Why Your Private Cloud Will Fail


Matt Asay

May 20, 2015

Are you building clouds, or are you building applications that run in the cloud?

This is a crucial question, and many companies are getting it wrong. By frittering away their time building out infrastructure rather than building on others’ infrastructure, the IT department’s quest to control infrastructure is causing companies to get left behind. 

With the primary reasons for moving to the cloud being agility and speed of development, the more enterprises spend on building out private clouds—i.e., ones they host in their own data centers—the less time they spend in the cloud. This contention is borne out by new Gartner data.

Why Go Public

Though most enterprises recognize the value of the cloud, 83% of CIOs are struggling to rein in unauthorized cloud adoption, according to a new Brocade survey. By “unauthorized” they essentially mean public cloud—i.e., computing services offered by the likes of Amazon and Microsoft—since it’s impossible to sneakily spin up a private cloud with all the required infrastructure. 

And yet the very presence of that unauthorized (public) cloud adoption should tell them something. It should tell them that their efforts to make agile infrastructure, whether delivered as traditional data center assets or private cloud assets, are failing.

It’s therefore not surprising that public cloud is growing much faster than private cloud (20x vs. 3x). 

But not all public cloud is created equal. And not all of public cloud providers will thrive. As Gartner analyst Lydia Leong highlights, “few providers have the financial resources to invest in being broadly competitive in the cloud IaaS market.” Such “relentless economies of scale”, as Redmonk analyst Stephen O’Grady cautions, don’t simply favor big providers over small, but also big over big. 

Amazon, for example, builds its own hardware, including networking gear. There is simply no way to compete without following suit, as Google and Microsoft have. 

Customers are buying in. As Gartner finds, 2014 saw the absolute growth of public cloud infrastructure as a service (IaaS) workloads exceeding the growth of on-premises workloads (of any type) for the first time. In fact, Gartner’s 2015 CIO survey shows that 83% of CIOs actively consider cloud IaaS as an infrastructure option, and 10% are already cloud-first with cloud IaaS as their default infrastructure choice.

What Do You Want To Build Today?

And yet some workloads will continue to find their way to the private cloud, for both good and (many) bad reasons. 

Some will succeed. You can absolutely make a private cloud work. 

But much of your effort is spent … building the cloud, rather than focusing on the applications that leverage the cloud. And in the case of OpenStack, you’re going to have to invest a heck of a lot of resources to make it work to get beyond what Gartner calls “the nine circles of [private cloud] fail.”

As Gartner analyst Alan Waite told an audience in Sydney recently, “[OpenStack] is a science project,” one that can be made to work but demands skilled resources to achieve success. 

This may be one reason Gartner is only able to find 740 production OpenStack deployments. If even remotely true, that’s a damning indictment of the private cloud’s poster child.

Some commentators rightly complain that such “comments about OpenStack could be applied to ANY private cloud software (proprietary or open source),” insisting that “There are very few successful platforms mostly because the guys designing them…never think through the actual use cases, so end up building a platform that has no value to end users because they are locked into the usual IT department echo chamber.”

Which seems to be an awkward defense, indeed.

Others, like this one, claim “For the average user, any cloud system, OpenStack included, is something which someone else puts together for you. It’s sort of like the next level of operating system.” You probably don’t want to cobble together your own OS (no, really, you don’t), and so those who embrace private cloud, and more often than not this will involve OpenStack, will do well to go with a leading distribution like Red Hat.

But this still isn’t as easy as AWS. 

Give Them What They Want

AWS has been winning because it was the first to really focus on developers. This gave them a surprising “seven-year head start,” according to AWS cloud chief Andy Jassy.  

Microsoft and Google also have the developer religion, and are making gains accordingly. But most others still seem to think that the primary customer for their clouds is the CIO, and they’re flailing and failing accordingly. 

OpenStack, for its part, counts 30,000 community members. That’s great. But what is really, desperately needed is a product that makes it easy for developers to get up and building apps, not infrastructure.

Lead image by George Thomas

This article was written by Matt Asay from ReadWrite and was legally licensed through the NewsCred publisher network.

There is 1 comment

  • Chris Ciborowski - 05/29/2015 17:17
    Nice article. Certainly agree when talking about IaaS. Unless there is a compelling reason to run onpremise (i.e., control, you have the highly technical team versed in *nix, economies of scale, etc.) public deployments make sense. But when moving up the stack there is more to this than just infrastructure. If the business is considering distributed applications which require new infrastructure services (beyond compute/network/storage) using those offered by a public provider beyond infrastructure can be a risk. Better to consider a design using a modular approach with open source components that can be used across deployment methodologies, thereby minimizing application dependency if a provider change is warranted.

Great ! Thanks for your subscription !

You will soon receive the first Content Loop Newsletter