Talk:Cloud computing

Criticism (light) on Cloud Computing

Trying not to be harsh, but... OK, check this: in cloud computing, if you mainly store your data online, even if you lock it and whatnot, it can still be hacked into and... well, then where's your file? someone took it, important or not, especially if it contains personal info. —Preceding unsigned comment added by 4.184.4.221 (talk) 09:13, 28 January 2010 (UTC)

and this is somehow different if you have your data on your own PC (and your PC is turned on regularly with your network cable plugged in). How? Do you think you are better equipped to safeguard your data than the online storage people? Mahjongg (talk) 23:13, 19 February 2010 (UTC)

This is actually a really good question. There is no way of knowing how data can be recovered without testing the recovery process. This is why I've been harsh on some people like Reliacloud who make indefensible claims about how their storage provides geographic dispersion when all their web site talks about is a SAN, which only spans a single site. For good data redundancy you need to store more than one copy of the data, on different sites. You also need to keep checksums of the data, and regularly verify that the copies of the data you have checksum to the same value as the checksum. If there is a mismatch you have a problem. Apache Hadoop, for example, has every data storage node, the datanodes, set up to continually read and checksum the blocks in their spare time, so that problems get found and corrected early. Even then though, go through their bug database and search for "Hadoop DFS data loss" and you find out the harsh truth: data can get lost in interesting ways. See [| HDFS Issues] for details.

Amazon S3 is replicated onto multiple sites, I have heard claims that they are on different fault lines, so it would take more than one earthquake to lose the data. However there is still the risk of data corruption, which is often the most insidious. The single documented event [|| 2008-07-20 outage] tells us much about its implementation.

A conclusion has to be: you can't trust any single site, local or remote. You need to keep a copy somewhere else, possibly offline. If you want it to be secure "in the cloud" or locally, encrypt it, don't forget the pass phrase. But also bear in mind the lifespan of DVDs and tapes, the long term availability of DVD and tape readers, and other potential problems. SteveLoughran (talk) 17:35, 21 February 2010 (UTC)

The conclusion is, storing data on your own computer may be just as unsafe as storing it in the cloud. Mahjongg (talk) 08:45, 22 February 2010 (UTC)

Guys, this is not the right forum for general cloud discussions. Unfortunately I don't have any suggestions for a better one, but this isn't it. -- samj _in^out 20:58, 22 February 2010 (UTC)

Agreed. Mahjongg (talk) 23:44, 22 February 2010 (UTC)

I disagree, this is the perfect forum to discuss cloud computing, exactly for the reason that you don't have any suggestions for a better location. Of course, location is used loosely because it can only be defined by an IP address and a position on the very "cloud" we are discussing. Lately I've been noticing in management system magazines the advertising lining the pages with companies offering their abundant data collection to everyone, for a pay as you go business model. The company I want to focus on from the magazine is AT&T, they not only offered space available to businesses, they also offered security. The only item that remains in question is, how do you govern promised security on a world wide market? Are there any laws supporting AT&T's claim of security? The truth is, if they actually devote the energy and time in securing and storing your information, it is less the expense for your busines or yourself. Think about the costs associated with retaining and developing encryption for your information. When I think about it I'm certainly not above cloud computing and outsourcing my storage space, as Nicholas G. Carr discusses, many businesses have already put the energy into making the decision to use these outsourced data centers.

-J.W. Watson —Preceding unsigned comment added by 150.131.70.175 (talk) 16:42, 17 March 2010 (UTC)

I think it might be worth pointing out the WP:NOTFORUM policy here: bear in mind that talk pages exist for the purpose of discussing how to improve articles; they are not mere general discussion pages about the subject of the article, nor are they a helpdesk for obtaining instructions or technical assistance. Letdorf (talk) 13:52, 18 March 2010 (UTC).

The lead, one sentence. That's all I ask.

Well, months later, the lede is still baffling. I just don't have the head to fix it. I hate computers, but would personally love to know what the heck cloud computing is. Here are three sources with definitions. All I am suggesting is a simple, one-sentence addition to the beginning of the lede to help people understand. Then, sure, the mumbo jumbo and talk of paradigm shifts can follow. [1] [2] [3] I've contributed to this problem as far as I can. Please help. Anna Frodesiak (talk) 04:35, 22 February 2010 (UTC)

The simple problem is that every party involved has some reason why "The Cloud" needs to mean what is (Commercially or Technically) "best for them", so an agreement on what "the cloud" means is almost impossible to enforce. This situation will sort itself out once people start to use one incarnation of "The Cloud" above all the others. Meantime this article talks about the principles involved, and that makes it so convoluted, also because everybody tries to give the meaning its own twist. Mahjongg (talk) 08:52, 22 February 2010 (UTC)

Okay. That's sounds reasonable. I would still like to add some sort of broadly agreeable definition and see what happens. I don't want to start an edit war though. Thanks for your input. I will replace the unintelligible first sentence with this, which is a composite of many broad definitions that I have come across:

Cloud computing is a way of computing, via the Internet, that broadly shares computer resources instead of having local servers handle specific applications.

Anna Frodesiak (talk) 09:31, 22 February 2010 (UTC)

Fine, I have just one practical comment, and that is that most users wont recognize that "local servers" in this instance means their own PC. So I would make it Cloud computing is a way of computing, via the Internet, that broadly shares computer resources instead of having your local PC handle specific applications.

Mahjongg (talk) 13:43, 22 February 2010 (UTC)

That's what local servers mean? I thought it meant computers in the vicinity. I know nothing of computers. Ask me about pies. I know how to make pies. Can you make the edit? Be bold. Thanks. Anna Frodesiak (talk) 13:53, 22 February 2010 (UTC)

Well, thats part of the problem, there may be people who argue that the cloud is an alternative to serving applications from a local server (intranet). But I that is really not what most people will think of when they think about running "cloud applications". I will be bold, and make the edit, But I fully expect I will attract "flak". We will see what happens. Mahjongg (talk) 18:55, 22 February 2010 (UTC)

Ok so you've picked up the "Internet" and "multi-tenancy" themes but missed "utility", "device and location independence", "scalability", etc. Sure what was there isn't perfect, but this isn't much better. -- samj _in^out 19:09, 22 February 2010 (UTC)

