an email to the Executive Committee of the IEEE Mass Storage Systems Technical Committee suggesting they take the time to edit the Mass storage article into something illustrative of the art as they have practiced it since 1974. So I would leave a stub, perhaps pointing to the IEEE Mass Storage history webpage rather than redirecting to an article that is off the point. If you want to update the Computer data storage article with any relevant material from the current Mass storage article including a link to the stub, then I suppose I can find the time to construct such a stub.  Tom94022 17:47, 24 October 2007 (UTC)

DeDuplication
It is not clear this section is appropriate to the article which is more focused on the architecture of the data storage and the components thereof. For example, compression was well established as a means of reducing the size of the stored information well before de-duplication appeared yet compression not been cited. To me, de-duplication is just another, albeit more sophisticated compression within a SAN or NAS. I propose deleting the section but would like to hear some discussion. Tom94022 (talk) 15:50, 24 September 2008 (UTC)
 * I agree, this is inappropriate for this article, and I'll remove the section. If there's further discussion which concludes otherwise, it can always be put back in (or more likely links added to Data deduplication and other compression articles). -R. S. Shaw (talk) 21:10, 24 September 2008 (UTC)

so funny —Preceding unsigned comment added by 69.152.53.126 (talk) 19:02, 25 September 2008 (UTC)

Main memory vs. Secondary Storage (2004 comment)
The following section is best understood from a historical POV where usage of the terms evolved. Economically, speed is an asset when the CPU is operating on the data, whereas size is an asset when reading or saving the data streaming to or from the CPU.

Main memory vs. Secondary Storage

In traditional parlance, main memory contains data that are actively being used (for example, the programs currently being run and the data they are operating on). "Main memory" is typically high-speed, relatively small, is often (but not always) volatile.

Secondary storage, also known as peripheral storage, is where the computer stores information that is not necessarily in current use. It is typically slower and higher-capacity than main memory. Peripheral storage is almost always non-volatile.

I propose this substituted text.Ancheta Wis 05:36, 11 May 2004 (UTC)


 * I have no idea why this section was added at all much less why it was added at the top of this discussion page. So I moved it into its own section.  I happen to think the 2004 proposed definition is not particularly good, since, for example one would ordinarily think of the data in a database stored in Secondary Storage is none the less "actively being used."  I suggest we ignore it and in the due course of time it will be archived Tom94022 (talk) 00:13, 14 August 2009 (UTC)

Where's MRAM?
Spintronics anyone?! —Preceding unsigned comment added by 87.113.16.28 (talk) 14:06, 7 December 2009 (UTC)

8000 000=1MB?
In the first section, this is said, this should be verified though I do not have time at this moment...


 * Yes, 8,000,000 bits = 1 megabyte. Your article edit suggests you think a mebibyte is 10^6; it's not, it's 2^20.  I've changed the article back to show the correct relationship. (Of course a computer would be more likely to come with 1 mebibyte than 1 megabyte, but that's a different issue.  And decimal multiples have been used; for instance some IBM machines came with 65K characters of memory, and they indeed meant 65,000 characters, not 65,536 (2^16) characters.) -R. S. Shaw 01:27, 24 November 2006 (UTC)


 * Actually, 1048576 bytes is one megabyte. -iopq 11:50, 24 February 2007 (UTC)
 * By one definition. By the SI definition (and the one used for most hard disks), a megabyte is 10^6 bytes.  See binary prefix.-- mattb


 * I believe you are confusing bit and byte. A bit is a "1" or a "0", a byte is a group of eight bits. 1,000,000 bytes = 8,000,000 bits. 1 megabyte is technically 1,048,576 bytes, but is commonly rounded to 1,000,000.
 * (1 kilobyte = 1,024 bytes, 1 megabyte = 1,024 kilobytes, 1 gigabyte = 1,024 megabytes...)
 * --Pokeman 02:33, 13 October 2007 (UTC)


 * No one was confusing bit with byte. You seem to have ignored mattb's comment; see binary prefix. -R. S. Shaw 19:56, 13 October 2007 (UTC)

There is a discrepancy with the ways a megabyte is measured. Companies which produce secondary storage tend to use the 1 = 1000 standard metric argument, probably because it is better for marketing. Meaning the MB count is higher using this method, making the storage device look like it holds more data. However, software companies such as Microsoft, as well as primary storage manufacturers use what is technically correct, the 1 = 1024. Remember that this is computer science, not SI mathematics. Numbers work differently than they work in the real world.

