I only know of one reasonable use for cloud computing on a large scale, and that's with MMO(RPGWXYZ) servers. Times such as when a "sub-server" is not necessary, it wouldn't even need to be loaded and can save resources for where users actually need decent response times. Perhaps educational servers, or web site that are literally HUGE (I don't think they exist unless you count the whole internet). Of course the obvious thing to most of us is that if MS servers weren't so unstable, there wouldn't be a need for virtual ones, becuase when a hardware system goes down in a stable environment, you don't want more than one service to be effected if necessary, or in future cases, entire web services going down for those who don't have the millions to have server warehouses with multiple blades running multiple virtual servers that spread across multiple servers to manage load. Who exactly need this yet besides the ones who already use it? I honestly have no idea. But as long as IT management continue to be migrated across from HR departments with the "anyone can manage IT" mentality, we'll continue to be overwhelmed with hype and bells and whistles, while functionality, efficiency, and reliability are sacrificed.

But hey, why break a trend :P It's not like the economy can't handle it, IT is full of cash <-(sarcasm) The great Moore's law will solve it all, along with world peace and world hunger, and the vacuous debt void the governments keep tossing money into. The future isn't bleak, it's just stupid.

So I move all my data onto a cloud server and select a number of apps to use - Office, Photoshop etc. I start paying, I dunno, £20 a month or so to use all the space, the apps and the CPU/memory grunt, all done over a (hopefully) very secure and (probably) very slow connection over the net.

And then when my ISP's network goes down, I can't do any work. When the Cloud Service Provider's servers go titsup, I can't do any work. When their own ISP goes down, I can't do any work. When their SAN goes haywire and it turns out their backups weren't working, I lose all my data. When I finally decide I'm fed up with paying £20 a month, I lose all my data.

Yeah, I can go for that. No, really.

Why would anyone with a choice hand their private data, and processing requirements to a third party?

Especially given the Governments inability to protect personal information, BT's complete failure to provide a private/secure UK communication infrastructure, the private sector's eagerness to buy/sell/swap any and all personal information about people, and the ICO staff apparently too confused by all this complicated modern computery stuff, and Richard Thomas wringing his hands over the surveillance society.

Cloud computing is *utter* nonsense. Its the emperor's new clothes. Its the next dot com bubble.

Stallman has it about right; "Do your own computing on your own computer... You're putty in the hands of whoever developed that software".

I'll get my coat, because its real, it has substance, and it provides more protection than the ICO.

Isn't this going round in circles a bit - use servers/web for stuff you need access to from wherever, use your machine for the personal stuff that you don't want others to see. I really don't see a future in "The Cloud" as it will take too much effort to supply the services that are required, imagine trying to get ALL applications onto the Cloud - that's what businesses and individuals will want.

There is mileage in having your DATA available anywhere, but applications? Why?

Also in the UK with the already overstretched broadband/internet access there are going to be bandwidth problems as well.

If you can run it locally, do. If you need access to data - whack it on a secure server or web host - but running the apps over the internet with effectively a dumb terminal at the end? Why?

We stopped doing it years ago - connectivity goes down, you loose the LOT, all your apps, all your services etc etc - not good for your business, and then who is to blame, hmm ISP's are going to love THAT clause in their contract - "Due to lack of service we were unable to do any work at all, therefore we are suing you, Mr ISP, for loss of money/business"

We make a software PBX platform based on Asterisk. Normally we install stand alone dell boxes on customer sites. Sites are typically 20 to 250 users/extensions.

A local hosting provider has invested squillions in a HP Blade / VMware cloud platform. Thought we give it a try and deploy our PBX system into a virtual server on it.

Didn't expect much in the way of results to be honest.

So guess what? - works like a dream. Complete enterprise PBX with all the trimings. In a could. Fantastic. Have just completed the first commerical roll out this week. Customer has 10 sites, each with a hardware PBX. Not any more...a single system with baby VPN routers on each site for the handsets to connect back.

Big wins for us are:

no more server hardware to support (major headache gone)

no more standby PBX systems required (VMware image can be restored through a web console in less then 15 mins)

much cheaper for customer in long term

server can be beefed up through a control panel

used to take us 4 hours to install and build a PBX (inc OS install). - now takes 30 mins!

If you make so sell software systems for a living then worth investigating..

so, yeah...thumbs up!!!!

I've been playing with Google's AppEngine for a while, on my own time rather than for work purposes. I had a couple of projects I wanted to try and get online, and AppEngine is (in typical Google style) free until your page gets popular, so I couldn't resist having a bit of a play.