Ah, there we go! Duck and cover! :-) . Where does it say "multi-Tenancy " in the first sentence (and what does it mean)? "utility", "device and location independence", "scalability" All these terms will mean nothing to the average potential user of a cloud service who wants to know what the heck it is. That these terms do not clutter up this first sentence in the lede means there is a fighting chance they will read on, and not be discouraged from the start. Its a very complex issue, lets start it simple. Mahjongg (talk) 23:52, 22 February 2010 (UTC)

Mahjongg: Well put! Let's not squabble. We're here to serve (ultimately) the visitors to the page. Let's keep the lead sentence as broad as possible, and then go into details later. And SamJohnston: At least the lead sentence approaches a definition. The previous lead sentence made no sense at all. Please tweak the lead, keeping in mind that visitors may be coming to the page for a definition, as well as further information. Happy editing. Peace. Anna Frodesiak (talk) 01:32, 23 February 2010 (UTC)

Look I agree the article needs to be fixed and have spent the intervening 8 hours working on it... the WP:LEDE should be an introduction and "summary of the important aspects of the subject" which means it should be fairly specific while also being readable. For some background I wrote the article & created the diagrams a few years ago (after spending countless hours poring over literally dozens of definitions) and have spent the last 3-4 years helping large enterprises understand what it's all about. Most of my (limited) Wikipedia time recently has been spent on pointless bikeshed painting here on the talk page and cleaning the article's various spamtraps. As you can see today I've been working from the bottom up, leaving the hardest stuff until last. If you're happy to help get the article in shape (ideally GA or possibly even FA status) then you're most welcome, but let's try to be efficient about it. -- samj _in^out 01:52, 23 February 2010 (UTC)

I agree. And, your substantial improvements to the article are well appreciated. My only involvement or concern was with the first sentence of the lead. As long as it's not what it was before, I'm happy. As for helping making it GA or FA, I'm unqualified to help. My meddling work is done here. I'm off to Pie to see if I can get my fingers into that. Thanks for your patience and understanding. Keep up the good work. Anna Frodesiak (talk) 02:40, 23 February 2010 (UTC)

Yes, if you look at how readable its now compared to say a week ago, its a huge improvement. The lede should be readable by the average Wikipedia visitor, technical details can be explained later, and non mainstream interpretations of the term should too. best regards. Mahjongg (talk) 22:59, 23 February 2010 (UTC)

And once again, already, the intro is not only meaningless (eg vague) but flat out wrong. "Cloud refers to the fact that exact pathways and exact servers need not be addressed by the user."? WTF? -- samj _in^out 15:29, 25 February 2010 (UTC)

Tweak it. Change it. As long as the lead sentence doesn't get into paradigm shifts, I'll be thrilled. Anna Frodesiak (talk) 15:41, 25 February 2010 (UTC)

Agreed that sentence does nothing to clarify what cloud means to an average users, it rather "clouds" the issue (no pun intended). It should go. Mahjongg (talk) 17:32, 25 February 2010 (UTC)

• The problem is every time we touch this (trust me it's happened like 1/2 a dozen times now) it goes into a state of fast flux and ends up completely ridiculous (as above) before I reset it to something sensible derived from the words themselves (cloud as in internet-based and computing as in, well, computing). Imagine the Internet is like the electricity grid with no power stations - you can connect on one side and consume power from a solar panel I've got connected on mine, but there's nothing else to it really. Enter cloud and you're adding the power stations. -- samj _in^out 00:01, 26 February 2010 (UTC)

Yes there seem to be forces at work that deliberately want to disorient people about what cloud computing could mean for them, (and what the potential pitfalls are) or they want to give it one proprietary meaning. We should not let this happen, and revert with malice and tenacity! Mahjongg (talk) 20:46, 26 February 2010 (UTC)

Ok so once again we have a WP:LEDE which is more precise but less accurate. Who ever said cloud doesn't need/use software & storage on the local PC? What's a browser? Gears? You see the problem now - it's not so easy to accurately describe in plain english. And you know what, it doesn't need to be precise, so long as it's accurate. -- samj _in^out 00:30, 8 March 2010 (UTC)

You seem to use very peculiar definitions of "precise" and "accurate", which I do not care much for. My only wish is that the lede is readable, and is not manipulated to tender for just one special interest group which wants do define "cloud computing" for their own limited purpose. As it is now, its much better understandable than it was before, and I think it would suffer greatly if we want to make it "less precise", and "more accurate". And no, I don't see any merit in your "objections". The lede isn't claiming that not any software or storage is used on the PC, it would be absurd to read that into it, and who doesn't know what a "browser" is, and why is that even relevant? The word "gears" (for "google gears", I think that is what you mean) doesn't even appear in the article (let alone in the lede), so what is your point? So no, i do not "see the problem". Mahjongg (talk) 02:29, 9 March 2010 (UTC)

Precision and accuracy are best explained using the target analogy. You want something specific, to go into some amount of detail for readability/understandability, and yet cloud is, by definition, nebulous and thus far evades a precise definition. Furthermore, it means different things for different people (and there are conflicting reliable sources to this effect) so we really need a definition that satisfies most/all rather than one that matches with your personal view of what cloud is (with or without an opinion piece as a source). This attempt is not bad - and certainly a lot better than many before it - but I think we can still do better. Skype, for example, uses local software almost exclusively to deliver a service that is, for all intents and purposes, cloud (the key thing is that you don't [need to] care how it is delivered). -- samj _in^out 21:14, 9 March 2010 (UTC)

Thanks for clearing up what you mean by "Precision" and "accuracy". Its precisely because the definition of cloud is still foggy at the moment that it would be unwise to be too "precise" in the definition, and it may well be that trying to be "accurate" is at this point unwise as well, as in fact the "target" (what the primary meaning of "cloud" will become) may move in the future. So IMHO we need to paint as global a picture as possible. The lede also should try to use language that the average reader who has a slight interest in the subject can understand, so do not try to be "accurate" by using "computers, phones and other devices", but use something they understand as "that what you are using now". A "Personal Computer" (I try to avoid the term PC, as many people think it exclude Mac's) seems to be a good enough description. Even better (more generic) would be to use the term "desktop", but its difficult to use that without causing confusion. This is why I am also against using specifics in the lede, such as mentioning Skype. I am not trying to write the lede to reflect "my personal view of what cloud is", but rather to what I feel is the current average idea that people have about the future/current meaning of "cloud". By "people" I mean people that have made an educated guess about what meaning "cloud" will get the future. But the picture is as you said still fuzzy. I think that therefore the lede should reflect that. It should not try to be "precise" and "accurate", simply because its still unsure which definition of "cloud" will prevail in the coming years. Especially I don't want to bury the lede under confusing highly technical specifics, as has been the case in the past. For that you can read the rest of the article, that mentions all the warring specifics. Mahjongg (talk) 01:22, 10 March 2010 (UTC)

