The Basic Economics of the Cloud

There's no question that cloud computing is the latest and greatest over-hyped technology--only five percent of the large-enterprise respondents in a new study plan to adopt a cloud-only IT strategy, citing security as a top concern.

That said, it's a mistake to relagate clouds to the JUNK heap.

The city of Los Angeles may point the way to the future of cloud computing. I recently spoke with officials there about their $7.25 million Google apps cloud implmentation, which will impact about 30,000 city workers. Rather than simply signing on the dotted line and passively "renting" biz apps, the city spent weeks hammering out SLA details with Google and reseller Computer Sciences Corp.

The result: a contract that replaces hype with penalties that have razor-sharp teeth. And by offloading the responsibility for commodity-class biz apps and support systems, the city expects to save $5.5 million over five years, compared to modernizing its current programs. 

Ironically, security concerns--now one of the biggest inhibitors to cloud adoption--may ultimately help to promote it. That's because the choice organizations now face isn't between bullet-proof in-house systems guaranteed to keep hackers at bay versus unproven cloud security. Despite their best efforts organizations remain vulnerable today, which is why the Hannafords of the world may eventually decide to roll the dice with outside services providers.

What happens when (not if) there's a large-scale breach at a cloud provider? In LA's case, its contract entitles the city to unlimited monetary damages in severe cases. That's a lot more insurance than it used to have.

Nobody - and I mean nobody - needs "cloud computing". it is a JUNK IDEA that was floated years ago and nobody could see the use of it then either.

* PCs are hypercommodified, and I can store what used to be warehouses full of data on single disks. Cloud is not needed or wanted for Backup.

* The idea that I should RENT the use of software on a cloud which I OWN on my PC is absurd, and the idea that "if I don't subscribe, I'll miss out on important features and upgrades" is also NONSENSE. People can still do 99.99% of what they do now using software from TEN YEARS AGO, for example. How many bazillion copies of MS Office are out there on PCs? Lots. I won't have a problem finding a Word Processing program when I need one, that's for sure.

* In today's world, with breakins to supposedly "secure" installations occurring every month, it's absurd to think vouchsafing my data with some remote internet company makes any sense either. It would make more sense for me to store my data with a personal friend.

This is nothing more than "we'll pretend you have these needs to fill" coupled with "we'll spend a lot of time telling you how great it will be when you can finally get it." As if anyone's actually waiting for it. (They're not.) And as I said, they tried to pull this same scam on us in the late 90's (remember NC's?) That died from the CO2 from people yawning.

I predict that this idea will die a conclusive death within 4 years of being launched. Too bad it can't be conclusively refuted before it is uselessly sold to people.

Don't spend your money on it - and for certain, You don't need Cloud Computing, do you?

Thought not.

it'd be kinda nice to know what the "economies of scale" discounts are as data centers grow, though.   do I save 5% if I operate more than 1000 servers?  and where do the savings max out?  I assume there isn't a clear difference between the operating costs of 100,000 servers versus 200,000 servers (on a per-server basis).