My colleague Phil Wainewright loves the cloud, but not the private kind. He recently surfaced a Microsoft report that makes the case for public over private clouds in a strong way. "The economic case to discredit private cloud has now been made,"he observes. "The other shoe still waiting to drop is a complete rebuttal of all the arguments over security, reliability and control that are made to justify private cloud initiatives. The dreadful fragility and brittleness of the private cloud model has yet to be fully exposed."
In the Microsoft paper, titled "The Economics of the Cloud," Rolf Harms and Michael Yamartino look at the economic benefits offered by the public cloud. "We see a long-term shift to cloud driven by three important economies of scale: (1) larger datacenters can deploy computational resources at significantly lower cost than smaller ones; (2) demand pooling improves the utilization of these resources, especially in public clouds; and (3) multi-tenancy lowers application maintenance labor costs for large public clouds."
As Harms and Yamartino explain, the economic benefits of public clouds are getting so compelling -- particularly in their ability to deal with variability -- that organizations will soon have difficulty justifying why they maintain their own IT infrastructures.
And, they add, in light of the advantages of using public cloud services, private clouds will also be difficult to justify. Public cloud offers advantages over private clouds in terms of 80% lower total cost of ownership for data center assets, a 40-fold cost reduction for SMBs, and a 10-fold cost reduction for larger enterprises, Harms and Yamartino point out. A private cloud data center of approximately 1,000 servers will cost 10 times more than the same capability delivered via public cloud. "As more and more work is done on public clouds, the economies of scale... will kick in, and the cost premium on private clouds will increase over time.... Supply-side economies of scale which allow large clouds to purchase and operate infrastructure cheaper; demand-side economies of scale which allow large clouds to run that infrastructure more efficiently by pooling users; and multi-tenancy which allows users to share an application, splitting the cost of managing that application." Issues such as security and performance are also effectively being addressed, they add.
What are the disadvantages of private clouds? Harms and Yamartino observe that private clouds allow for pooling of resources, and therefore reduce issues with load variability, and also ensure greater security. But the movement of virtual loads between machines may get hung up by incompatibility and complexity, they state. "Public clouds have all the same architectural elements as private clouds, but bring massively higher scale to bear on all sources of variability. Public clouds are also the only way to diversify away industry-specific variability, the full geographic element of time-of-day variability, and bring multi-tenancy benefits into effect."
A couple of thoughts on this:
Microsoft has stakes in both the private and public cloud worlds, so it can be assumed this report was developed objectively. As Phil points out, Microsoft shoots itself in the foot in regards to some product lines -- for example, Microsoft Dynamics is very much part of the on-premises computing world.
I would like to see more examination of the economics of private cloud in the context of the blurring lines between software providers and software consumers. That is, what is a software "vendor" these days? Many non-IT companies are assuming the role of cloud provider, whether they call it "cloud" or something else. Amazon certainly broke this mold a few years back. The US General Services Administration runs an app store, called Apps.Gov, for all its agencies. Is the US government a software vendor? Think about companies that are gaining business advantage by extending private cloud services to partners and customers, particularly industry-specific services that a generic cloud provider may not offer.
The service-providing organization may even be tapping into a public cloud infrastructure on the back end to power such offerings. Who knows? It's all possible. The point is, companies are all becoming both producers and consumers -- and brokers -- of IT services. And thus, we see a hybrid of private and public cloud resources becoming available to many organizations provides opportunities for businesses and their IT departments -- enabling them to craft entire new lines of technology-focused business services that they can provide to new and existing partners or customers. The provisioning of external services not only means extraordinary customer service, but potential new business models and revenues in the long run – not part of current cloudonomics computations.
In the Microsoft report, Harms and Yamartino cite the development of the first automobiles, and how they were essentially "horseless carriages," even with buggy-whip holders still attached. That's because innovators viewed the new invention through the existing lens of their time, when the horse dominated the streets. Likewise, the tendency is to view cloud computing through the traditional divided -- and often adversarial -- vendor/customer lens that's been in place for decades. Will a hybrid "vendorprise" (based on a term coined by Thomas Otter) be the wave of the future?