When you ask "people" they say "Internet": it's like computing on the internet innit[4] -- samj _in^out 02:46, 10 March 2010 (UTC)

LOL, a "taxi-driver definition" I like that! As I said I don't mean those people, I mean people who have taken the time and trouble to get a little bit informed about the subject, enough to generate their own (well educated) idea of what they -hope- "cloud computing" can offer them in the future. I think that this "hope" is ubiquitous, stress-less, effort-less and non frustrating use of a computer like device and its applications and services, anywhere and anytime, for less money than they pay now for their current computer and software. Mahjongg (talk) 23:31, 11 March 2010 (UTC)

Watch this video (and part 2) of Canonical's Simon Wardley explaining cloud definitions, utility computing, commoditisation, etc. - it's not too long and it's fairly entertaining (not to mention insightful). -- samj _in^out 11:33, 12 March 2010 (UTC)

Break

I think the reference is destabilizing the lead sentence. It has already been tagged as an unreliable source. I suggest that both the ref and the tag be removed. Also, it seems that some are unhappy with the first sentence, but most agree that it is better than before. So, I suggest that those who feel they have something better, why not present it here so that we can see what an alternative definition might look like. Anna Frodesiak (talk) 21:39, 9 March 2010 (UTC)

Current

Cloud computing is a way of computing, via the Internet, that broadly shares computer resources instead of using software or storage on a local PC.

slight modification

Cloud computing is a way of computing, via the Internet, that broadly shares computer resources located across the internet rather than using software or storage on a local PC.

make it clearer where the resources may be. --Salix (talk): 21:54, 9 March 2010 (UTC)

The main issues are that a> software is required on the PC (usually a browser but often enough a dedicated application) and that b> it doesn't matter where the storage is... nearly all cloud systems use local storage either implicitly (browser cache) or explicitly (offline access via HTML5/gears, dedicated clients, etc.), not to mention compute & network resources (esp where a local client is involved, and volunteer computing like Skype). -- samj _in^out 23:51, 9 March 2010 (UTC)

• Reference (IEEE paper): Volunteer Computing and Desktop Cloud: The Cloud@Home Paradigm -- samj _in^out 23:52, 9 March 2010 (UTC)

I think something like this would be more explanatory but still not too specific:

Cloud computing is Internet-based computing, whereby software and information is consumed by computers, phones and other devices on-demand, like a traditional utility.

Perhaps we don't need all the links, but they do provide a good explanation of the concepts for a beginner. Is that more like what you had in mind Anna? -- samj _in^out 00:23, 10 March 2010 (UTC)

Keep it simple please, do not obfuscate the lede! Don't be obtuse for the sake of "precision", the term "cloud" is err... too clouded still to do that, its not helpful to the average reader. Remember this is only the lede, for all the warring details and definitions of "cloud" you can read the rest of the article. I only added this reference because it was the first relevant entry I found when literally entering the lede into google. If its is destabilizing the lede (whatever that means), please remove it, or add a better one, the net is humming with similar "definitions". Mahjongg (talk) 01:36, 10 March 2010 (UTC)

Simple, sure, but it also has to be right. -- samj _in^out 02:42, 10 March 2010 (UTC)

Yes, but what is "right" in this context when the definition of "cloud computing" is still a moving target.Mahjongg (talk) 03:03, 10 March 2010 (UTC)

The phrase "...whereby software and information is consumed..." doesn't quite make sense to me. From what I understand, resources such as processing and storage are shared. Further down in the lead, I see the phrase "...shared pool...". That phrase seems to hit the mark for me. Correct me if I'm wrong, but in laypersons' terms, this whole thing is about using a smidgen of storage and processing power on one's own PC, (basically a bit of software to interface with the cloud), to access and share raw computing power and tons of storage that exists out there in the cloud. Am I way off? Anna Frodesiak (talk) 01:24, 10 March 2010 (UTC)

No, that isn't way off, obviously you need some elementary computing power to get on the net and make use of cloud services. Still "shared pool" does nothing for me other than conjuring up an image of people sharing a jacuzzi Keep the language simple please. Mahjongg (talk) 01:39, 10 March 2010 (UTC)

Ok so how about this then:

Cloud computing is Internet-based computing, whereby shared resources, software and information are used by computers, phones and other devices on-demand, like a traditional utility.

It is also a good deal more explanatory, and largely technically correct. -- samj _in^out 02:42, 10 March 2010 (UTC)

Okay, what about this: Cloud computing is Internet-based computing, whereby shared resources, (like software and information) are used by computers, and other computing devices, in an on-demand fashion, like traditional utilities such as water and electricity.

Mahjongg (talk) 02:53, 10 March 2010 (UTC)

In fact I meant "shared resources" (compute, network, storage) in addition to software (code) and information (data) - you need those three things for computation, which is why those three things are listed (albeit in human-friendly forms). I'd also steer clear of giving examples of utilities... they are listed in the LEDE of the linked article anyway and some people have an aversion to the electricity analogy specifically. I honestly don't think it needs to be much longer than what I've suggested anyway - as you say, we've got the whole article to go into detail. -- samj _in^out 03:19, 10 March 2010 (UTC)

Could probably drop "phones" in retrospect:

Cloud computing is Internet-based computing, whereby shared resources, software and information are used by computers and other devices on-demand, like a traditional utility.

Better? -- samj _in^out 03:20, 10 March 2010 (UTC)

Now I think it is taking a nice shape. How about this tweak to the last bit to openly disambiguate, else people (like me) will probably think of gas and electric:

Cloud computing is Internet-based computing, whereby shared resources, software and information are used by computers and other devices on-demand, like traditional utility computing.

(It's strange seeing the words "traditional" and "computers" in the same sentence. I am getting OLD!!!!)

