Can we trust big cloud?

Cloud-based services are all the rage. It seems like everyone and their sister is getting in on the action. The big plays include hosted SaaS (software as a service) applications such as Salesforce.com, Google Apps, and the forthcoming Microsoft Office 365, along with hyper-elastic application hosting environments such as Amazon EC2.

These services and a multitude of others like them have one common characteristic: They allow you to take an isolated chunk of your application infrastructure and move it off-premise, often at lower price point and with (hopefully) better reliability.

The next wave of cloud-based activity will center on ITaaS (IT as a service). Instead of seeking to host a single business function in the cloud, these offerings promise to absorb your entire server and storage infrastructure, while providing cost and reliability benefits similar to those of SaaS — plus complete operating system and application support.

A “total cloud” solution might sound like science fiction to you, but I recently encountered two businesses that have gone the ITaaS route.

When you think about it, ITaaS is a logical extension of the SaaS model. Economies of scale dictate that a provider operating a thousand infrastructures can do so at lower cost and with greater reliability than you can. Similarly, the same elasticity and pay-as-you-go benefits that have made services such as Amazon EC2 so successful apply — providing a powerful technology and business argument for total outsourcing.

But SaaS and ITaaS are different in one very crucial way — and it’s not technical. It’s psychological. Trusting a cloud hosting provider to operate your corporate email or a single line-of-business application is one thing. Trusting a third party to operate everything you have is quite another. The risk of your cloud provider going off the rails — either technically or financially — applies in both cases, but a catastrophic failure of your entire infrastructure would have far more impact on your ability to do business.

If you accept for a moment that ITaaS offerings can provide the same or better technical functionality than your existing on-premise IT infrastructure (this is often but not always the case), the real difference is almost entirely human. Whose throat can you choke when something doesn’t work right?

If you have your IT department and infrastructure onsite, you know everyone’s name, cell phone number, and home address. If one of your admins flakes out, you can fire him or her and hire another. If the equivalent happens in a hosted infrastructure environment, you can kick and scream all you want, but you don’t have any real recourse other than to migrate your entire environment to a different provider. Depending upon the severity of the problem, you might not even be able to do that.

A strong reputation for reliability, financial stability, and good customer service will go a long way toward alleviating potential customers’ concerns about these issues. But if we’ve learned anything over the past few years, it’s that no company, no matter how well-regarded, is too big to fail — and I sincerely doubt that the federal government is going to bail out your hosting provider.

Even if you trust a potential hosting provider’s reputation and SLA enough to look past the catastrophic failure scenario, the much more mundane results of a faceless corporation hosting your infrastructure still exist. Many of the cloud infrastructure offerings available today come with full operating system and sometimes even application support — essentially seeking to replace the day-to-day infrastructure management tasks you perform in-house now.

There, the devil is most assuredly in the details. You know that your line-of-business application needs to be manually started after the server is rebooted — does your hosting provider? If you tell them, can you guarantee that the new guy who happens to be doing patching on your systems this week will know that? Will they know that the third Friday of the month is when you run your overnight transaction processing batches and that an outage during one would result in disaster for you? Do they know that installing .Net 3.5 on your Web server will break your payroll application? Can you get their engineers on the phone with one of your software vendors to work through an application performance issue?

It’s this sort of direct knowledge of your business that I think will make it difficult for large cloud hosting providers to really gain acceptance in the marketplace. Doing it correctly requires a great deal of communication and follow-through. Even the best large companies are not known for being good at either task.

Given this, I wonder if there is a play for smaller, locally focused cloud hosting providers to make a niche for themselves. A smaller regional hosting company may not be able to achieve the same degree of elasticity or economy of scale that larger hosting providers can, but they can do one thing the big guys can’t: actually know who you are and what makes your business tick. Likewise, you know where their offices are and can call their sales team on the carpet in person if an SLA isn’t met.

This article, “Can we trust big cloud?,” originally appeared at InfoWorld.com. Read more of Matt Prigge’s Information Overload blog and follow the latest developments in storage at InfoWorld.com.