Defining “Cloud Computing”
By Jeff Vance
What exactly does the term “cloud computing” mean? For every simple definition (say, computing delivered as a service), there is a detractor claiming that this misleading or downright wrong. It doesn’t help that pretty much every vendor in the world is slapping the term “cloud” on its products, whether or not these are true cloud products. In times of hype, words get twisted so much that they start losing meaning.
Depending on where they are positioned in the market, some vendors define “cloud computing” narrowly as virtual servers available on demand. Others have a definition as broad as the previous one is narrow, arguing that any resource served up from outside the traditional enterprise perimeter is a form of cloud computing. Neither of these extremes is terribly helpful, with each leaving out or including far too much.
What, then, should our current definition of cloud computing be?
I would argue in favor of a broad definition (although not as broad as the one above) as the starting point. As with many umbrella terms, there will be sub-categories that define specific cloud variations, so the task at hand is to start drawing mental Venn diagrams to figure out what SaaS, utility computing, Web services, Platform as a Service, etc. all have in common.
Here’s what I’ve come up with so far:
1. Each trend was driven by the desire of disparate businesses to get out of the IT business. In other words, do insurance companies, banks and auto parts manufacturers all need to be IT experts too? No. IT is better handled by IT-focused organizations.
2. Efficiencies of scale: Early forms of cloud computing, such as ASPs and MSPs, defined their existence by being able to handle IT more affordably and effectively at scale. It took a while for this vision to pan out, and it took a major technological innovation – virtualization – to deliver the efficiencies that are making the switch to the cloud possible.
3. Virtualization as a foundation: In the early 2000’s, ASPs and MSPs were VC darlings. Yet, most of them went out of business almost as quickly as they raised funding. The problem was that they simply weren’t efficient enough, nor could they scale as much as they initially believed without some technological breakthroughs. The key breakthrough was virtualization, which radically changed the economics of being an ASP, MSP, SaaS provider, cloud platform provider etc.
4. Freedom from underlying hardware: Of course, with virtualization as a foundation, it’s important to remember that applications and services are freed from underlying hardware. This freedom drives many of the basic benefits of cloud computing, from near real-time scalability to flexibility to cost savings.
5. Pricing models based on usage: Cloud pricing models are all over the map – anything from per-transaction pricing to weekly, monthly or yearly subscriptions – but they aren’t shrink-wrap software that organizations own indefinitely, buy on a per-CPU basis and install in each user’s individual machine.
6. Scalability and flexibility: Because virtualization is a foundational technology for the cloud, by definition, organizations have the ability to scale resources and services upwards and downwards at will, moving in near real-time to adjust to actual demand, not theoretical peaks.
7. Geographically independent (in theory): You should be able to access cloud services wherever, whenever and on whatever device you choose. That’s the Platonic ideal of the cloud. The truth is that the public Internet still isn’t up to the task to achieve this ideal, nor are all end devices, but the industry is slowly but steadily working towards this goal.
8. Providers are forced to focus on users: Okay, you might say, “Wait, don’t all technology companies focus on users? They wouldn’t be in business if they didn’t.” Au contraire. It wasn’t so long ago that a savvy sales staff could convince a few big companies to buy crappy software and that all but ensured success for that software for years to come. The purchaser was convinced that the software was valuable, often after a few lazy demos, and then the rest of the company had to suffer with this decision for years. Those days are also slowly coming to an end. If Salesforce.com drops the ball, you can switch to SugarCRM or even Oracle without nearly the pain you would have had to incur five or ten years ago. If users can’t use the software, if it is too cumbersome, too expensive or too buggy, it will die a quick death.
That’s a start, and I’m sure I’ve left out a key concept or two. Go ahead and add your two cents in the comments field. In the next post, we’ll look ahead to some things that aren’t included in our (or most anyone else’s) current definition of the cloud, but which should be in the near future in order for cloud computing to reach its potential.
Guest blogger Jeff Vance is the founder of Sandstorm Media, a copywriting and content marketing firm. He writes about cloud computing for a number of publications, including Forbes.com, Network World, Datamation, Cloudbook and many others.