Wow, I just looked at the dates and realized that this argument has long passed, but I am posting this already, only because I already typed it. —Preceding unsigned comment added by 68.32.4.83 (talk) 03:40, 5 June 2010 (UTC)

Storage vs memory
I think something like this would be more correct. Normally memory is used to refer to RAM, or virtual memory is used to refer to all addressable storage space on architectures and operating systems that support virtual memory and something like swapping. Normal hard drive storage space would not be called memory, as it is only accessible through system calls and sequences of I/O instructions, not through memory operations. Although certainly operating system mechanisms like virtual memory and RAM-disks can blur the boundaries. NTK 18:05, 5 Dec 2004 (UTC)
 * virtual memory is also volitile memory and hence should fall under the same category.--Kim Nevelsteen 22:36, 21 August 2005 (UTC)
 * The Multics operating system really blurred the separation. In it everything had a "memory address", regardless of where it physically resided (disk, tape, punched card, printer, etc.). There were no system calls for I/O, except to map and unmap files and devices to their addresses. Once mapped the file or device was accessed just like any other memory as far as the programmer was concerned. -- 205.175.225.5 00:15, 23 September 2005 (UTC)

I'm english and I use the word memory far more often than storage. Is this just me? or should the distinction between US and UK words be kept the way it is?
 * Okay, U.K.ers, is "memory" not used significantly in your part of the world? "Storage" is used in the U.S., but it is uncommon to use it for RAM and such.  -R. S. Shaw 20:05, 1 May 2005 (UTC)
 * I'm from England and "memory" is far more common than "storage" in my circles. -Lisa
 * Two votes to nought is good enough for me. I'll remove the UK/US distinction at the top of the article. -R. S. Shaw 22:38, 28 May 2005 (UTC)
 * In the US most computer illiterate people refer to hard disk space as memory, as a geek it is extremely irritating. They are two totaly different things that should not be used interchangabley.  Storage implies that the data is there long term RAM is not storage.  IMHO another title that encompass's both needs to be created.
 * I agree; should not the link to this article on the page memory_(disambiguation) take the reader to the article on RAM? Geekosaurus 2 July 2006


 * Hard disks do implement a form of memory... You're just debating semantics. "Long term" is relative to some time frame... Relative to a femtosecond, DRAM certainly does implement long-term storage.  The primary reason this article is called "storage" is because the word lacks some of the aformentioned connotation that "memory" has with solid state RAM.  Really this article could equally well be titled "Computer memory"; it's a fairly arbitrary decision.  -- uberpenguin

In my personal experience, as per all the posters above, the term 'storage' isn't used to refer to volatile memory. Could someone provide a reference for this statement

To whoever made the above unsigned comment, a quick search for "real storage" will give you a definition. Since virtual storage is being discussed, and also to distinguish between hardware and memory content, I'd suggest adding a section that clarifies the distinction.

For a historical example and one that's easy to find references to, you can check IBM's z/VM command reference manual, which I believe is available on line. The CP (control program) query command, which is the layer below the virtual machine level, will give information about memory, not disk, if you query storage. For a privileged user (class A system administrator) "QUERY STORAGE" will give information on real storage, or real memory. For a general (class G) user, it will give information on virtual storage, which of course looks like real memory to the user's virtual machine. A class A user would need to "QUERY VIRTUAL STORAGE" to get information on memory being used by the virtual machine that's issuing the command or would end up querying real memory by using the class G form of the command.

I use this as an example because it is a documented use of a command that was used the same way in 1972 as it is today. In addition to calling memory "real storage," it was also commonly referred to as core storage. When I first learned about core storage, it literally consisted of a bunch of wires passing through iron cores, each one used to store a bit. But programmers talked about what's "in core" for many years after that. So there's no question that storage referred to volatile memory going back to the early days of computers, and nobody used it to refer to disk drives.

In the example above, the distinction between what the computer reports as storage based on context is simple. Virtual storage refers to the content of memory as the user's application sees it, whether or not the OS has it there at any given instant. If we step out of the mainframe world, real memory is only a concern for typical end users with respect to whether their total hardware capacity can support what they run. If virtual storage is not discussed in the main article, there's nothing giving readers a clue what's meant when referring to what's in memory, because some of it might be in real storage and some might be swapped out to auxiliary storage.

