Cloud Connect's Alistair Croll argues that internal enterprise clouds are a temporary phenomenon, to be followed by a big migration to public cloud infrastructure.
Over the last few months, there’s been growing discussion over private and hybrid clouds. At first blush, a "private cloud" sounds like an oxymoron, particularly if you subscribe to Amazon CTO Werner Vogels' definition of cloud computing: "a style of computing where you have massively scalable IT-related capabilities that are available as a service, over the Internet, to multiple customers."
There's a massive difference between clouds as a business model--outsourced, third-party computing on demand--and clouds as a set of technologies (virtualization, automation, and so on). Vendors who blur the distinction between the two in order to jump on a bandwagon make me mad. We had a vigorous debate on the subject at Interop New York, and I’m sure it will be front and center at Cloud Connect in March.
So I'm going to try and set out why I think the distinction is important, and not just a semantic one. That way I can stop interrupting people and going off on a rant every time the subject comes up.
For a long time, IT was inefficient. Companies scaled their computing vertically--that is, they bought bigger machines, rather than changing the way applications worked so they could handle more users with more, smaller machines. IT could be lazy--there were simply no alternatives. The IT organization had a monopoly on computing. We've known how to make IT more efficient for a long time. Here's a shortlist of disruptive, quick-ROI technologies:
Virtualization, which not only increases the utilization of machines and reduces the number of physical devices the company runs, but also makes speeds up IT operations.
Service-oriented architectures (SOA) that break up an application into distinct, pluggable services like storage, messaging, and authentication, making them easier to diagnose and upgrade independently.
Automation, which reduces human involvement (and human error) while speeding up the time it takes to bring new resources online.
Scaling horizontally, by avoiding the use of joins in databases and other storage approaches.
While these were available for a long time, IT wasn't quick to adopt them. Just ask anyone selling SOA software how hard it was to find buyers. These were good ideas, but IT had no real pressure to embrace them. So in many companies, these things got lip service, but didn't get deployed.
Then along came cloud computing--in the true business sense of the word. It changed expectations in the rest of the company. The CEO called the CIO into his office and said, "Why does it take us nine months to deploy a new version of our CRM when the VP of sales tells me he can have a new CRM from Salesforce.com running in a week?" And it wasn't just SaaS; as testing, QA, and R&D started playing with Amazon's EC2, that changed the expectations.
In many ways, cloud computing is the child in the Emperor's New Clothes, pointing out that the monarch is naked. So now enterprise IT is doing what it should have been for a long time: deploying those disruptive technologies. That's great news; it'll save money, and help the company focus on its core business rather than being delayed by IT obstacles.
But it doesn’t make internal IT a cloud. Because of this, enterprise cloud computing is a temporary, transient thing that's just a stepping stone to clouds-as-a-business-model.
IT Service Management Must EvolveThe idea of technology being delivered as a service appeals to the 409 IT pros responding to our Service-Oriented IT Survey. But cloud providers are competing for that work, and CIOs are being selective.