Anna Frodesiak (talk) 04:47, 10 March 2010 (UTC)

Actually gas and electricity is the comparison we want people to draw, we just don't need to be so explicit. One could argue that cloud is "traditional utility computing", because we never really got it right before. -- samj _in^out 11:21, 10 March 2010 (UTC)

Well, indeed its progressing, but I still have a few doubts, I don't much care for "computers and other devices", it reads as if washing machines and such will use cloud, which may well start to be true, but it isn't what is intended here. I objected against "phones", because it was too specific, and at the same time not all phones (will) have internet connectivity. So I still think that "other computing-devices" is better. I don't like the term "utility computing", in the lede, it is another highly specialized term, so at one end you are explaining the user what the concept of "cloud computing" is, while at the same time introducing a just as nebulous (no pun intended) new term. I didn't even like the fact that "utility" in fact linked to "utility computing". What is wrong with mentioning the fact that computer power will become as easy and accessible as drinking water, I think that for many people that is exactly the main promise of "cloud computing".

So my preferred sentence would become:

Cloud computing is Internet-based computing, whereby shared resources, plus software and information are used by desktop computers and other computing-devices, on-demand, just like traditional public utilities. Mahjongg (talk) 11:38, 10 March 2010 (UTC)

Ok so I think you're over-thinking it. For a start, there is no need to call out resources specifically when they are useless without software and information. There is also no need to be more specific than "computers" and in any case desktops are less and less relevant these days as people move to mobile devices ranging from laptops through phones - notebooks have been outselling desktops since 2005. My choice of the word "devices" was deliberate - smart grid meters are one example but ubiquitous computing in general relies heavily on cloud. Finally, when we have a perfectly good article bridging the gap between computing and utilities, why would we want to throw our readers in the deep end? In any case, if we start talking about "public" utilities we're sure to have the private cloud brigade on our case in a heartbeat. Perhaps the linking could be extended but I really don't think we're going to get much better than this:

Cloud computing is Internet-based computing, whereby shared resources, software and information are used by computers and other devices on-demand, like a traditional utility.

I think you are trying to find extra precision when there is none to be had - specifying 'desktops' for example, and 'public' utilities. I honestly believe this is as precise as we're going to get while maintaining accuracy. -- samj _in^out 13:57, 10 March 2010 (UTC)

Well the boldening of "shared resources" and "are used" wasn't my input, and I have no problem with removing it. I can live with the other changes too, because of the "less is more" principle. My one remaining reservation is why do you insist in linking to utility computing, its not helpful here to link to this defunct and specific (in the wrong way) subject, while we are trying to purify this first sentence from any non generic definitions of what cloud is, It smells like an attempt to sneak in a non generic definition again. I would remove the link, or link to the neutral public utility instead. I'm sure we could tweak the sentence indefinitely, (for example the end "like a traditional utility" is a bit ambiguous), but its not too bad, most people will understand what is meant. After removing two other somewhat distracting links, leaving only the two which may clarify non obvious terms we end up with this (which is very close to what you started with):

Cloud computing is Internet-based computing, whereby shared resources, software and information are used by computers and other devices on-demand, like a traditional utility.

Mahjongg (talk) 23:15, 10 March 2010 (UTC)

Ok by me - I did feel there were a few too many links. -- samj _in^out 23:35, 10 March 2010 (UTC)

Ok by me too. Clear and concise. A cirrus improvement. Anna Frodesiak (talk) 00:10, 11 March 2010 (UTC)

Great, if nobody is objecting I think this will be a good first sentence, just one point, Traditionally the recurrence of the title in the fist sentence should be bolded. Ill add that to the edit now. Mahjongg (talk) 00:13, 11 March 2010 (UTC)

Sure, if we're down to that level of detail we're doing well... -- samj _in^out 08:47, 11 March 2010 (UTC)

"Utility" is a bit of an ambiguous word. What is meant by a "traditional utility"? If we're comparing cloud computing to a public utility, then why not say that? If we're not, exactly what kind of "utility" are we comparing it to? Letdorf (talk) 13:17, 11 March 2010 (UTC).

Yes, we mean public utility, but you can be sure that if we use the term "public" we'll have the private cloud parade on our case ("electricity" is another one guaranteed to upset people who still want to sell kit). I don't see a problem with linking to the utility computing article here as it provides a good bridge between the two concepts (and some have argued that cloud computing is just utility computing with a new name - something I don't fully agree with but a sentiment I can nonetheless understand). I think using the term "utility" and linking off to the utility computing article would do nicely, rather than dropping people in the deep end by linking to public utility. -- samj _in^out 13:41, 11 March 2010 (UTC)

• Cloud computing is Internet-based computing, whereby shared resources, software and information are used by computers and other devices on-demand, like a utility.

How about this? I think utility is a bit ambiguous so "traditional" was serving a purpose. If we don't like "traditional" then we may have to say "public utility" or give the "electricity" example - both of which I think will cause unnecessary contention. In any case I think the utility computing article is vastly better than the (very poor) public utility article for getting across our message. -- samj _in^out 13:48, 11 March 2010 (UTC)

Why not just drop the utility bit, it seems to confuse more than elighten, comparison with utility computing can go lower down. Hence:

''Cloud computing is Internet-based computing, whereby shared resources, software and information are used by computers and other devices on-demand.

--Salix (talk): 15:22, 11 March 2010 (UTC)

If "utility" is a problematic word, maybe services will (err...) serve us better. What about this variation.

''Cloud computing is Internet-based computing, whereby shared resources, software and information are served to computers and other devices on-demand.

As I said, I object against using "utility computing" for the reasons stated above, and don't understand why this option returns each time, if "public utility" is a poor choice, then "utility computing" is infinitely more so. In fact I have no problem ending the sentence with "like a public utility. What is wrong with the term "public utility", or with "public"? Please enlighten me, is there a ethical/moral issue against using it here I am not aware of? Comparison to a "public house" perhaps, or what other sensibilities are there with the term? I simply don't understand the objections against "public utility", except when it comes from "private cloud" proponents, i would say they are oversensitive if that is the case, we hardly imply that cloud computing cannot be delivered by "private entities", only by "monopolies". IMHO comparisons with the services that cloud computing provides with public utilities like the provision of water, gas or electricity only clarifies what cloud computing is. Mahjongg (talk) 23:07, 11 March 2010 (UTC)

It more that I just don't think its needed in the first sentence. It clearly needs a little more explanation as there is subtleties on quite what is meant by utility public or otherwise. This to me sugests a second sentence.

