What you think about cloud-computing?

…i like more the good old way…my hardware, my software and not be afraid of “marketing-file-watchers”…:wink:

It’s a great concept, but I would never, ever use it in practice. Like you, I want MY hardware, MY software … and most importantly, MY DATA … under my control. I don’t want it stored on a server 3,000 miles away.

The whole point of personal computers is that they’re personal. :slight_smile:

I’m sure there will be some people who use it, just because it’ll be convenient: no matter where you are or what computer you’re sitting at, you can log in to your remote desktop and gitterdone. But that will change as soon as the first major data breach occurs.

And it will, count on it. Take it to the bank. :wink:

Our company runs its own mail and web servers, in spite of the fact that there are some great deals available nowadays from the ISPs and hosting providers. In our case, it’s not a question of saving a few dollars here and there. Security is the #1, overriding issue. The mail server sits about 30 feet from my office where I can keep a nice, beady eye on it. :slight_smile:

Fad is a fad is a fad.

Yes thats right!

well… cloudbusting is music to my ears (kate bush) while cloudcomputing is… evaporating to me…

I wouldn’t be so sure. You can definitely make the case that netbooks are an example of computers going full circle, and becoming more like dumb terminals again - and I can’t see the netbook concept disappearing soon.

I don’t think the ‘cloud’ will replace conventional computing, but I could see it having a place, parallel to the powerful stuff we use now - which in most cases will probably end up being run entirely networked anyway, just with mini blobs of centralisation under the end user or business’ control.

Lots of small clouds. Droplets, or something, we might call them…

It reminds me of computing in the early 60’s. 1960’s that is. In many ways we are all just terminals attached to a massive host. As I see it “personal” computing is still in its infancy.

Moving forward to the past.

I like having stuff locally but cloud computing is hard to avoid. I don’t really use Google docs much but I use a lot of other Google services (e-mail, calendar, reader, voice). I have a Smugmug account which I use a lot and another account on a UK photo site. I also have data on Foxmarks and Zotero servers. And I use Evernote to clip stuff. And I have an Amazon S3 account, which I haven’t used much but the storage is dirt cheap and easy to use with JungleDisk. I’ll probably use to sync my Zotero PDF files. A lot of these services work across platforms and allow access to key files anywhere or allow data syncing on computers at different locations (regardless of operating system).

Working in a data center / ISP I thought I would chime in.

I firmly believe cloud computer is going to have a huge impact over the next few years. For example, virtualization has taken very firm root in the enterprise and data center spaces whereas a few years ago one heard similar arguments about running VMs. However, tThe benefits of virtual machines, when correctly deployed, are considerable.

Virtualization has a primary draw back. No matter how dynamic the hypervisor is, no VM can exceed the resources of the physical server it is hosted on (basically). That is where clouds really come into play.

Let’s say you run Linux Rocks magazine. You host your site in a data center and have redundant servers in differing geographic locations - of course. Your traffic is fairly steady through the month, with a few very significant exceptions. When a few issue comes out, for a few days, you are swamped. And twice a month you have pod casts which are very popular. On these occasions your load balanced hi availability clusters are nearly brought to their knees.

So what can you do? Lease / buy bigger servers of course! Only that can get very expensive, very quickly - especially when you are talking about replicating this in multiply physical sites. Why buy bigger iron that will run underutilized 80% of the time?

Enter cloud computing: Now you can run some of these services in a dynamic cloud that can allocate more resources as needed, perhaps even provisioning more servers automatically - but only for as long as you need them. The cost savings can be considerable.

At the data center I work in we are experimenting with cloud computing via Eucalyptus and are examining offering a cloud solution to our customers, some of which have nearly identical situations to that above.

And with Ubuntu 9.10 - Karmic Koala - focusing heavily on Eucalyptus and the ability to quickly deploy and manage private clouds, I think cloud computing is here to stay:

Eucalyptus

https://lists.ubuntu.com/archives/ubuntu-devel-announce/2009-February/000536.html

Clouds are not at all intended for PCs as by definition they are clusters of computers - so running one at home is unlikely. But in the enterprise, where servers can be put into a pool while being re-provisioned, or in data centers - they make very good sense.

Home users may fear putting personal data into a cloud - understandably so. But this is not really what Clouds are used for.

I would love to see Suse embrace Eucalyptus (or similar projects) and integrate it with the ease they have done with Xen and the hypervisor available through Yast in 11.0.

Though there is some hype involved, I believe Cloud computing is about to take a very huge leap forward. Many sites you visit and many web based apps will very likely soon be serving you from the lofty heights.

Cheers,

Pete

You raise some valid points, but I think it will depend on the business and the application. Your example of the “Linux Rocks” magazine might be a case where it’s useful. A company that edits, moves and writes a lot of documents might decide that it’s worth it.

But another large business with lots of proprietary data (user databases, account and sales records, etc., etc.) is likely to want that under lock and key in-house. It’s just part of the cost of doing business. If you need another server, you buy it.

I don’t think cloud computing will take off like its proponents think it will. VMs are making an impact, but only in specific environments. The average desktop user certainly doesn’t use them (and in truth, probably has never even heard of them).

