Have any of the major Linux desktop distros taken an aggressively anti-cloud stance yet? I know Mark Shuttleworth of Canonical (Ubuntu) keeps talking about how important cloud computing in, and that seems to be the general trend in both free and proprietary software.

It seems like where you have a large-scale movement to take software and data away from people's hard-drives and throw it up on a webserver where it is someone else's property and could potentially be closed-source and/or available at a fee, there is going to be some pushback. I haven't seen it yet, though, at least not at the level of an entire popular distro saying "this isn't for us". But I don't keep close track of trends, so I am wondering if one of the distros has said that and I've missed it.

I think it'd be great if a distro did take that stance, if for no other reason to be used as a fallback for people -- i.e. "I'm going on the cloud with my primary computer or OS, but I'll use this one for caching data locally and with a full suite of software I can use locally if need (like a lack of net access, servers going down, service fees being unaffordable one month) and/or preference strike at any given time". And, heck, I think there have to be some folks out there who just hate concepts like software as a service period.

It's taking a while for people to notice, yeah.

I doubt a distro will take such a stance. Though there may be "cloud distros" later, which are very light to install because all the work is done online. However, expect this idea to become unpopular fast as these servers start losing peoples data. This has already been an issue, as well IIRC as some companies going bust and their clients data ending up sold to third parties. (Not to say, if the server is in the USA, the govt can just order that your hosting co turn over all data to be examined by a third party.)

The main argument I've been hearing in favor is that it allows your data to be available anywhere. You've articulated the main argument against. It is difficult far some people to see the difference between doing, say, your word-processing online and the way FOSS project share their code.

Personally, I don't see the need to have your data available everywhere. I carry the files I'm working on with me. BFD.

One of the reasons for the GPL 3 was the introduction of web apps since GPL was created. However its focus is on preserving the freedoms of users and not the security issues you mentioned. Some Software as a Service projects may be created under the GPL and the license needs to protect users and contributers from some of the abuses you mentioned.

There was an article in one of the linux magazines about how the prevalence of web applications and cloud computing will cause a paradigm shift in security away from relying on closing ports in a firewall and towards a data centric approach. The information being safeguarded is at the core and application security is just around it. In a self contained LAN, the diagram may look like circles. If the data is in two locations, it will look like a dumbbell (which I imagine you think is fitting).

The standards for this non firewall centric approach are still being developed, but as this will probably be a larger concern for businesses, I think that rather than a particular distro resisting cloud computing, the Enterprise editions of each distribution will be more likely to develop the tools needed to maintain security as the network becomes extended to include SAS services.

The solution may involve extending the RBAC approach of SeLinux to encompass the entire network including offsite SAS providers. Companies may prefer only using providers that use OS web applications and open protocols. This will allow auditing how data security and authenticity is maintained instead of relying on blind faith which you have to using propriety software and standards.

One thing I've never liked about technology is that it seems like a trend develops and everyone jumps on it as quickly as they can. It actually, in some ways, would offer users greater freedom to have different operating systems and so forth take difficult approaches. Linux and the open-source movement allow for that in spades, and intuitively seem to get that more than proprietary software, but I sense a little bit "follow the leader" in this whole cloud thing.

Rather than having Windows and Apple both do it, and all the Linux distros hop on board (or try to beat Windows and Apple to the punch and get there faster), it'd be nice to have some Linux distros and aps take a different path. That way, people can run different distros for different reasons, or pick which approach they prefer.

It seems to be that a Linux distro or two could really differentiate itself/themselves in that way. Basically be saying "Everything's going to the cloud but us. Don't like the cloud? Have reservations? Want a second OS where you own all the software and get to control it more fully, like you're used to? Give us a try!". Rather than being, I don't know, let's say the 5th or 10th most popular cloud based OS, an OS could create market share by being the top non-cloud based (and non-cloud software based) OS in the market.

I've got to think that there *will* be a decent chunk of people not happy with the cloud who'll be looking for something more like what we've had, but yet still modern and updated.

Personally, and call me crazy on this, all other things being equal, I'd like to have more control over my software and documents by keeping them locally, and to not have another bill I have to pay each month. One issue with software as a service is that it often takes something that is now without recurring costs and makes it like your Internet bill or your cable bill where you have to keep paying and paying and paying through the nose. I'm not surprised businesses like that idea in general, but I think it could be profitable for a business to be the one guy not doing it to provide an alternative, or a good community service (ironically, by not being service-based ) for a community-based distro not as tied to a corporation to embrace.