Hagrinas (talk) 01:31, 1 October 2010 (UTC)

Magnetic cards
No mention of this primitive storage device, as used by HP-65, etc. I remember spending longer convincing my boss that it worked than I did programming it! SombreGreenbul (talk) 16:17, 24 November 2011 (UTC)
 * Cards, in general, added to section on magnetic storage. IMO no need to go to the detail of the HP65.  Tom94022 (talk) 16:42, 24 November 2011 (UTC)

CRC for EDAC
Does any system built in the past 20 years use CRC in storage? I would be interested in -- and surprised by -- any examples. (Ideally, examples in the form of a reliable source WP:RS that we could use as citations in this article).

Recently, I changed the sentence that previously said "The cyclic redundancy check (CRC) method is typically used in storage for error detection and correction."

That edit was reverted with the comment "CRC was used and probably still used in storage".

Really?

My understanding is that, while a typical CRC is excellent for detecting errors and widely used in communication, CRC is useless for correcting errors in storage.

The current hard drive and random-access memory and ECC memory articles never mention "CRC". Instead, the hard drive article mentions "Reed–Solomon error correction" and "low-density parity-check codes", which (unlike CRC) correct bit errors. Likewise, the ECC memory article instead mentions "SECDED Hamming code" and "triple modular redundancy", which (unlike CRC) correct bit errors.

I propose that we begin the sentence "... method is typically used in storage for error detection and correction." with some method that actually is commonly used in storage for error detection and correction. --DavidCary (talk) 05:08, 8 August 2014 (UTC)
 * I think it is fair to say Storage uses both CRC and ECC, CRC with retry for some errors and ECC for other errors. CRC is only the ED part of EDAC.  With regard to HDD storage, for many years the data and servo sector headers were typically only protected by a CRC; not sure if that is still true for servo sector headers.  It maybe similar in optical storage.  It is true in floppy storage although that is pretty much obsolete.  It maybe true in some older tape storage.  Futhermore, many of the serial storage interfaces also were only protected by a CRC and I think that remains true to this date.  So at least in some aspects, I think Storage today uses CRC in much the same way as Communications, to detect some errors and possibly correct them by a retry.  Storage unlike Communications does provide additional error correction and detection for most errors thru ECC.  So the article as it stands is correct although perhaps incomplete, depending upon how Storage ECC is handled in the article (don't have time to read it right now).  Tom94022 (talk) 03:31, 9 August 2014 (UTC)

SSD SSHD OSD
Article needs to be brought up-to-date. Secondary data storage is no longer just traditional HDD-based, that is spinning disks. There are now SSD, SSHD and OSD options. Obviously, solid-state-based date storage devices do not have seek-time latency. Should also mention object-based storage (OSD) as a next-generation data storage architecture, particularly for unstructured data, such as images and video, linked with the rise of companies such as Facebook and YouTube. 66.155.23.67 (talk) — Preceding undated comment added 12:16, 30 July 2015 (UTC)

Cloud Storage?
I didn't see mention of cloud data storage i this article. Did I miss something? 70.247.162.60 (talk) 15:24, 8 November 2015 (UTC)
 * No it's not there becasue IMO its just way of connecting to and/or owning Secondary or Tertiary Storage. I suppose one could so note in the appropriate sections, but I wouldn't add it as another layer.  Tom94022 (talk) 19:33, 8 November 2015 (UTC)
 * Good point by 70.247.162.60 that it isn't OK for this article never to mention it or link it even once. Good point also by Tom (not separate from secondary and tertiary), but (as he said, "could so note") it really out to be mentioned and linked in those sections. I may go add a line. Quercus solaris (talk) 02:41, 12 November 2015 (UTC)
 * Agreed, it clearly isn't a separate layer, but should be mentioned in appropriate sections. &mdash; Dsimic (talk &#124; contribs) 21:53, 15 November 2015 (UTC)

Nanowire research
(Copied from tom94022 talk page)

This is authoritative referenced info. Where should this go? This shouldn't be deleted.

