Evolution is supposed to be slow-moving and steady, giving those affected plenty of time to react. Revolution, on the other hand, is rapid and disruptive and usually leaves casualties in its wake. Nevertheless, revolution is still a process, which means it unfolds in stages, and it isn’t always easy (in fact, it rarely is) to see how one stage will affect another or how it will all come out in the end.
In that regard, the cloud is a typical revolution. What’s surprising, though, is that each stage so far has been a mini-revolution in itself, and it doesn’t look like we’re even close to finishing.
NASA CTO Chris Kemp spelled all this out plainly enough at the recent OpenStack Enterprise Forum, where he noted that demand for the cloud is quickly propelling the demand for hyperscale infrastructure, both within the traditional enterprise and in the hosted service provider market. The thing is, hyperscale is set to revolutionize some longstanding IT concepts, such as abandonment of top-end systems for legions of low-power ARM and DRAM components. Not only will this produce a different kind of data center, but it will enable new applications and new business opportunities that will seem completely alien from a traditional enterprise perspective.
As I noted in a previous post, hyperscale provides a range of benefits even to enterprises that do not need to push scale into the, well, stratosphere. No matter what your data load is, hyperscale should allow you to handle it at a fraction of today’s infrastructure costs. Improvements are realized not only in standard capex/opex terms like reduced hardware and lower and power/cooling costs, but in harder-to-define ways that come from increased infrastructure flexibility and rapid deployment of dev/ops environments.
And from a management standpoint, of course, the cloud requires a revolutionary shift in perspective from infrastructure to services. As EMC’s Jay Snyder points out, CIOs are already clamoring for solutions that allow their internal IT departments to deliver services at the speed and scale of cloud providers like Amazon. This has already moved beyond mere ERP and CRM solutions to full IT as a Service (ITaaS) capabilities that allow end-to-end compute environments to be brought up wherever and whenever they are needed. To get there, though, IT will have to implement a number of uncomfortable features, such as self-service deployment and select-your-own service catalogs.
This is why many organizations are turning to Platform as a Service (PaaS) infrastructure as reliance on the cloud grows, says Apprenda’s Sinclair Schuller in a recent Wired post. PaaS provides just the right amount of flexibility and centralized control to suit the needs of both knowledge workers and the enterprises they serve. PaaS allows the enterprise to implement core data environments on a software footing, where they can more easily adapt to the growing diversity of data operations, such as mobility, hybrid clouds and software-defined everything.
This revolutionary cloud, then, has no real interest in preserving what was, but it has only provided the barest glimpse into what’s to come. No doubt, many setbacks will crop up in the transition to such a highly dynamic digital ecosystem, but that is true of any revolution: The plan is hazy at best, and the end game is not always clear.