It’s a familiar story – an industry spends its time in the sun, fortunes are made and lost, competitors battle for ‘real estate’, categories and sub categories are established, IPOs happen, stocks go up –and then it all changes. The product and the means of production become standardized, the sub-categories become purchase patterns, economy of scale rules, consolidation happens, competition is limited, prices go down and are regulated by the government. Boom towns turn to ghost towns, specialty degrees are no longer offered, workers migrate to the big city, superfund sites are established.
Telephony, electricity and other utilities, energy, steel – all these businesses have gone down a variant of this path. And guess what – cloud computing will too. There is too much demand for basic compute and application infrastructure across the industrialized world for it to go any other way. And many industry behemoths, who are fashioning themselves as the future computing utilities are beginning to bet their business models on it.
So, how do we, the IT and tech guys who are inventing this stuff, keep our jobs? After all, when it’s all consolidated in mega-geo-plexes, they won’t need as many of us.
The first answer that comes to mind is – don’t impede progress and be roadkill. As the Derailers (great Austin band) sang, “you can’t stop a train”. You will not win if you design, build or operate bad cloud infrastructures. It is the surest way to accelerate the substitution process and have your work farmed out to another provider. This is particularly true in the ‘operate’ side of the equation, where cloud infrastructures spend most of their time. Unreliable, poorly performing, non-transparent cloud operations will kill things fast. Whenever I read opinions that suggest that clouds, because of their uniformity, standard components and modern code are inherently more reliable than traditional infrastructures, I cringe. The moment it begins to fail through scale, motion or end user demands and expectations, there’s an easy substitute to go buy out there. To avoid an early exit, it’s mandatory to get on the good side of cloud during these still early days. Taking a look at BMC’s Cloud Operations solutions is a good start.
The second answer is – don’t believe that cloud infrastructure is the destination. Those who work inside the cloud utility (picture the power linemen or a lonely Homer Simpson) will soon be the vast minority. As an example, in a previous life, I was involved in a west Texas electricity wind farm. After it was built, patched into the grid and launched by marketing, the entire farm, all 160+ turbines and 200 Megawatts of it, was run by a single guy in a rather modest control building complete with sleeping quarters. Clouds are enablers of applications, and the larger opportunity will be in those apps, social interactions and the business activity they represent. Engineers and architects who understand how the cloud really works will have the edge when it comes to the apps that are designed to exploit it.
Anyone else have any better ideas?