Researchers at University of Southampton discovered a way to store 360 terabytes of data on a quarter sized disc for billions of years preserving humans' learned information. The scientists use a sophisticated laser to encode the information into minuscule formations, known as nanostructures, inside fused quartz. The structures alter the way light travels through the glass, allowing the data to be read by special optical devices. --Wyn.junior (talk) 18:01, 18 February 2016 (UTC)


 * There are all sorts of research projects that don't necessarily belong in Wikipedia and this maybe one but it certainlhy doesn't deserve a new section by itself. Perhaps it could appear as a new item in Other Storage Media or Substrates following "Molecular memory," but please note that each of the potential future memory technologies have reached Wikipedia's level of notable and have a corresponding article.  IMO it is not yet notable and shouldn't appear at all but I would like to hear from other editors.  I suppose it could appear in the linked section without its own article.  What makes u think it is notable? The technical journals are full of not notable material.  Tom94022 (talk) 18:39, 18 February 2016 (UTC)


 * My sense of notability is only by the source. CNN Money. This isn't a technical journal.--Wyn.junior (talk) 02:25, 19 February 2016 (UTC)


 * CNN Money is even less of a source for notability. Tom94022 (talk) 01:23, 20 February 2016 (UTC)

Using processor instead of CPU
Since now GPUs are also becoming widely used, should we use processor as a general term (to refer to both CPU and GPU) instead of CPU alone? Previously CPU alone was synonymous for processor, but it is no longer true. --Bhattasamuel (talk) 12:51, 9 April 2016 (UTC)


 * Processor refers to a general processing unit unless specified otherwise ie GPU. A GPU is only good at specific tasks and thus the need to specify what kind of processor one is intending when not referring to the general one. Bytesock (talk) 16:05, 9 April 2016 (UTC)

Please split Computer_data_storage into separate article
I did not expect this gem hidden inside trivial Computer_data_storage.

This page is huge and most of the statement are historical/trivial. Please don't merge articles about professional terminology with historic articles about Punched cards. It should be stated at History of computer data storage, not at Computer data storage (or Memory hierarchy). Ushkin N (talk) 21:16, 27 July 2016 (UTC)

"Primary storage topics" vs "Secondary, tertiary and off-line storage topics"
Data deduplication, Data proliferation can be used at any level.

Yes there in-memory deduplication.

1 Kernel same-page merging

There so many unsourced inaccuracies in this article, that I'm simply using Memory hierarchy. Ushkin N (talk) 03:47, 30 July 2016 (UTC)

External links modified
Hello fellow Wikipedians,

I have just modified 1 one external link on Computer data storage. Please take a moment to review my edit. If you have any questions, or need the bot to ignore the links, or the page altogether, please visit this simple FaQ for additional information. I made the following changes:
 * Added archive https://web.archive.org/web/20090822172353/http://www.dvddemystified.com:80/dvdfaq.html to http://www.dvddemystified.com/dvdfaq.html

When you have finished reviewing my changes, please set the checked parameter below to true or failed to let others know (documentation at ).

Cheers.— InternetArchiveBot  (Report bug) 08:19, 29 November 2016 (UTC)

External links modified
Hello fellow Wikipedians,

I have just modified 2 external links on Computer data storage. Please take a moment to review my edit. If you have any questions, or need the bot to ignore the links, or the page altogether, please visit this simple FaQ for additional information. I made the following changes:
 * Added archive https://web.archive.org/web/20081221131054/http://springlightcfl.com/consumer/energy_savings_calculator.aspx to http://www.springlightcfl.com/consumer/energy_savings_calculator.aspx
 * Added tag to http://www.simpletech.com/content/eco-friendly-redrive
 * Added archive https://web.archive.org/web/20110823052120/http://news.inventhelp.com/Articles/Computer/Inventions/apple-macbook-air-12512.aspx to http://news.inventhelp.com/Articles/Computer/Inventions/apple-macbook-air-12512.aspx

When you have finished reviewing my changes, you may follow the instructions on the template below to fix any issues with the URLs.

Cheers.— InternetArchiveBot  (Report bug) 21:05, 11 August 2017 (UTC)

Business aspects
The article now is mostly physical: speed of access, energy use, hierarchy, etc. There needs to be something about the business end: concentration and competition of manufacturers, buying storage hardware vs. renting online storage, license and privacy restrictions, etc. Let me know if there is another article which gives an overview of the business end. If not, I could start adding a section here, and if it gets big it could then become a separate article.

For example Western Digital and Seagate each has about 40% of the hard drive market; Toshiba has most of the rest, and the total is declining. The market for solid state drives is rising. The biggest flash memory manufacturers are Samsung, Toshiba and Sandisk (Western Digital).