I do like you version with served which seems to capture the relationship better than just used.--Salix (talk): 00:15, 12 March 2010 (UTC)

Utility computing is central to cloud computing (some argue they are one and the same - I would say cloud is a form of utility computing or that it provides utility computing) - indeed Nick Carr's "The Big Switch" arguably got the ball rolling by comparing it to electricity. It should absolutely, without doubt be in the first sentence and I don't think there's much contention about that point, rather what form it appears as. Looking at the quality of the public utility article though, and knowing that there's already a bit of a jump to get from computing to utilities, I think we either need to spoon feed our readers with the utility computing article (which concisely describes the relationship between the two in the LEDE) and/or give the electricity example. If we really want to dumb things down then the latter is probably the better option, but with the advent of cloud, utility computing is hardly "defunct and specific (in the wrong way)", rather alive and well. I really don't know how we could improve on this definition:

• Utility computing is the packaging of computing resources, such as computation and storage, as a metered service similar to a traditional public utility (such as electricity, water, natural gas, or telephone network).

On the subject of using the verb "serve" - the first thing I think of is client-server. Other terms I considered were "transferred" (again Internet/networking/client-server) and "consumed" (apparently too technical) so I think "used" is functional, without being too specific (and wrong). -- samj _in^out 00:46, 12 March 2010 (UTC)

Another option would be to drop the "utility" ref altogether (as suggested by Salix), but run with "electricity" instead (which is a potent analogy that needs no explanation whatsoever):

• Cloud computing is Internet-based computing, whereby shared resources, software and information are used by computers and other devices on-demand, like electricity.

Sure we may rouse a few "oversensitive" cloud-as-a-supermarket types by we're armed to the teeth with reliable sources - even the Macmillan Dictionary agrees with us. Say "public" and it'll be on for young and old. -- samj _in^out 01:08, 12 March 2010 (UTC)

I like the "like electricity" analogy, but the sentence as a whole still doesn't really convey the message too well, it seems to be the "whereby". But perhaps its because I'm not a native English speaking person. I thnik this is slightly more understandable:

• Cloud computing is Internet-based computing, where shared resources, software and information are offered to computers and other devices on-demand, like electricity.

Mahjongg (talk) 12:10, 12 March 2010 (UTC)

Ok so we're nearly there... the used/transferred/consumed/served/offered verb needs to be vague I think, because there are so many different modes of operation (datacenters, peer-to-peer, volunteer computing, etc.) - and it's more than "offering", it's actually making use of the resources. I think "consumed by" is technically correct, but "used by" is ok if we want to dumb it down a bit. offering/serving/transferring is more like client-server/Internet, which focuses on the transfer of data rather than the actual application. -- samj _in^out 12:47, 12 March 2010 (UTC)

I would maybe have said "provided" rather than "used" or "served". IMHO, "electricity" isn't much of an improvement though... is it like electricity in that you can't see it, or it can be stored in batteries, or it can electrocute you? I think we need to be more specific about the electrical power industry, if that's the analogy we want to use. Letdorf (talk) 13:29, 12 March 2010 (UTC).

"Provided". Nice. Seems obvious now, and while there's still a hint of client-server in there, that's indeed the reality for much of cloud today. I think the electricity analogy is fine as is but if it's a problem for you then how about "the electricity grid"? Hmm... readability suffers. One could argue the [generic] electricity analogy also satisfies the private cloud parade (where in this case it's a generator rather than a grid), in which case the extra precision is maybe unnecessary/unhelpful? I still think utility computing is by far the best article to link here, and all the others I've looked at (including electricity grid are really throwing our readers in the deep end. -- samj _in^out 14:25, 12 March 2010 (UTC)

I've made the change suggested by Letdorf above and think what we have right now reads well and is far better than anything we've had previously:

• Cloud computing is Internet-based computing, whereby shared resources, software and information are provided to computers and other devices on-demand, like electricity.

I'm not sure there's much more improvement necessary (or indeed, possible). -- samj _in^out 14:30, 12 March 2010 (UTC)

The last word "electricity" is not too clear. Maybe "household electricity" or "tap water" or something to indicate definitively that it is something sold on a metered, per unit basis. Just my two cents. Anna Frodesiak (talk) 22:57, 12 March 2010 (UTC)

Yes, "tap water": is better, as at least you can't be "electrocuted by it (alone) ", and you can "see it" (just like "information"<grin>). Oh ,and using "utility computing" is certainly "throwing the reader in the deep end", as it obviously needs an explanation of its own. For many its just another word for "cloud computing", and I don't understand what is helpful about using another similar word that is just as, or even more, unknown to the reader, while trying to explain what the first word means. Also "cloud computing's" first sentence should not directly link to an old and defunct variation of itself. Mahjongg (talk) 00:34, 13 March 2010 (UTC)

P.S. I simply added "or tap water" to the end, as I think it makes it extra clear we mean "electricity" to be a service, disambiguation it from other meanings. Mahjongg (talk) 00:39, 13 March 2010 (UTC)

No, "tap water" is original research, and utility computing being "an old and defunct version of itself" is your opinion. The utility computing article is exactly what we are referring to:

• Utility computing is the packaging of computing resources, such as computation and storage, as a metered service similar to a traditional public utility (such as electricity, water, natural gas, or telephone network).

If you don't agree then please tell me why, and what we need to do to fix the other article. Honestly I think what we've got is good enough, linking to the utility computing article would be better, and whatever we decide the discussion will be brought up again in 6-12 months with different participants and quite possibly a completely different result too. -- samj _in^out 04:13, 13 March 2010 (UTC)

To put things in perspective, we have spent well over 5,000 words discussing 20, and this is definitely not the first time there has been a large scale debate about the lead - probably about 1,000 words of discussion for every word that appears in the article - trying to find a consensus where there is quite probably none to be found. -- samj _in^out 04:19, 13 March 2010 (UTC)

