Four Reasons Private Cloud Adoption Initiatives Fail

While it may make sense to get a private cloud up and running as fast as technologically possible, technology is just a small consideration compared with user needs and business culture demands. When organizations go pell-mell at their cloud deployments, disappointment inevitably waits around the corner.

"There are some large companies that can get a cloud up and running fast. And some have done that, and then are disappointed that the company doesn't adopt it or users don't adopt it or costs actually go up, not down, and provisioning takes longer, not less, time," says Jay Seaton, chief marketing officer at GlassHouse Technologies. "And so a lot of what we're seeing is if a cloud is set up without the right objectives or without the right configurations and all of the pieces that go behind that, then bad things happen and people are disappointed."

He says this is why, to a large degree, private cloud adoption has been slower than we might have expected several years ago. And it's why his company announced this week a new Accelerated Cloud Enablement (ACE) service, meant to thoughtfully circumvent private cloud pitfalls in a speedy but not breakneck pace of 45 days.

"If you just talk to software vendors, for example, they could set up a cloud in a manner of days, but what we're talking about here really is not just deploying hardware and software, but making that cloud effective and ensuring that that cloud achieves what the client organization wants to do," Seaton says, adding that making sure configurations, policy setup, self-service and automated provisioning all work to "get at the nub users want out of a cloud that they don't typically see as being available in a private cloud."

According to Seaton and his colleague, Ken Copas, IT service management and cloud practice lead at GlassHouse, this is because organizations tend to fall into a number of the following private cloud pitfalls:

1. IT views cloud and virtualization as the same.

"I've heard a lot of people say that cloud is just a new buzzword for virtualized environment," Copas says. "While it's true that a virtualized environment is key to building a cloud environment--you can't have a cloud environment without a virtual environment underneath it--they are two very different things."

According to Copas, it's important for IT to take a step back and define the difference between a virtualized compute environment and a cloud environment. Doing so will help IT define the objectives for success in a private cloud deployment.

"Anytime you can spin up a virtual machine and make that available to a business user, that is a virtual environment," he says. "But until a user can go to a portal, select among various services through self-selection and then choose to automatically provision that service to themselves and then begin paying for that service as they use it, then they haven't met our definition of what cloud is."

That orchestration on top of a virtualized environment is the essence of what cloud is, and until IT can deliver that successfully, its private cloud efforts will fail.

Next: Private Clouds and the User Experience2. User expectations are skewed.

IT can often meet user resistance to private cloud initiatives not just on deployment but also at conception due to the hype surrounding the "ease" of public cloud computing. The best way for technologists to get users on board with private and hybrid cloud initiatives is through long-running expectation setting and a little bit of marketing finesse.

"Because of what they've heard about public cloud, business consumers think they can whip out their credit card and have a server and their applications up and running in an hour. But there's really very few instances where that's true," Copas says. "Most companies in Fortune 1,000 are using applications that were not built to be deployed in a cloud model. They were built relying on enterprise-level infrastructure underneath of them, and therefore it is not even an option."

It is IT's job to not only let users know about these limitations but to also sell the advantages of the private cloud solution over a public cloud scenario. As Copas puts it, when business consumers engage a public provider they're getting the same service as everyone else.

"What they're offering is not specialized, whereas if your own IT department is offering it, they know your business, they know what you're trying to accomplish, they're on your side," he says. "If you think about it, if a company is trying to go with a truly sustainable competitive advantage with a cloud platform, they would have to do that themselves because anything you can go buy from someone else, your competitor can, too."

3. IT never takes customers into account.

This education and marketing campaign among users may be stymied, though, by an adversarial relationship between business users and IT.

"[What] we run into a lot is user departments and users that are so cynical about IT that the view is if it is a private cloud and IT is running it, then it will reduce my productivity," says Seaton. "We definitely see a lot of divided camps of people looking for any reason to bypass IT."

Often this cynicism is borne from a "geeks know what's best" mentality. This IT tone deafness can kill a public cloud initiative more thoroughly than any kind of technology failure. Unfortunately, it's a slow death that typically comes after investment and long deployment hours.

"We've seen cloud implementation plans for various companies across all the verticals take anywhere from between six months to 12 months and then fail because they weren't implemented in collaboration with their business consumers," says Copas.

He adds that as with many IT initiatives, private clouds depend on healthy business cultures where IT departments aren't technology-centric cost centers but customer-centric business partners.

"That means changing the attitude of the IT department from 'We know what you need and we'll just give it to you' to 'You tell us what you need and we'll build it and enable it for you,'" he explains.

4. IT creates an all-you-can-eat buffet.

One of the key differentiators between virtualized computing and cloud computing is the capability to granularly meter departmental and individual usage and employ effective chargeback models to charge business units exactly for what they consume. Unfortunately, many business don't build charge back into their models, largely as a function of the aforementioned IT-centric decision-making.

"Sometimes they don't put any thought into a chargeback model, and so they provide the end user the ability to create these services that are defined in an IT verbiage versus business verbiage and if they don't have any type of chargeback involved, what they really have created is an all-you-can-eat free buffet," says Copas. "So the end user just goes out and creates things at a whim; there's no cost to them, and they end up with this huge virtual server sprawl."