I've been talking about cloud computing a lot recently, both at an analyst conference and at seminars organised by a cloud provider / systems integrator.
Gathering feedback from attendees at the seminars was enlightening. My job at the events was to educate the audience, consisting of CIOs and IT managers, about the various definitions of cloud computing, the various categories such as public, private and hybrid, and types of services, such as infrastructure, platform and software as a service, along with the the benefits and challenges.
Like others in or around the IT industry, I've been talking about cloud computing for years -- pretty much since the industry came up with the concept. And while it's been a few years now, my experiences suggest that there's still a lot of people out there who feel under-educated about cloud, what it is, what it's for and why they might -- or might not -- want it.
Back to basics
So let's go back to basics. Cloud computing describes the flexible delivery of online services. The essential difference between cloud and the now outdated (of course -- this is the IT industry after all!) application service and managed service provision (ASP and MSP) models, is the word 'flexible'.
Enabled by virtualisation, a datacentre can deliver cost savings through the economies of scale when servicing a number of users or customers. That works because the peaks and troughs in demand for resources, such as computing, networking and storage, are spread across all users.
The end result from the user's point of view is that resources can be turned on and off or scaled up and down according to need without having to buy the equipment to provide that service. As soon as you buy hardware, you are locked into the maximum performance that hardware can deliver, whether or not you need it now.
A classic example is storage, where buying to meet the demands of the foreseeable future, such as three years, means you could end up buying three or four times as much as you need now. It's the difference between leasing and buying. Not only are those mostly empty, spinning discs consuming energy, they are costing you either interest on the money you borrowed to buy them, or foregone interest or other return on that capital.
If you buy cloud-based storage (or any other cloud-based resource), the theory states that you buy only what you need when you need it, and this saves you money in the long run.
Now all this is the cloud industry's yada-yada. There are of course problems with cloud, especially around security, availability and performance of networks, and the issue of needing to know where your data is, something a lot of cloud providers either won't tell you or won't give you any control over.
It's also a bit of a jungle out there, with incompatibilities, data lock-ins and incomprehensible SLAs just a few of the gotchas waiting to strike the unwary.
Further, research predicts that spending on cloud computing services by 2020 is unlikely to exceed 10 percent of all IT spending, so the whole cloud phenomenon does need to be put into perspective; in nine years time, you'll still be spending some 10 times more on the kinds of IT stuff you buy now than you will on cloud services.
That's cloud in a nutshell -- it's early days for the phenomenon so there's still plenty of room for change and for definitions to shift. Above all, don't think about cloud as the be-all and end-all, think about it as one more tool in the box. If the tool fits the job, by all means use it but don't get carried away by the latest shiny-shiny.
If you have any thoughts along these lines, I'd be glad to hear them.