Cloud computing is all about the applications, so pretty much by definition it's a non-OS issue. Distributions are operating systems, so there is pretty much no relevance to cloud computing.

Don't like SaaS? Don't sign up for it. How much more simple can that get?

Meanwhile, the entire computing world will pass you by because the economy of scale means it makes much more sense to do many things in a shared-tenancy environment where resources are allocated on-demand.

Does this mean you should run out and sign-up for Google Apps? No. Google Apps in particular are rather poorly done, from what I've seen, and Google doesn't exactly have a comforting track record of how they handle personal data. On the other hand, there are quite a few companies in the commercial space that have done very well with SaaS, such as the ubiquitous Salesforce.com .

If you resist change, you're just going to get run over by it later. It's the early-adopters who get to help shape the direction.

In part, that's a good point, at least in the near-term. In the long-run, though, they may move core components of operating systems onto a cloud so you are basically running a terminal skin. It might not happen, but I wouldn't be surprised.

Having said that, even in the near-term, we could see certain things moved out of the typical install CD bundles. Not that getting them from a repository is a big deal, of course, but it is a slight hassle, and then down the line they may move from a repository to having to download a package and do it yourself from a web site. And then, ultimately, you may see popular aps like OpenOffice and so forth move to the cloud in newer versions -- and they'll only be forked if a distro or two actively support and push non-cloud software solutions and would consider backing forks to keep non-cloud options available.

For the most part, I probably won't. I exempt a few things where it makes sense for me (like e-mail).

This is exactly my issue. For me, and I don't think I'm completely alone in this, my needs and wants are probably mostly better met by software and data that's based on my hard drive, and a net connection to do things that can only be done through the Internet, or that I prefer to do through the Internet. But folks like me may in the end not be given the option to stick with the system that works best for us, because big businesses will decide what is better for them, and, granted also for some end-users who have different preferences than I do, and new software versions and so forth will ultimately be geared toward that end. The end result is that folks like me will be left with systems that are subjectively worse than what we have now -- either because we feel forced into the cloud and don't want to be there, or we have to use outdated aps.

Technology needn't march forward in only one way. It can take several different paths. That's in a sense part of the point of the open-source movement and the ease of forking that's part and parcel of it. But it seems like many of the movers and shakers right now just want this cloud thing and seem insistent on ramming it down people's throats instead of saying "Here's an option you can try, we'll make sure you're covered if you don't like it also".

I don't see anyone going for an anti-cloud distro simply because most people don't see a point in limiting people's choices. The cloud stuff has so many issues to resolve; just look at this situation for example:

Bob connects to the cloud via LaptopA. At some point the internet connection is lost. Fine - the software provided a local copy of his files and he's now running local programs to edit them. Bob is now on LaptopB which is connected to the cloud although he didn't have a chance to sync LaptopA. Bob edits some of the same files (why not, since he has access to them) and these files are synced to the cloud. Now Bob is back on LaptopA and the cloud has returned. What happens to Bob's files now?

Now how about all the security and privacy issues which come up when you talk about corporate data? Do you need to communicate only with certified clouds? Backup datacenters have been providing acceptably secure backup of corporate systems for some time now, so storage and transmission of data is no great hurdle. Running remote software to process that data is also nothing new. The only thing new is that people using the services can be assumed to work from multiple machines which do not have reliable access to the datacenters.

Technology evolves and typically only the strongest ideas, or strongest implementations survive. Strongest might not be "best" in a purely theoretical sense, but certainly market backing, ability to execute, stability, and interoperability all play a big part. It's pretty clear that for the providers of the technology, and for many of the consumers of technology, moving to large-scale implementations with resource sharing and virtualization is the way to go.

Asking someone to dedicate resources to providing a dying and out-dated model isn't very realistic. It's basically like asking ship yards to keep building sailing ships because you don't like steam power very much. It would make no sense for them to do that. If you want a sailing ship, build it yourself. Just don't be surprised when there isn't much demand for your product and the ship yards that converted to producing steamers are making all the profit.

If you're scared about the direction, start using the technology and provide feedback to influence the direction. No one is going to pay attention to a few lonely voices demanding that progress stops. The best way to protect your interests is to cooperate, because opposing the inevitable is a sure way to lose.