Are you serious? "electricity" is fine, but "tap water" is original research? Well that is the most contrived reason I read on WP for a long time. I don't know why "utility computing" is your "hobby horse", but you seem to be on a mission to slip it into the lead sentence, which just wont do. As a word alone it does nothing to explain anything to the average reader, and as a link Utility computing doesn't link to your neat definition sentence, but redirects the reader away to an article that tries to give a specific definition to "cloud computing", the very thing we want to avoid. Removing "tap water" isn't a big deal, but it just makes it a little clearer we are talking about utilities without actually using the word "utility". That we need so much words discussing the first sentence is because the first sentence is so important. For many people this will be the first definition they read of what cloud computing is. So many institutions have an need to define "cloud computing" to fit their purpose that we cannot allow them to define the first sentence. Therefore the first sentence must not only be readable, and factual, but also should be devoid of any specific definition. That includes a specific definition such as "utility computing", which for example claims that "its a metered service", which well may not apply in the future. There is no reason to include it in the first sentence, and there is plenty of opportunity a few sentences later to mention it, just not in the first sentence. Mahjongg (talk) 21:06, 13 March 2010 (UTC)

Actually being a measured/metered service is pretty much universally considered an essential characteristic of cloud computing, and you have yet to demonstrate why the utility computing article does not do an adequate job of explaining the message that we are trying to get across. Rather than repeatedly accusing me of trying "to sneak in a non generic definition again" et al, how about you stick to the facts. Do you have any examples of where this is not the case as saying it "may not apply in the future" sounds like original research and/or opinion too... -- samj _in^out 04:17, 14 March 2010 (UTC)

To be metered or not metered, that is the question. (is it?) Listen, I do not have a crystal ball, but neither do you. I gave just one example by expanding the definition to that of "utility computing" in the first sentence is a bad idea, actually Anna Frodesiak has put it quite well. For a visitor to have to read a second article just to understand the first sentence of the lead will be too much. . That, and that alone, should be more than enough reason not to put the link to utility computing into the lead sentence. Mahjongg (talk) 14:36, 14 March 2010 (UTC)

Links are helpful (except perhaps if there are too many) and are very often included in article leads. Most readers will understand what we're talking about at first glance and those that don't can click through to find out more. Let's not forget that a second ago you were promoting linking to the public utility article. -- samj _in^out 18:21, 14 March 2010 (UTC)

Yes, this certainly is the mother of all threads. I just thought concerned editors would throw out a few tweaked opening sentences, without having to state their rationale only in the edit summary.

As for the last bit, as a non-partisan layperson, linking to "utility computing" in the lead is problematic. For a visitor to have to read a second article just to understand the first sentence of the lead will be too much. I suggest excluding the last bit if a clear example cannot be found. I am almost out of substitutes to suggest. Maybe "...household utility", "...public works utility", "...electricity from the power grid", "...electricity from a power company", hmmmm, okay, I'm fresh out. What else is metered? Actually, the way it is now using "electricity" is not bad, except that readers might not understand that it means "metered electricity". Anna Frodesiak (talk) 00:02, 14 March 2010 (UTC)

If only it were that simple... anyway "electricity from the power grid", or rejigging to get "the electricity grid" in there would presumably be satisfactory (albeit IMO completely unnecessary). -- samj _in^out 04:07, 14 March 2010 (UTC)

That is exactly why I added "or tap water", as its another public utility, just as in some places natural gas is, Mahjongg (talk) 00:58, 14 March 2010 (UTC)

...but it is not used in this context, which makes it original research at worst and a poor choice at best. -- samj _in^out 04:07, 14 March 2010 (UTC)

What "context"? I'm just trying to convey the meaning that "electricity" must be read as a public service, not as "an electric potential", or "power source". Is that so hard to grasp? Adding "or tap water" (another utility service) is just a way to accomplish that the reader understands we are talking about public services here, without using the word "public" (as that may "offend some" editors here, as you claim). Your claim that adding "or tap water" is "original research" seems to be only based on mis-reading ("wishful reading"?) of an aborted entry in my edit summary that read "added or", while it should have been ' ' added "or tap water" ' ' (I wrestled with how to solve the double use of quotes, and inadvertently pressed enter) , In your zeal you seem to have thought it was a "self confession" of me that I "added Original Research". Otherwise your obsession with calling my edits "original research" makes no sense to me, and I think not to most other editors. I think this silliness has taken enough of my time, and this thread is long enough now. Mahjongg (talk) 14:36, 14 March 2010 (UTC)

Actually I just thought that using an edit summary of "added or" for original research was ironically entertaining... it was that or a tag. -- samj _in^out 18:23, 14 March 2010 (UTC)

Yep, "electricity" could mean so many things. It is ambiguous. If the suggested disambiguations are "out of context" and "original research", please explain why. I'm just not following that. Anna Frodesiak (talk) 19:47, 14 March 2010 (UTC)

It's not ambiguous when you link to the utility computing article (which explains very clearly, exactly the message we are trying to get across). I've not found a reliable source comparing cloud computing to tap water, and for every one that exists there would be 100 comparing it to electricity. Hence original research. -- samj _in^out 21:08, 14 March 2010 (UTC)

Fair enough. But it would still be better to compare it to something other than utility computing for the aforementioned reasons. Can you think of something we can compare it to that is self-evident -- something that does not need a reference? I am not seeing any difference between tap water and electricity, as they both are simple examples of something that you can tap into, and that is paid for in a metered fashion. I know this is a lot of energy to spend on a single word, but this is really the crux. This is the single term that will help everyone understand. Your suggestions are very welcome. Best, Anna Frodesiak (talk) 01:18, 15 March 2010 (UTC)

The "single term that will help everyone understand" is "utility" (even the public utility article says "usually just utility"). "Traditional" utility is vague and "public" utility is arguably wrong (both because it could be a private utility and because for many this refers to government ownership rather than public accessibility). These "aforementioned reasons" are still eluding me - I see I just don't like it (literally "I don't like the term 'utility computing'") and the claim that "For a visitor to have to read a second article just to understand the first sentence of the lead will be too much". Check out the top 5 Featured Articles for engineering & technology (35_mm_film, Atomic_line_filter, Autostereogram, Matthew Boulton & Construction of the World Trade Center). ALL of them use links in the lead extensively and the WP:LEDE article specifically encourages this. If the word "utility" is linked, most will understand - those that do not (or simply want further explanation) can click through for more details. The only question that remains then is what article to link to, and when you have two concepts (utility and computing) and a perfectly good article bridging them (utility computing) then why on earth would you not use it? The only justification provided for this is that it "may not apply in the future", which is nonsense - and even if it were true we could always revisit the article were that ever the case. -- samj _in^out 04:53, 15 March 2010 (UTC)

The "single term that will help everyone understand" is "utility" Agreed, and I don't object so much to using the word utility at the end of the first sentence, I object to using utility computing at the end., not because "I don't like it", but because it's not needed to make the sentence more understandable, and because it introduces an alternative meaning to "cloud computing". The examples you give to explain "why it is allowed to do so", only explain that it is allowed to use a lot of links in the first sentence, each a "breakdown" term of the subject involved. There is indeed nothing wrong with that, and we are doing that. Let me try to explain my objection, lets pretend we try to explain what a "chicken" is. This is the normal way:

The chicken (Gallus gallus domesticus) is a  domesticated fowl.

