Some vendors and commentators have coined a new phrase — the “private cloud” which may turn out to be more oxymoron than anything else.
To review, cloud computing refers to what were once called on-demand or SaaS applications combined with applications that live in the cloud because that’s where they can do the most good. Salesforce.com is a SaaS application and its broader footprint of Force.com and hundreds of applications built on it represent a part of the cloud. So do Facebook, Amazon and numerous applications that only exist in the cloud and that provide meaningful interconnection of daily life for millions of people. I call this second category WebNecessary applications because they really can’t do their jobs without being cloud resident. Cloud applications are typically multi-tenant, easily configurable and integrate well with one another.
What’s this private cloud thing?
The private cloud is an interesting twist on the idea of cloud computing and in my mind it is a way to extend the older paradigm of data center computing. Let me explain. Over the last ten years, we have seen a shift in conventional computing from client-server to leaner applications in the data center (DC). There has been dynamic tension between on-demand — up to and including cloud computing — and conventional data center style computing. Over that time, cloud computing has become more robust and conventional computing has adopted many of the improvements brought about by on-demand systems. If you look at a conventional application in a browser today, it’s almost impossible to tell if it’s being served across the Internet or from the data center (DC). But there are differences.
Multi-tenant cloud computing delivers economies of scale to host many applications and many users on a single architecture. It was once called utility computing because the capital-intensive infrastructure was centralized and, much like electricity or water, a service to the customer was the net result. Most people don’t have their own electric generators and multi-tenant architectures mean they don’t have to run a data center either. This brings us to the private cloud.
The private cloud is in some ways a contradiction in terms brought about because some vendors have decided to spruce up their conventional computing architectures with the improvements pioneered in cloud computing but they stop short of supporting full multi-tenant architectures. This is not simply a philosophical difference. Single tenant architected solutions cannot be simply or cheaply converted to full multi-tenant architecture. So the next best thing is to adopt some improvements and a name that blurs the distinction between the two. The private cloud is pure marketing genius.
You might ask, Why this is happening? Why now? The answer is rooted in economics. Simply put, we are in the middle of a paradigm shift and the shift is accelerating. The last ten years were a slow build-up to this point and we can expect things to continue accelerating from here.
The economics of software heavily favor cloud computing. For the customer it is significantly less expensive but for the vendor, it requires a semi-prodigious investment in servers, data centers, smart people and re-architecting products. Companies that have done all this over the last decade are at a decided advantage over companies that resisted change. In this economy, the latter companies face what the politicians in the last administration in Washington once called an “existential challenge”.
If you can’t beat ‘em, change the subject
Recessions are a time when everything is tested or re-evaluated to one degree or another. Today (April 29, 2009), the New York Times reported a government announcement that the US economy shrank at a 6.1% annual rate in Q1 2009. The article went on to say that “…business investment cratered as companies slashed their spending and their inventories as they rushed to cut their costs. Business investment in software and equipment declined by 33.8 percent….” In this environment finding more cost effective ways to do everything, including supporting the IT infrastructure become paramount.
The recession is one of the factors that will put more pressure on organizations of all types to move from conventional computing to cloud computing. In this environment the idea of a private cloud is simply an attempt to extend the old paradigm and deliver at least some of the benefits of cloud computing.
Vendors with large customer bases want to hang on to them while they re-tool their businesses to adjust to the new paradigm. As a result, we get these half-way measures that attempt to prolong the existing paradigm while giving vendors some breathing room. That’s the private cloud. Its purpose is to extend conventional computing long enough for the big guys to catch up. Will it be enough? That’s hard to say. For an analog you might want to look at Detroit.
For thirty years the American automakers knew there was a need and demand for fuel-efficient cars but their business models were oriented on selling large, fuel inefficient models. They made some attempts at paradigm extension, for example, skirting the CAFÉ fuel efficiency mandate by spawning a fad for trucks and SUVs, which were exempted from CAFÉ. It worked for them for a long time. But then, seemingly without warning the bottom fell out of their market. Fuel prices went over $4 per gallon and the recession hit.
A year ago GM and Toyota jousted for the title of world’s largest car company but today two out of the big three US automakers are teetering on bankruptcy.