Amazon's cloud crash is turning out to be the IT world's equivalent of the rubberneck-attracting, traffic-snarling dramatic highway disaster of the season. There's a promising new technology that has been promising (or threatening, depending on your point of view) to shake up the structure of the whole industry. It has a celebrity victim that built a bulletproof network of data centers to support its own business, then opened them up to the public. It has innocent bystanders flattened by the impact, several of whom suffered serious and long-lasting injuries. It also has an object lesson to anyone who thinks of cloud computing as a quick path to problem-free, low-cost, top-flight computing without the trouble of building or maintaining the technology to make that possible. Here's a tip you can pass on to the business-unit managers who are either buying cloud services themselves, or can't understand why you can't get rid of your expensive data centers and just lease some from Amazon, like the shiny BMWs the sales staff pay rent on for three years, then give back without any equity to show for it: Clouds are data centers. They're big data centers that are at least as complex as your internal data centers and it takes serious technology chops to build, distribute and run n-tier applications across their many facilities to get the kind of performance and reliability the business units think comes for free. Apps running in the cloud are inherently multi-tier – housing data on one set of servers and SANs, applications on another set, databases on a third, networking, load balancing, resource-management, backup, disaster recovery, security and a dozen other functions on a third or a fourth or a fifth. In large-scale clouds, those sets of servers aren't grids or clusters, they're data centers. At Amazon one of those data centers -- the one housing the Relational Database Service that is to Amazon's cloud what an ordinary Oracle RDBMS is to a normal business application – stopped talking to the apps and APIs and message-queuing and other services that all relied on it. That brought down, slowed down or stopped in their tracks sites that built their whole businesses on the assumption Amazon knew what it was doing when it built the cloud and when it said the cloud was a reliable platform on which to build. That was true, actually. Amazon did. It may have made a mistake in this case. It's still trying to isolate and squash the problem that brought down Reddit and Quora and others, so it's not yet clear whether the proximate cause was a mistake Amazon made, a hardware or software failure it didn't anticipate, or some unpredictable glitch that wasn't its fault at all. It could have been the fault of a telecom vendor, an app vendor, a customer who mis-configured a database or application, or something that went wrong in any of the thousands of bits of IT in any data center – each of which has a tiny bit of evil at its core. That evil lets it pretend to be working perfectly until even a slight stutter would cause a big problem, then explode dramatically. The central skill of data-center gurus is not in computing; it's in disaster prevention. Power failure, environmental disasters, hardware failures, software failures, sabotage, telecom problems, security problems, environmental problems, zombie apocalypses, onslaughts of fully authenticated BYOT devices – there are backup systems, redundancies and preventative measures for all of them. Any company running its apps in the cloud has to do the same. Relying on someone else's IT to keep your business up and operating perfectly is like walking across Times Square without looking at traffic and relying on every bus, taxi bike and pedicab to give you the right of way automatically. It would be a good experiment, but don't try it without a good medical plan and an extra couple of video-game lives in the bank you can fall back on. It's not that Amazon's outage isn't a big deal. It is. Amazon will get a black eye and will deserve it. But cloud computing as a category might also get a little more realistic reputation, rather than the idealistic sheen that blinded people to the idea that cloud computing is a technology, not a miracle.