Now what you seem to want to do is claiming there is a need to change the sentence to this

The chicken (Gallus gallus domesticus) is a  domesticated fowl, like the Leghorn

Yes, I know its a slightly absurd example. But what I mean by it is that you use another slightly different and more specific version of "cloud computing" in a sentence designed to define cloud computing in as abstract terms as possible. Doing that destroys that the sentence is meant to be as abstract as possible, and suddenly points to one specific example of it, in the case of "cloud computing" to the specific version of "utility computing", and in the (slightly absurd) case of the chicken into the specific Leghorn. As I said, I don't object to the word "utility" (But I do think the reason for not using "public utility" is silly), I object to a link to the subject utility computing at the end. Mahjongg (talk) 11:53, 15 March 2010 (UTC)

If the difference between "utility computing" and "cloud computing" is subtle and uncertain, then it's probably best not to compare the two when trying to define cloud computing in as few words as possible. As for the phrase "public utility", since this is only an analogy, and a very broad one at that, I can't see how the fact that some clouds may be considered "private" (which can probably be considered as something of a special case anyway) as being a valid objection to using it. Letdorf (talk) 14:42, 15 March 2010 (UTC).

Ok we've just cracked 5,000 words discussing one. Run with "public utility" and I'll ignore its alternative meaning (government run) and hope the private cloud parade will also ignore it so as to avoid yet another time wasting debate about the lead. It's not like I even believe in private cloud so it suits my personal preferences just fine. -- samj _in^out 14:54, 15 March 2010 (UTC)

For what its worth, I too hope that cloud computing won't be "government run", (many utilities are NOT ran by governments you know). Concerning "private cloud", what is there to "believe", you have "private virtual networks" too. "private cloud computing" will probably at best be an obscure diversion from the norm (but I don't "predict" anything), if the "private cloud" evangelists make a fuss about "private" in the lead sentence, (used in a wholly different context) then thats their hyper sensitivity and problem. Mahjongg (talk) 23:17, 15 March 2010 (UTC)

I doubt there'll be much "government run" cloud (most of it will be partially/fully outsourced) - my point was that by talking about "public utilities" some readers will assume we are talking about "government run" utilities. Regarding private cloud, a Merryl Lynch MD at Cloud Connect last week said "I'm just going to call a private cloud a data center.", which is about right. -- samj _in^out 16:29, 20 March 2010 (UTC)

Well maybe that mirrors that you have learned (by example) that most "public utilities" are run by the state, in my case I have learned (by example) that very few (actually I know none) of the "public utilities" are run by the state. I think it will be different from region to region. In fact I don't expect that many people who read "public utility" would automatically assume that this thus -must- imply that its "state run". Its certainly not stated or implied in the rest of the lede, Its a moot point I think. If there are signs that this really is an issue, then you can explicitly express it somewhere else in the lede. Mahjongg (talk) 18:52, 21 March 2010 (UTC)

I can't think of any state run utilities off the top of my head... though I did one work for one that was in a past life (Telecom Australia aka Telstra). Probably more true of developing countries than the west these days. -- samj _in^out 10:21, 23 March 2010 (UTC)

I'm puzzled, if you do not connect "public utilities" with "state run", then why do you suppose the readers will? But no matter..., I'm just curious that is all. Mahjongg (talk) 16:38, 23 March 2010 (UTC)

Well, most of the public utility article carries on about the things being government run for a start - mentions governments about a dozen times over the space of a few short paragraphs! -- samj _in^out 00:00, 26 March 2010 (UTC)

Ah, well then its a good thing we do not link directly to that article. In defense of the article, it does have a "The examples and perspective in this article deal primarily with the United States and do not represent a worldwide view of the subject. Please improve this article and discuss the issue on the talk page." header. But yes, the narrow-mindedness of this bad article explains some things. It does indeed need some good cleaning up. Still I do not expect even Americans to have a pavlov reaction to Public utility such that they automatically assume it must be government run, one can only hope. Mahjongg (talk) 00:12, 26 March 2010 (UTC)

I guess it depends on what you're used to... here in Europe I'll bet there's plenty who have different ideas. Anyway "public" as a prefix to anything (including "cloud") is problematic - I'm glad someone removed it from the article. -- samj _in^out 11:58, 30 March 2010 (UTC)

Criticism section redux

The criticism section is once again a coatrack for random disparaging comments from individuals (which do not necessarily reflect the opinion of the communities they represent). It includes criticism by a CEO whose sales are directly affected by cloud computing (following a pattern of "abuse and embrace" that we've seen before), whose company has now completely embraced cloud computing with a 50 city tour, and which analysts have rejected as "complete nonsense". Following this there's an analyst whose opinion differs from that of his company (and which appear to have changed anyway) and finally a diatribe from a free software activist. Oh and it closes with an irrelevant ad for "Fabasoft Folio Cloud".

