Before Apple, MP3 music players were everywhere. So, it wasn't that the iPod was such a revelation; what Apple did was figure out that people would prefer to buy individual songs―if they could download them easily, quickly and cheaply. By thinking of music as a service instead of a product, Apple was able to identify and provide a new method of consumption. One that, because of Internet delivery, completely changed the way people purchased and consumed music. Now, with the iPhone, that paradigm is expanding to all manner of Internet-enabled services.
Cloud (also enabled by the Internet) will have this same affect on IT. Not for IT's customers―they will probably never know that the application they are using is coming from the "Cloud"―but on IT itself and more importantly it's relationship with the business. Cloud is potentially the game-changing paradigm that will rewrite the way businesses view, use and pay for technology.
According to the report: "The most important value the cloud brings is not lower costs. It is improved agility, not just for IT, but for the business as a whole." This is accomplished by allowing companies to off-load anything and everything about their IT infrastructures to an outside provider. While this is not new, the way it's done is. With Cloud you contract to only for the services you need as you need to consume them. You don't have pay for 24/7/365 anything. A great example, said Neal, is dev/test.
At pharmaceutical giant Eli Lily last summer, a researcher went to IT with a request for 50 servers to help in the development of a new drug. Given the importance of the project IT agreed but said it would take 90 days to get up and running. Well, at big pharma delaying a big drug project can cost upwards of $150 per second so, needless to say, the researcher was a little adamant about getting the servers sooner. So, after he put in his request, instead of waiting, he went to Amazon's EC2 Cloud offering and provisioned the servers there. After a day or so of trial and error they were up and running, said Neal.
What they discovered was a bit embarrassing but, ultimately, very rewarding. After running through some testing they discovered that to do the research internally would take 250 servers, not 50. "So, how would you feel to be the guy who beat up everybody to get your servers only to find out you were off by a factor of five? Pretty stupid, huh. But this is Amazon so they changed 50 to 250 and clicked 'Go' and ran the who shebang that day."
That is the power of Cloud. Change one parameter and get the power of 200 servers at your finger tips. No provisioning. No waiting for UPS. No waiting for IT. You need it, you click it, pay for it, and you're done ... all for around $.10 per minute.
"What the promise of the Cloud is, in part, to begin to segment apps so that ... for those few times a year I need extra capacity, I reach out to the cloud," said Neal. "So, I pay with occasional OpEx for what used to pay for with CapEx and, in the process, I could probably double, tripe, quadruple the bang I get out of my data center" which is still runs at 10% of capacity at most organizations― " ... a level that would not be tolerated in any other expensive capital asset," the report states.
Wake Up Call
For IT, this is a shot across the bow. Cloud promises to finally remove the barriers between CPU's and the services that depend on them. It's like Moore's Law applied to service: every 18 months the level of Cloud-based services you will be able to offer could double. Think of it like a RAID for CPU's, said Neal: a whole bunch of cheap servers, provisioned as needed to handle whatever load is required as it is required. You know that a few of the servers will fail (maybe more than few) but that doesn't matter. In the Cloud, everything just fails over to the next available processor. For all you know, you application and are being served up from servers all over the country. The point is, it doesn't matter. Location in physical space is irrelevant. All that matters is service delivery.