Licenses from different companies have different pitfalls: There are articles on very specific storage businesses, such as Comparison of file hosting services and Comparison of online backup services, so maybe there's a general article on the computer storage business already. Numbersinstitute (talk) 19:27, 1 May 2018 (UTC)
 * Sandisk/Western Digital "When you visit our websites or cloud environment, we may collect ... usage information, ... file attributes (including attributes for photos, videos, music, and documents), and location information from ... devices on which you ... access our products. https://www.sandisk.com/about/legal/privacy
 * Seagate can use information stored on their products: "When you ... use our Offerings, we collect ... information that you provide ... through the use of various Offerings ... Any information you submit ... through our Offerings will be used for a variety of business purposes, including for example, to ... provide ... user experiences;" "Offering" is defined to include Seagate products. https://www.seagate.com/legal-privacy/privacy-policy/
 * Samsung only collects, " information about you, your device, or your use of the services", not the information placed on the device. https://www.samsung.com/us/account/privacy-policy/
 * Unfortunately Computer Data Storage is more than just SSDs and HDDs (i.e. secondary storage) so the added section is somewhat redundant, very incomplete and inappropriate for this article. The market shares of SSDs and HDDs are already coverred in those articles and it is not clear that the licensing issues belong here at all.  So I am deleting it for futher discussion here.  I suppose brief summaries of the business aspects of the four hierarchical levels of computer data storage would be appropriate but they should wikilink to articles dealing with each level. Tom94022 (talk) 17:02, 30 May 2018 (UTC)
 * Too true the section is incomplete. That's not generally a reason for deletion, but for expansion. Most topics start incomplete. The HDD article has a bit on market share, and the SSD and SDRAM articles have a bit on which companies reached various milestones first (which I would not suggest citing here). There are lots of overlapping articles on the technologies and hierarchical levels. This seems a central article in this constellation, and it should have some coverage of business and link to wherever there is more information. Computer memory is a big business, like much of technology. Briefly distinguishing the sizes and license rules of major producers (which produce multiple technologies and levels) is part of describing the business. I think we can agree on brief, with links elsewhere. Numbersinstitute (talk) 23:36, 31 May 2018 (UTC)
 * If by brief with links elsewhere you are proposing such for each of the four levels (possible sub-sections on the section) then we are in agreement. I don't see any need for the licenses issues in HDD/SDD in this article at all.  Tom94022 (talk) 18:35, 1 June 2018 (UTC)
 * Well I don't know if I have enough information for sub-sections, but I'll do something, and you and other editors can reorganize if needed. Sorry you don't see the need to cover license issues, but they're certainly important, especially privacy aspects, with GDPR just going into effect. This article has any number of topics which some can call unneeded, but others see as useful parts of a general encyclopedia article, such as the conferences, history of each type of media, which is presumably also in other more specific articles, and some of the definitions. Each is useful to some people, and it's convenient to have them together in this article. The article is well enough organized that people can read what they want and disregard the rest. Numbersinstitute (talk) 22:27, 1 June 2018 (UTC)
 * I don't see much in common in the business aspects of the four hierarchical levels and as near as I can tell coverage will vary as they don't have much in common. For example, primary memory is dominated by a few companies making DRAM in very expensive facilities but I don't see anything about the business aspects in DRAM or Computer memory so before you add something here you might consider adding something to one or both of these articles.  Likewise tertiary storage is dominated by tape libraries and off line storage has articles on tape drives, DVDs and Flash drives all with different business models and issues not well covered in the articles.  Actually the more I think about it, there is no need for this section at all and to the extent there exists information in related articles such as hard disk drives and SSDs the information can be summarized and linked from the appropriate subsection of the Hierarchy of storage section, in the case of SSDs and HDDS that would be secondary storage subsection.  In the future as we add business aspects are added to the main subject articles we can add links and summaries herein but adding the section as a separate section, particulary when addressing only one segement of the hierarchy makes little sense to me. Tom94022 (talk) 23:31, 1 June 2018 (UTC)

Dedicated article on "overprovisioning"
Searching for "overprovisioning" currently leads to a disambiguous page, which has a link to this article. However, this article currently doesn't mention overprovisioning. Should "overprovisioning" be added to this article, or should it be a separate article? I think there is enough information about "overprovisioning" for creating a dedicated article about it. The article can be named "Overprovisioning (computer data storage)". The following is a list of some items that we can include:


 * 1) definition of overprovisioning in context of computer data storage
 * 2) benefits of overprovisioning for NAND flash
 * 3) history of overprovisioning being used, not prevalent in HDDs, but more prevalent in SSDs
 * 4) marketing and overprovisioning found in consumer SSDs, etc.
 * 5) general confusion with storage capacities
 * 6) lawsuits regarding actual vs available storage capacities