This is an obvious NPOV violation, giving undue weight to the negative faction, yet every time I touch it I get abused for repressing others opinions etc. If it must stay (I would argue that it shouldn't - integrate with the articles of those quoted or if the issue is notable then include it in the "Issues" section, which I created in part to avoid these useless soundbytes) then we should balance it with a good handful of quotes from notable individuals gushing about cloud and how it's going to cure AIDS and deliver world peace. Thoughts? -- samj _in^out 12:55, 11 March 2010 (UTC)

In general "Criticism sections" in Wikipedia articles are a bad idea, and should be discouraged, they generally tend to be "troll magnets" at best. In fact wikipedia has a policy that says its better to integrate any negative points into the article itself, instead of creating a "criticism section" . Mahjongg (talk) 23:14, 11 March 2010 (UTC)

While I agree with you 100%, I'm not touching it with a 10 ft pole until we have a clear consensus here. I would however appreciate if one of you could remove the blatant advertisement and ideally warn the user for their unhelpful contribution. I also agree that we should discuss valid points in the issues section (security, privacy, control, etc.) and leave the soundbytes and name dropping out altogether - it adds nothing and for every -ve quote we can find a dozen +ve ones. -- samj _in^out 00:58, 12 March 2010 (UTC)

Agreed, (also the ten feet pole sentiment, I'm fed up with bickering, thank you). Mahjongg (talk) 23:20, 15 March 2010 (UTC)

Ok, so let me put this another way: is there anyone who thinks that the dedicated criticism section needs to remain? If so, why? Why not just integrate the valid/notable points? -- samj _in^out 16:30, 20 March 2010 (UTC)

First, I think its strange that the section is called "criticism of the term", as nobody is criticizing the term "cloud computing", they are criticizing the idea(l) behind it, not the term. Secondly, I don't care much about criticism of Oracle's CEO, as it is obvious where it is coming from, and at the same time it's just as quickly retracted if that fits their "bottom line" better. Quite a different thing are the opinions spouted by Richard Stallman, yes he is something of a Zealot, but he does say what a lot of people think, (and fear) that cloud computing can become something that detriments the freedoms they have with Personal Computers. I do agree that these views and opinions should be integrated into the article. Mahjongg (talk) 22:13, 20 March 2010 (UTC)

Ok so unless anyone has anything to add I think we can call consensus fairly soon... -- samj _in^out 10:19, 23 March 2010 (UTC)

Yes, "criticism sections" always give me acid eructations, they simply always are a very bad idea. Mahjongg (talk) 00:19, 26 March 2010 (UTC)

I disagree with that generalization: sometimes a criticism section is appropriate, and this is such a case. Also I disagree with your statement that nobody is criticizing the term, since that really is the point of Larry Ellison's critique, as well as some of Richard Stallman's remarks. However, because the criticisms go beyond arguments purely about terminology and semantics, it would be appropriate to shorten the title from "Criticism of the term" to simply "Criticisms". Yes, one could bury these points elsewhere in the article, but I don't see that as an improvement. There are legitimate arguments to be made that the term "cloud computing" has become a heavily over-hyped buzzword, and that the cloud computing model, whatever that may mean, is not necessarily a panacea. For the sake of giving the article some balance, it would be best to retain the criticism section, and I encourage anyone who feels so inclined to improve it. JCLately (talk) 02:47, 26 March 2010 (UTC)

Also I would say that the quote from John Rymer is pretty lame, and that sentence should be removed. The entire rest of the article stands in opposition to the criticism section. Do we really need a criticism of the criticism within the criticism section, and such a weak one at that? As to the preceding line about Oracle launching a cloud computing tour, well that's pretty amusing, but it really doesn't contradict Ellison's point, so it's a bit of a non sequitur. I wouldn't object to taking it out or leaving it, for amusement value. JCLately (talk) 03:02, 26 March 2010 (UTC)

Larry's criticism is nonsense, and it's not the first time he's trashed a technology before fully embracing it. I believe comments like John Rymer's as well as the company's subsequent actions (not to mention his recent change in heart all but invalidate his earlier criticisms and as such they should be balanced per WP:NPOV or ideally removed. -- samj _in^out 07:50, 29 March 2010 (UTC)

RMS raises some interesting and valid points but generally goes overboard (as he is known to do - and I'm as big a fan of free software and the FSF as the next guy). The issues section was created to address these in turn, and it's doing a pretty good job I think, without becoming a WP:COATRACK for out of context soundbytes.

In any case, if you argue that the "criticism" section needs to be kept then in order to avoid giving it WP:UNDUE weight and maintain WP:NPOV we'll need to up some gushing quotes about cloud from Amazon, Google, Microsoft, etc. and add a "praise" section. -- samj _in^out 07:50, 29 March 2010 (UTC)

Diagram

At the top of the page is a diagram which is supposed to help explain what cloud computing is. There's a wavy outline; that's the cloud, I guess. Inside the cloud are names of a bunch of vendors (I think, some of them I don't recognize); all of them are connected to each other (meaning they cooperate with each other?), but none is connected to anything outside. Outside are a bunch of unlabeled computer symbols, which I suppose are personal computers (but they all look alike; isn't one of the ideas of cloud computing that you could use a computer-like tool with very little power to run a program in the cloud that takes considerable power? or am I just thinking of a dumb terminal?). There are lines drawn from each of these computer symbols to the cloud, meaning what, they get rained on? It reminds me of one of those diagrams where everything is connected to everything, and IMHO it's not helpful at all. I don't know if there are better diagrams; maybe this one :-). Mcswell (talk) 21:32, 10 April 2010 (UTC)

Unfortunately there's copyright problems with that image ;) I created the diagram a year or two ago to indicate that the cloud computing services were provided from "within" the Internet, that there was some connectivity between the various services (think Intercloud) and that the many edge devices were just clients (be those PCs, mobiles, etc.). Painting a picture of cloud without getting into original research is *hard*. -- samj _in^out 14:37, 11 April 2010 (UTC)

Comparisons

I would find it helpful if the "Comparisons" sxn could be expanded, or if there were links to somewhere that expanded on the comparisons. I have a particularly hard time understanding the diff between "Client-server" computing and cloud computing, and the couple-line comparison didn't enlighten me. The hyperlinked Wikipedia article on client-server computing doesn't mention clouds, so it doesn't clarify the situation. Can someone help me, by expanding this sxn? Or else I'll conclude that Larry Ellison is right in saying that "cloud computing has been defined as 'everything that we already do'" (under Criticisms, at the end of the article). Mcswell (talk) 21:44, 10 April 2010 (UTC)

I was trying to keep a lid on this section, but in terms of client-server specifically, cloud computing can be seen as the second major paradigm shift (the first being from mainframes to client-server and the second being "back" to a centralised "cloud" of computers). Look at it like this - the Internet kept us amused by passing packets between edge nodes (some clients, some servers, some both) - like the electricity grid without power stations - but now we're adding computing capacity to the core. -- samj _in^out 14:35, 11 April 2010 (UTC)

Mr Stallman

the text says ``Richard Stallman, founder of the Free Software Foundation and creator of the computer operating system GNU`` ... GNU is no OS, it's a licence ... Is it ok to modify that? Ph.