Why Public Clouds Could Win in the Enterprise


Companies — from local booksellers to offsite hosting providers — are fighting to persuade us that their vision of the cloud is the right one, and the only one.

But to be perfectly honest, no one has all the answers yet. Neither we as commentators, nor they as providers, know how the cloud market is going to play out, and the philosophical differences of opinion are important and worthy of airing as we find a set of solutions that makes lasting sense. Labeling a competitor as peddling a “false cloud,” just because it isn’t like yours? That’s as unhelpful to the caller of names as it is to those who are named, as it provides ammunition for those who wish to dismiss any suggestion that a particular cloud is of value.

Regardless of the philosophical distinctions, it seems apparent to me that the recent growth in various types of private clouds must surely be a short-lived phenomenon. Sooner, rather than later, the majority of CIOs deploying private clouds because of security concerns will see sufficient evidence to accept that the public cloud can be just as secure. The CFOs insisting on private clouds so that the capital expenditure on hardware shown on the company accounts is neatly amortized over three to seven years will have written those former assets off. The public cloud is cheaper. At scale, the public cloud has almost infinite scope to cope with the long tail of peculiar requirements — and to charge those edge cases accordingly.

At the end of the day, the public cloud must surely be the most viable way forward for almost every aspect of mainstream enterprise (and SME) computing. The question is whether this inevitable transition takes us two years or 10.

For more of my thoughts on the eventual ascendance of the public cloud, see my latest weekly update on GigaOM Pro (subscription required).

Image courtesy of flickr user Bob B. Brown

Related Content From GigaOM Pro (subscription required)


Sacha Labourey


Part of the confusion comes from the fact that no naming conventions are shared by everybody yet. To me, the most complete and yet simple one relies on a 2×2 matrix: one one axis you have [public; private], on the other axis you have [internal; external] (would be on-prem vs. off-prem in NIST’s definition). That covers most cases.

What you seem to describe in this article – which I would agree with – is not so much about public/private, but more on the second axis i.e. on-prem vs. off-prem. As the move to off-prem happens, I think most cloud providers will provide both models (public and private) through a continuum of solutions, from a simple virtual private cloud (VPC) à la AWS to fully dedicated private environments (with dedicated network&SAN infrastructure).

Ultimately, the market will decide where the cursor of a “good/safe/secure enough” private cloud with proper cost-leverage will sit on that continuum. But it will take a few years to settle down.

In the meantime, on-ramp solutions that help enterprises move from on-prem+private to off-prem+private/semi-public/public will be the true enablers of that market shift.



Paul Miller


thanks for that. We are certainly facing a lot of issues here in which there is rarely a sensible black or white distinction to be made. There’s a lot of work to do in explaining the choices in terms that make sense to those required to make these decisions… and a need to get past both excessive marketing spin *and* excessive zeal on the part of enthusiasts backing one or other of the available options.

Beth Cossette

Great article. At Qwest, we see a lot of businesses looking at the hybrid approach where an organization provides and manages some resources in-house and others externally. By leveraging a public and private network, there’s less network disruption, more control over data protection, and existing investments in infrastructure can still be utilized. According to a recent online poll of business and IT leaders conducted by IDG Research services, almost all enterprises have adopted or are planning to adopt a hybrid cloud model for at least a portion of their computer applications. You can review the poll here: http://www.qwest.com/qptcms/qCmsRepository/resources/pdfs/cloud-security.pdf?refCode=ENT000000489. Additionally, 45 percent of the respondents in the poll think that a hybrid or mixed approach to cloud security is the way to go. Will be interesting to see how it goes.

Paul Miller

Thanks Beth.

Do you think the hybrid approach to which you refer will be these respondents’ long-term solution, or is it driven more by the need to address current concerns and integrate with current infrastructure?

Beth Cossette

Paul, I would guess medium-long term. It will likely come down to the fact that there is so much embedded code, architecture, and processes that shouldn’t be moved to full cloud. I think Sacha makes a great point regarding ‘clouds’ that can be implemented on-prem or off-prem. The latter can be a mix of multi-tenant shared services and more dedicated services, depending on the customer’s needs for availability, security (per their definitions), etc. So a great amount (maybe a majority) of computing infrastructure will be outsourced in the future, but the methods will still come from a continuum of services that meet customer needs, and it will be a while before all those needs can be met mainly through a multitenant shared environment.

Paul Calento

Public vs. private cloud isn’t an either or debate. I’d argue that most orgs with a private cloud aren’t necessarily doing it becayse the public cloud isn’t viable. Certainly some are.

The private cloud movement (which previously went by other names like “new data center” before becoming a buzzword du jour) starts when organizations first virtualize at scale (servers) and begin upgrading their infrastructure to support that virtualization.

Eventually this spills over to other areas (apps, desktops, etc.), leading to the realization that org’s should be able to access services outside of their company, when needed. But when and if does this happen? I’d argue that the so-called private cloud is the norm for today … with access to the public cloud, as needed, an inevitability.

Paul Miller


broadly, I’d agree. Private cloud is the default answer for most enterprise deployments today. My argument is that this default position is unsustainable, and that *most* (not all) private cloud solutions will migrate to the public cloud over the next few years.


Comments are closed.