I'll admit, my understanding of technology is far from perfect. So, maybe people can walk me through this a little bit. Right now, things are basically the way I like them, new features would be nice, more bandwith would be nice online, I'd like flash to work on Linux, but I don't need anything fundamentally overhauled about my computing experience. I like having a music player and my music on my harddrive, a word processor, etc.. If I lose my Internet connection at some point for whatever reason, my computer is functional and a non-brick. What does the cloud have to offer me? I don't want to access my stuff anywhere, I take my laptop with me if I'm going to need my computer somewhere. I don't want to pay more fees, I don't have much money.

The cloud is basically making my aps slower, limiting me to accessing them when I have a net connection, and taking the control away from me and putting it in the hands of giant corporations to alter or apply their own terms of use whenever they want..

It's said that by putting stuff in the cloud, you get the most efficiency possibly, basically, but who needs efficiency when they'll be 12 and 16 core processors and 8 or 16 gigs of RAM pretty soon? Some people already have terabyte hard-drives. Unless you turn the PC into a Star Trek: The Next Generation style holodeck, I don't see why I need more efficiency than a PC is going to be offer in a few years. The limiting factor on me now with my little dual-core 1 gig ram machine, is actually most slow bandwidth speed, not even the hardware. This runs a touch slow on booting up Vista, but Ubuntu boots up at very acceptable speeds (and each build is getting faster). A 12 or 16 core with 8 or 16 gigs RAM and a terabyte harddrive would almost be overkill, but I'm sure people will find uses for them with HD video and 2 second boot times and whatnot. But what's the killer ap that makes even those too slow? Why do we need the cloud to process things better?

In the end, it's just one of those deals that hurts a lot of the end users, other than those with specialize needs. It's a higher cost and less control over our software and data.

I understand that in the end, it may be my only real option. Resistance may be futile. But aren't escaping the Borg-like tactics folks attribute to companies like Microsoft and other companies a lot of the reason for free (as in freedom) software in the first place? Isn't free software supposed to be about choice?

I mean, there are Linux folks and distros who still don't use a gui. I don't have an issue with that, it's kind of cool, actually. But they've basically rejected the gui, which, let's face it, something that's very basic to modern computing in many respects. They can be using a more modern highly evolved unix interface and people keep that up to date, but there's no room for a couple of non-cloud distros and a full set of aps that are kept up to date and keep adding features? Isn't part of software freedom having the option to say "No, they can't take that away from me"? Or is the group of people who are going to reject the cloud really so small that there won't be developers to support it?

I really don't see what you're worried about. People will work on things if it's fun; I don't see traditional apps disappearing no matter what the hype on the "cloud computing".

Actually because cloud computing is more efficient, services can actually be delivered more cheaply, passing the savings along to... guess who? This is the exact reason why a lot of businesses are moving to buying Software as a Service rather than hosting it themselves. It's fundamentally the same for consumers. Essentially it means you need to invest less in hardware and spend less time fiddling with software, saving you both money and time (which is money).

You are formulating your opinions based on a flawed understanding of how the technology ecosystem works and how products and services are delivered to consumers.

For instance, the company I work for sells a product that performs a particular task in under 20 seconds, 90% of the time, guaranteed. No other company in our market has this kind of guarantee, and in a lot of circumstances our competitor's products can take several hours to perform the same task.

What's the difference? Nearly all of our competitors offer the product as an on-site solution for businesses and it requires the customer to invest in tons and tons of hardware in order to run at anything approaching acceptable performance. We, on the other hand, require very minimal on-site hardware and perform all the resource-intensive functions in our datacenters running huge clusters of servers. The trick is that every time one of our customers needs to perform a task, they get many, many servers worth of resources to run their task on--far more servers than they could ever afford to purchase and maintain themselves. Since our customers perform these tasks relatively infrequently, we can have hundreds of customers using the same resources, essentially taking turns.

You can easily imagine scenarios where the same is true of consumers. For instance, right now editing digital video requires very fast CPU, fast GPU, fast hard disk drives, and a huge amount of storage space, not to mention very expensive software licenses. Imagine if you could use a sub $1000 laptop without a dedicated GPU to simply upload your raw video to a computing cluster, and run AJAX-based editing software through a web interface to edit your video. You would save thousands of dollars on hardware, your video would be edited faster, and the software would probably be cheaper as well since the distribution model is a lot easier for the manufacturer. That's just one example, there are dozens of others.


