What makes a cloud a cloud? The ancient Greek philosopher Plato might tell you“cloudness”, but what exactly does that mean?

  Long before human scientists figured out that DNA was the basic building block of everything living, philosophers spent long eons being satisfied with Plato’s (and his equally famous student Aristotle’s) explanation that there is some inherent “ness” in everything that makes it what it is. One of Aristotle’s dialogues deals with the answers to questions like, “What makes a cat a cat? And why does a kitten never have a duck?” as he explains the concept. Retroactively applying our knowledge of DNA to Plato’s principle there’s a certain simple but elegant logic in his answer. DNA carries “catness” or “humanness” as it passes from generation to generation and, thus, his principle is actually a fairly sound one. It’s just not very satisfying if you actually desire any kind of detail.

The answer to “what makes a cloud a cloud” can also be viewed as an Platonic one: cloudness. For many who view the cloud from a black-box perspective, “cloudness” is a good enough answer because all that really matters is that cloud computing fulfills its promises: delivery of applications in an efficient, on-demand fashion for less cost than would be possible in a traditional corporate data center. How that happens is not nearly as important as the end result. But for many it is important because they, too, want to partake of the goodness that is cloud computing but want to do so on their own terms, in their own data centers.

"Our belief is the future of internal IT is very much a private cloud," says Gartner analyst Thomas Bittman."Our clients want to know 'what is Google's secret? What is Microsoft's secret?' There is huge interest in being able to get learnings from the cloud."  -- Jon Brodkin, Private cloud networks are the future of corporate IT, NetworkWorld

That isn’t a recent question. The article and questions were raised in November 2008. It’s now April of 2010 and the notion of private cloud computing, if not still contentious, has still not really been answered. That’s likely because “cloud computing” isn’t a “thing”, it’s an architecture; it’s an operational model, a deployment model, even a financial model, but it’s not a tangible “thing” with a specific “secret ingredient” that makes it work. Cloudness is, in fact, very much like DNA: it’s the way in which the individual strands of genetic material (infrastructure) are intertwined (processes) and the result of that combination that make something a cat – or a cloud - not the existence of the individual components.

So like Google and Microsoft and Salesforce and a host of other cloud computing providers across the “aaS” spectrum, they all have the same ingredients – they’ve just architected them in different ways to make what we call “cloud computing.” Their secret is ultimately in the operational integration of servers, storage, network, and compute resources smothered in a secret sauce called “orchestration” that gives it cloudness: a dynamic infrastructure.

That’s really the core of what underlies the mechanism we call cloud computing: a dynamic infrastructure providing the means by which components are integrated and automated, processes orchestrated, and capacity/provisioning enabled on-demand. Resources, whether compute, storage, or network are consolidated and shared, which produces efficiency of resources and their underlying physical platforms. Tasks are automated, which results in operational efficiencies. Processes are orchestrated, which reduces the overall costs associated with dynamically managing the multitude of applications, services, and infrastructure required to support a large number of constituents. It becomes easier and hence faster to deploy new applications because the underlying processes required to provision the resources across the infrastructure becomes standardized. 

All of these require integration across infrastructure components and with management systems that monitor and act upon data shared through that integration. What matters is the architecture, because everything is depending on how all the moving parts interact and collaborate. When an application is launched, it kick-starts a process that has to deploy, enable, or inject policies across the entire infrastructure to support that application. From assigning an IP address to adding the server to a load balancing solution to allowing traffic to flow through the firewall to reach the application (so users can use it, go figure) there are myriad cogs in the great wheel known as cloud that make it possible for the infrastructure to work in concert just to get an application up and running. The architecture, in part, determines whether this process is efficient and whether or not it’s fast enough. The architecture in part determines how well the application will perform, whether it will failover correctly, how fault-tolerant it may be, and how well it will scale.

The architecture and the way in which its components are orchestrated to automate data center processes is the “secret sauce” of a cloud computing environment, and together that’s what makes a dynamic infrastructure. This is not a trivial task. The more services required – security, acceleration, optimization, load balancing, encryption, persistence, etc… – the more difficult it becomes to integrate, automate, and orchestrate the data center.

But obviously it can be done as it is being done right now by both providers and enterprise organizations alike. No two architectures will be exactly alike because every organization has differential operational parameters under which they operate. Some will be designed for peak performance, others for fault-tolerance, some for secure operations, and others purely for scale and availability. What they all have in common, however, is “cloudness” : a dynamic infrastructure.