Please feel free to contribute! Somerandomuser (talk) 16:04, 21 July 2021 (UTC)
 * My apology for editing your talk; it makes my response simpler and really didn't change anything. Tom94022 (talk) 00:17, 22 July 2021 (UTC)
 * FWIW I don't think overprovisioning is generic to computer data storage so perhaps removing the link from the disambiguation page is more appropriate. As you note in 2, 3 and 4 it is a term of the art in SSDs and NAND and could be in those articles, it is implied in link to Write amplification from Flash Memory.  Items 4 and 5 are really not related to overprovisioning.  So IMO probably a section in Flash Memory covering 1-4, with links to/from SSD, disambiguation page and Write Amplification would be appropriate but nothing seems appropriate to this article.  Tom94022 (talk) 00:17, 22 July 2021 (UTC)
 * See: DEFINITION overprovisioning (SSD overprovisioning) -
 * Overprovisioning, in a storage context, is the inclusion of extra storage capacity in a solid state drive (SSD). "
 * In general, to provision is “to provide” or make something available. Historically, IT departments have overprovisioned IT infrastructure because systems had to be big enough to handle peak demand.
 * Tom94022 (talk) 01:32, 22 July 2021 (UTC)


 * I agree that this article is too general a place to have mention of overprovisioning, as is the integrated circuit article, the food article, etc. I've been bold and removed the Computer data storage link from the overprovisioning disambiguation page. --R. S. Shaw (talk) 20:27, 25 July 2021 (UTC)

Confusing and inconsistent nomenclature
The article seems to assume that primary storage is volatile, and delay lines certainly were. However, the 1950s and 1960s were dominated by nonvolatile primary storage, e.g., core memory, magnetic drum, magnetic rod, magnetic thin film.

Devices such magnetic tape drives served the role of secondary storage on early machines without disk storage, but on machines with disks they served a role closer to tertiary storage, although usually without an automatic tape library.

The 1960s and 1970s had several technologies that, while slower than disks, were faster than the 5–60 seconds in the description of tertiary storage. These included

Shmuel (Seymour J.) Metz Username:Chatul (talk) 14:48, 2 January 2023 (UTC)
 * NCR Card Random Access Memory (CRAM)
 * IBM 2321 data cell
 * RCA 3488 Random Access Card Equipment and 70/568 Mass Storage Unit
 * IBM 3850 Mass Storage System (MSS)
 * Your statement of history is quite accurate but I really don't see much of any such historical inaccuracies with the article, other than the misleading template which I removed. For example, the Primary storage section mentions delay lines, Williams tubes, rotating magnetic drums and cores. If you see any other inaccuracies, please either fix them or at least point to them.  Tom94022 (talk) 18:03, 2 January 2023 (UTC)
 * I'm thinking of text like In the 1950s through early 1970s, faster persistent technologies were referred to as both memory and storage, depending on the vendor and sometimes the device.
 * BTW, IBM still uses the word storage for primary storage, at least on z, and NCR used memory for secondary storage.. --Shmuel (Seymour J.) Metz Username:Chatul (talk) 20:22, 2 January 2023 (UTC)
 * So why don't you fix it by changing it to something like
 * or some such lede summary sentence. I suggest the IBM and NCR examples are outliers not worthy of inclusion but if u want to work them into the article go ahead.  Tom94022 (talk) 20:36, 2 January 2023 (UTC)
 * or some such lede summary sentence. I suggest the IBM and NCR examples are outliers not worthy of inclusion but if u want to work them into the article go ahead.  Tom94022 (talk) 20:36, 2 January 2023 (UTC)
 * or some such lede summary sentence. I suggest the IBM and NCR examples are outliers not worthy of inclusion but if u want to work them into the article go ahead.  Tom94022 (talk) 20:36, 2 January 2023 (UTC)

SD cards?
Should the article mention SD cards, or have they been supplanted by USB thumb drives? Shmuel (Seymour J.) Metz Username:Chatul (talk) 13:23, 9 January 2023 (UTC)

Adjacentl links
I changed to to eliminate redundant use of the word memory.

Tom94022 then changed it to, which introduced adjacent links.

Can somebody come up with wording that is clear, concise and complies with wiki guidelines on adjacent links? --Shmuel (Seymour J.) Metz Username:Chatul (talk) 20:52, 16 January 2024 (UTC)