You should really try to understand something first, before you decide whether you're for or against it.

I truthfully don't have much trouble with software. Firefox and OpenOffice are free and work great for me. On Windows I use itunes, on Linux I use Rythmbox, ditto AIM and pidgin. All free as in beer, most free as in freedom. I like games like nibbles and rattler race. I like desklets and that sidebar Vista thing (and I like having them on my harddrive -- google gadgets is not for me). What does the cloud offer someone like me? Granted, Civilization 4 has had patching problems and one of my CDs is scratched, but I'll bet there will be issues with games on the cloud too here and there as well -- and even if I have to pay for it a second time to get a non-scratched replacement copy, it'll still be about 36 times less than I'd have had to pay for it as a service. Sometimes it's nice to be able to pay for something and forget about it, and not have to worry about if you can dig up the money to pay a bill every month.

You're assuming that I have a particular task I need done faster. And then you're further assuming the next generation of computers won't be able to take care of any issues with speed I might have.

I could see the need a little more on a corporate level where there are sometimes complex specialized tasks that need to be done, but even there it's a trade-off between time versus control and money. I could see where it'd be worth it for some corporations, though.

That's a very specialized need. I don't know many folks who edit digital video. I don't doubt that it's an important market segment -- there are folks who like to edit home videos, and filmmakers, and so forth. But I'm not sure it's the average user. I have a $600 laptop. I'd probably be happier if I had a $1200 one instead, but not *much* happier. I'd be a lot less happy if I were running what might amount to a terminal skim and had less control over and access to the software I was running and the data I accumulated.

chort: are you saying is that customers should sacrifice their freedom and control if it means they can lower other costs?

You have been arguing economic and technical advantage. The argument against is based on social and ethical cost. As you urge others to understand the economic basis for "cloud computing", I will urge you to understand the moral basis of the opposition. After all, every one of your arguments could be applied as well to the use of black slaves in the cotton industry. Economic arguments, alone, are not good enough.

Sounds like your business does not value your customers liberty, and your customers will sacrifice all our freedoms in return for their, short term, profits. This may be a profitable way to run a business, now, but is it ethical? It is certainly short sighted.

In the end, though, the term "cloud computing" covers a wide variety of uses. Not all of them are explicitly detrimental to liberty. So we really need to assess the services for their individual merits and costs rather than bandying blanket terms about like this. And let us factor in the social costs when we do this analysis.

Define "efficient". It certainly is not more efficient to put your main processing in another building - owned by someone else. It also puts my work at the mercy of the service provider.

Because a distributed license is cheaper than a per-seat license. But with free software, there is no per-seat. It is also only a matter of time before the proprietary vendors realize they are missing out and recoup.

Which already happens with free software.

However - there is a risk that "consumers" give up their control over their work. There is, for example, a strong motivation to increase lockin to make sure customers stay with one software-service supplier.

On the contrary - you have your opinion based on a short-term goal ... making money. You have not assessed the social consequences, nor have you considered the effect as non-local computing moves into the mainstream. How your competitors will respond. How your software suppliers will react.

Which task and how do you guarantee this? This statement is meaningless without specifics.

Again - specifics lacking. What kind of guarantee?

Clearly exaggerating for effect on zero data - surely customers are investing in kilograms of hardware... not "tons".

Since you won't provide specifics here - I will...

It is easy to imagine a thin-client system served over the internet. The devil is in the details... how easily can customers change service providers? Can they get access to their files? Can they access the service with free software? What happens to their files in the event the service provider proves insolvent?

There are ways of running these things which mitigate the social and ethical effects. What does your company do to address these issues?

While it is possible to imagine an ethical cloud computing service, there is no mechanism which helps guarantee one. It is all too easy to imagine an unethical system, however.

Not so with free software. Instead of promoting remote use of proprietary systems, we could be developing and improving free software.

Apart from the social costs, there is also a more specific opportunity cost involving developer man-hours.

Note: professionals (Pixar, Weta) prefer not to risk their "property rights" to a third party. They use inexpensive, local, applications to do their rendering on cheap hardware.

But - when the video can only be rendered to a closed format and requires a specific client to view it, you've lost almost all the advantage. There is an economic motivation to run your service like this... it makes it difficult for customers to change providers. Then you get to ramp your fees.

you certainly should. You need to think beyond short-term convenience, and try out a larger picture. What sort of society are you helping build?