Cloud Computing: Hey, Hey, You, You, Get Off of My Cloud!

October 15, 2012

The economics of cloud computing are so compelling it is clear, over time, that it will reengineer the entire landscape of enterprise IT. It is that little phrase—over time­­—that poses challenges, for customers and vendors alike. Different applications will drive cloud adoption at different paces and in different directions. To help sort out your priorities, here is a cartoon of the various types of clouds in nature, with a companion cartoon mapping them to the cloud applications getting the most traction these days.

That’s the Earth Sciences view. Here is the Computer Science view:

Let’s take these one at time, beginning with the low hanging fruit, and moving on to increasingly ambitious undertakings.

Virtualization in the cloud is simply an extension of virtualization in the data center—same motives, same economics, with a bit more risk and complexity and returns that warrant taking both on. Not much more to say here, except that if this is what your organization calls cloud computing, it’s probably still in denial about the disruptions coming your way.

Enterprise SaaS, with Salesforce.com being the canonical example, is the first enterprise cloud application to reach Main Street. The business drivers are compelling, beginning with a “consumption economics” business model that is far more attractive than traditional licensed software packages, but extending to a globalization and maintenance model where you are always on the current release. The big mistake early adopters made here was to overlay heavily customized processes and code onto this standard core, thereby undermining most of SaaS’s benefits. Later adopters should avoid making this mistake.

Web Apps are inside the tornado these days within industry segments where digital customer service and support have become table stakes—banking, retail, and travel & entertainment for three. And they are becoming core to consumer product launches and branding campaigns, following the powerful example of movie trailers, working to achieve that elusive “engagement” that marketers so desperately desire.

Email in the cloud is, in my view, either a head fake or a non-event. It certainly makes sense to outsource something as non-differentiating as this, but over-focusing here I think will miss the key point, which is that email is so last century. That is, unless it has been integrated into . . .

Collaboration. This, I believe, will be the next big thing for enterprise IT, what other folks have been calling the “consumerization of enterprise IT,” what I and my colleagues have been calling “systems of engagement.” The core idea here is to leverage all the collaborative technologies that have grown up in the Wild West of consumer computing—Facebook, Twitter, Skype, YouTube, and the like—and make them “enterprise ready,” a journey of civilizing that will take many iterations but will be justified by generating the next big productivity returns from IT. These will come from empowering the middle of the organization (not the bottom, OLTP applications did that, and not the top, Business Intelligence apps did that) to work across organizational, enterprise, and national boundaries to coordinate disaggregated supply chains and sales and service channels.

Test and Development is the only remaining cloud application that is across the chasm. Amazon has been a complete game changer here, and it is hard to imagine anyone else playing more than a second source, neutralizer role going forward. The great news is that, for any entrepreneur inside or outside the enterprise, the barriers to entry for building compute- and data-intensive applications have dropped to virtually zero.

Rounding out our tour of the stratosphere, Big Data Analytics and Real-Time Optimization are both still pre-chasm, Early Market, project-oriented affairs. It is getting a ton of attention precisely because it still confers dramatic competitive advantage to early adopters, be they in financial services, retail, or advertising. The challenge for most enterprises is that the go-to partners for these sort of projects—the top tier systems integrators—have not yet demonstrated they have the chops for this sort of thing, and it is virtually impossible to attract Big Data talent to work in an enterprise IT shop full time.

So the next time someone in your organization says “cloud,” channel your inner earth scientist and make them tell you which kind of cloud, and why.