There was a story on Slashdot just this past week about a data breach in Google Docs. Turns out it was minor and it was quickly fixed, but it illustrates the very reason why people like me aren’t interested in it. If you give a crook a single target that promises a large payback for the effort, it’s going to be tested more than many smaller targets, each with its own security.

You mention netbooks. They’re handy, but if you look online, one of the most common questions is how to sync the data on the netbook with the “real” computer on the desk in the office. :slight_smile:

Ergo, et sum: will Cloud computing become popular? Sure, for certain applications. Will it ever become so popular that it eclipses the personal desktop computer? Heck, no. Let me quote ionmich here and make one more point:

(ionmich) - It reminds me of computing in the early 60’s. 1960’s that is. In many ways we are all just terminals attached to a massive host. As I see it “personal” computing is still in its infancy.

… and I would argue that there are still people in computing who apparently miss those “good ol’ days” and want to return us to them. Some of these people are Cloud Computing’s biggest proponents. :slight_smile:

I think smpoole7 makes the big point - cloud computing has its applications. Will it become the dominant paradigm for computing? Probably not, for the reasons smpoole7 mentions - cloud computing aggregates valuable data in fewer places. Obviously that stacks the incentives for unscrupulous crackers.

If anything cloud computing (on the consumer-level) is just a larger scale version of what companies & institutions have been doing for a long time: Hosting some/all the data their users create and use on servers. The only difference is many companies restrict access to machines adminstrated by the company, while cloud computing is being touted as ‘any computer, anywhere’ - desktop, notebook, netbook, cellphone, public terminal etc.

Most people will use it for datatypeA but not datatypeB as they are comfortable. I chuckle to myself everytime someone says cloud computing has no future and then goes on to check their e-mail with IMAP. The idea is the same, it’s just the extent and specific implentations that change.

I think that’s true. There’s probably a lot of stuff running on Amazon’s cloud services. If you want to run a Linux server on Amazon there are lots of prebuild Amazon Machine Images (AMIs).

Smugmug is a good example of a company that makes heavy usage of Amazon Cloud Services and saves a truckload of money.
Amazon S3: Show me the money
Amazon S3 = The Holy Grail
SkyNet Lives! (aka EC2 @ SmugMug

With this sort of stuff and some Amazon Machine Images you could run a large Internet company from your home ;-).

Nice that here are so many posts about cloud-computing…

And what you think about the a future-fiction with cloud-openSUSE? Would you use something…?

For me is only a personal computer a real pc and cloud-computing has not a chance…no way!

As a simple home user, I find that I am using net apps (I think…) more and more. Also, as I don’t put any sensitive data on the internet (unless I’m more naive about data than I realised), I find the convenience of clouds greatly outweighs the risks. To be able to check my email from anywhere with a phoneline is awesome :).

Of course, if I want to access data at home I could always ssh…

Cloud computing = a poor attempt at a ‘buzz word’ to use to describe at a poor excuse for a ‘reinvention’ of the Dumb Terminal :wink:

I may be wrong but I think cloud computing will be like shopping online. People will shun it at first because they don’t understand it but as they sample what it has to offer they will use it more and more. Most people I know do some sort of shopping online.

I can’t see my employers putting their financial statements on the cloud, though. Like others has said, it will have it’s place and it’s uses.

Cloud services are good for Linux as a lot of the big providers are using Linux to run massive server farms. Google is supposed to use somewhere between 500K and 1M servers–I don’t think anyone knows for sure–but they run a custom Linux. I think both Amazon and IBM also use Linux. Not sure about Yahoo. I don’t know what other operating systems are widely used. I can’t imagine Windows is used much except on Microsoft’s own server infrastructure. Presumably they sorted out the Windows scalability issues that plagued NT4 and gave Linux its big break way back but probably not the license fee issues. :wink:

This really isn’t “cloud computing,” but it’s analogous from the user’s perspective: LTSP - openSUSE.

We currently use LTSP to get new life out of older PCs, primarily for those users who just need a workstation to check email and do simple document work. We had a bunch of older Pentium III’s that are now back in service, thanks to LTSP. Saved the company a BUNCH of money. :slight_smile:

We’re not using the Opensuse variant, but we may give it a try this year when we get time. The primary difference between LTSP and true cloud computing is that the latter is Internet-based, where LTSP essentially sets up “thin client” LAN workstations that are all in-house and secure. (Which is why we use it; see my comments above. :))

But to answer your question, I don’t see any reason why Opensuse couldn’t be the backbone for a cloud approach. The biggest limitation to making it look and run the same, IMNHO, would be limited bandwidth. LTSP is surprisingly fast on a LAN. I don’t think it would be usable over the Internet.

This is a great discussion!

I had struggled with accepting online services that are running on cloud-computing-like services (Gmail, Calendar, …etc). But it seems like it is going to take place. For example, companies like Salesforce.com provide hosted CRM services and are doing great business. Imagine that! … companies are trusting hosting customer and sales information on a third party just because it is cheaper, easier to manage, and provides ubiquitous access.

I think what we need to start thinking about is how we can allow data and applications to be hosted on a remote data center and still provide the confidentiality properties that we get from hosting it on our own hardware.