The Real Cloud Imperative Is Greater Simplicity

I saw the look of consternation on some faces: Doesn't cloud computing involve a lot more complexity? Haven't we already managed to create too much complexity on our own, without your assistance? And you're advising us to adopt more -- just so we can call it cloud computing?

Some attendees to the InformationWeek event, "The Cloud Imperative" March 29 at the Grand Hyatt New York, were well acquainted with the definitions and principles of cloud computing. But there's a lot of loose terms and revisionist product information being thrown around out there in the marketplace and the blogosphere. When the third speaker--Wally Pereira, technical architect of Intel's Data Center Group--came up, attendees revealed, by a show of hands, that only a few had actually started cloud projects. That untouched-by-cloud portion of the IT profession may be larger than many believe, given the trendiness of the term "cloud" combined with the looseness of its definition. Then there's the willingness of some in IT to suddenly discover that a few virtualized servers in the data center can now be considered a cloud. Not so. It takes more than virtualization, but some IT staffers have not had the chance to debate that issue with anyone.

All they know is that they're under constant pressure to do more with less, and the scuttlebutt on the cloud says it will help you do so. It's probably the reason why over 50 IT staffers from metropolitan area companies had braved a windy, rainy morning in the city to learn more about the cloud.

I saw this same desire to learn on the opposite coast. At the Cloud Connect 2011 show in Santa Clara, Calif., several hundred would-be cloud practitioners attended an Amazon Web Services presentation, "Moving to the Cloud, Step by Step." This was a more practical, implementation-by-example gathering under the tutelage of Jinesh Varia, technology evangelist for Amazon's EC2.

He laid out the abstract principles of cloud computing in a talk, then proceeded to lead the assembled believers in a trial implementation. Each was registered with Amazon Web Services and had carte blanche to use up to $25 of cloud services, more than enough to keep them occupied during a two hour dry run. As he led the crowd through various implementation steps, he heard the grumbling in the crowd that things were not working as expected, and one attendee submitted that lack of response to his best efforts was one of the reasons cloud computing remained a difficult sell in his own organization.

Varia tried to help the situation by asking who in the crowd had launched a workload in EC2 before, and about 10% of the hands went up. "Everybody note who's done this before. They will be your proctors," he said. He had volunteered 30-40 people in one dramatic swoop. I was sitting next to one of those proctors.

The demonstration continued to more advanced steps, and the undercurrent of muttering continued as well. I would have liked to ask my proctor a question or two, but he was now among the discontents, his own muttering moving into more explicit expletives on the nature of the cloud with each step in the demonstration. Varia bravely plowed on.

Perhaps it's best if the prophets and technology evangelists don't get too bogged down in the concerns of terra-firma -bound individuals. If they did, they would surely encounter delays in achieving their lofty goals. But for those actually responsible for implementing company workloads, unexpected delays, glitches and setbacks are the routine, not the exception, and they cannot ignore the problems in front of them.

I am struck over and over again how easy it is to discuss cloud computing in high sounding terms, while those plunging into the cloud are thrust into a welter of new technology processes and complex responsibilities. Both the technology and processes associated need to be simplified further, documented and explained further and made more accessible to those who have the skill but have not necessarily mastered the terminology, the disruptive style of thinking about hardware reliability, the brave new world of running applications remotely while not losing the data that make up the heart of cloud computing. The entry of the IEEE into this rapidly emerging field April 4 with a promise to standardize some portability and interoperability functions is welcome news. Cloud suppliers are willing to solve each problem in their own way, but that's part of the problem. They threaten to add a new layer of complexity and potential lock-in for users. A common way of getting systems in different clouds to work together or executing movements between clouds will be much more useful in the long run.

I think the Cloud Security Alliance, the Telecommunications Management Association, and DMTF, the former Distributed Management Task Force, will contribute to solving these problems as well.

But all parties, including the journalists and bloggers, need to do a better job of realizing the complex reality that IT managers face and not rationalize away the difficulties of implementing the cloud. The true cloud imperative is the need for greater simplicity, and we're nowhere near achieving it yet.