I'm pretty impressed with the platform. The tutorials are good, and there's a decent community to help out with queries. I managed to get a simple test page online within an hour, even while trying to wrap my head around Python (which I'd never used before).

The database architecture is probably the most interesting aspect, it pretends to be SQL-like in syntax but under the bonnet it's very different. You have to unlearn everything you know about relational database design in order to create something that will scale past a few users.

I do share other posters' qualms about putting my data in "the cloud". I wouldn't trust business critical systems to it until Google backs it up with a solid service level agreement. But for a beta service it's pretty damn reliable, and the team always responds quickly to issues.

Overall, I really like the concept, and Google seems to do it well (although I've never tried the competition). I hate the term "cloud computing" though - it's just web hosting that scales automatically, no fancy name required, but I guess it's too late to put that genie back in the bottle.

a long long time microshaft tried to get, what was then thin-client computing in the field , that would have meant charging people for usage . That failed .

Now as the sales of Microshaft products have gone thru down thew pan ( check out VISTSA Sales) and there are many FREE antlernatives, microshaft again has come up with a 'bright' idea, to charge customers for 'renting' space on thier servers, and likely to 'rent' applications

Lol the market goes round in circles, once again

PS in a lot of cases the FREE products are less buggey, more reliable the the MICROSHAFT ones !

Azure has some warts, but it is at alpha. The concept of storing your data both locally, so you can work on it when the network is down, and in the cloud is compelling. If you are worried about sensitive data, either don't put it in the cloud (most companies have tons of public info) or encrypt it (it's simple to add that step to a .NET app).

Save your data center for your core biz.

Yep, me too - music is on mp3tunes (until they get sued out of existence), photos are on smugmug, emails are on fastmail, contacts are on Plaxo, bookmarks with delicious and docs are with google - all tied up with a Netvibes homepage. All available locally on my MacBook using iTunes, iPhoto, Mail.app, Address book, Firefox & gears. A bit of manual synchronisation required sometimes, but wherever I am working or staying, I have access to everything. Suits me perfectly!

Hmm clouds - clouds in the sky - networks in the sky.....

I for one will welcome our new Terminating AI overlords.

To echo the positive comments above I think cloud computing has awesome potential! Check out here to find out more about Amazon's proposition, https://blogs.conchango.com/johnbrookmyre/archive/2008/12/09/get-on-my-cloud.aspx.

I'm another Appengine Kool-Aid drinker, but I'm still very much alive. My company has provided web apps on it's own Linux servers for years, and I know the cost of scaling a popular app - in terms of dollars, stress and customer dissatisfaction.

So when one of the more reputable (well last year, anyway!) online giants offers to host my apps on a platform with scalablilty and redundancy I can only dream of, and at a cost far less than hosting it on a single server, I give it a whirl. There are definitely challenges in moving from an RDBMS to Google's data store, but so far they have been surmountable, and still easier than designing an app for a scalable database system.

Python, once you've taken the plunge, is excellent. Efficient, logical, concise and readable - the code is practically self-documenting, reducing costs and making maintenance a pleasure.

We are currently completing testing of our first major appengine application, which has gone well - the platform has performed perfectly. We expect this one to be fairly big, generating most of the revenue for our company next year, and for once there is no trepidation about how fast it grows.

We have some gripes - currently there is a quota system in place which limits resource usage, effectively cancelling the scalability benefits. This should disappear when the billing system arrives, which hopefully will be in the next couple of months. If something goes seriously wrong, it will be very frustrating that the only support is a user group, hopefully Google will provide better support for paying customers.

As far as fears about giving your data to Google, I'm fairly sure the only access they will have (or want to have) is via the good old Googlebot, checking our public pages. It's a little like the phone company - we all know they could listen to our calls, but anyone who has worked for a telco knows that they can't be bothered - the value of any information they could reap is far out-weighed by the cost of gathering it.

Only time will tell, but to date the Kool-Aid tastes very refreshing.

Those who buy the outsourcing story will go for clouds, too, for seamless exchange between themselves and their partners.

Pain? What pain?

Could somebody tell me is it "cloud computing" if the data and apps are all stored on your own servers internally and are accessed across the private network?

Or is that just called using Thin Clients??

Cloud seems to imply it is stored somewhere that it can be accessed from any computer with an internet connection?

Whats the difference between cloud computing and settings up a few Thin Clients?