Analysts use US storms to question the future of cloud computing

Here in the UK we are enduring "The Year Without a Summer" as the jet stream sits stubbornly hundreds of miles south of where it normally is and pulls endless wet and windy Atlantic depressions in from the west and across the country. We have just had the wettest quarter (April to June) that Britain has experienced since detailed scientific meteorological records began first to be kept in 1860.

The vile weather all but did for the Queen's Diamond Jubilee celebrations and the royal horse racing at Ascot. Now the Wimbledon tennis tournament is underway, an event that almost always loses several days to rain - even in good years. Then in 25 days it's the London Olympics and the long-range weather outlook is grim. Synchronised wading on the athletic tracks looks like being a distinct possibility.

But we are not alone. The US too is suffering a topsy-turvy summer and on Friday night a series of massive thunderstorms, set off by extreme heat, ripped across six eastern states leaving devastation in their wake.

The storm damage was massive and some of the victims of the weather were customers of  Amazon Web Services, the trading name of the company's cloud computing system.

Hundreds upon hundreds of American enterprises such Fox Entertainment, Intercontinental Hotel, Instagram, Netflix, Spotify and Unilever, together with 187 US government agencies rely on Amazon to store their precious data and provide uninterrupted access to it and, although the storm did disrupt some services the hysterical response by the Doubting Thomases and cloud naysayers employed by analyst companies were soon out in force, ranting and raving about how the cloud exposes commerce and consumers to unacceptable risks and predicting that reliance on ephemeral, virtual cloud-based services will bring about the end of the world as we know it.

Except that it didn't.

None of Amazon Web Services's big corporate of government clients experienced any major interruption to their mission critical services and one might have thought that such a test, and the swift reaction to it, under extreme real-world conditions might provoke a round of applause for Amazon's resilience and redundancy, but no, the loudest voices were those of the doom-mongers embedded in the analyst community who, as we now know all too well, pursue their own agendas rather than acting in, to and for the wider good.

Thus the US media was bombarded with unsolicited analyst opinion to the effect that it is time for the US to take a step back from cloud computing, re-evaluate its efficacy and robustness and lessen its dependence on the technology.

Yes, once again the sky is falling even as Google pursues a huge PR and marketing push to provide cloud services at half the cost of the self-same services now being offered by Amazon. The Cookie Monster, rattled, partizan analysts, strident warnings, scare-mongering all involved simultaneously. Quelle coincidence!

What really happened is this: late on Friday evening an electrical storm caused a power failure at the Amazon Web Services centre in Northern Virginia, a facility housing a big server farm. It went off-line and, seemingly, the data centre's backup system was so overloaded that it could maintain only critical functions. However, by noon on Saturday, Amazon was able to announce that full service "has been restored to most of our impacted customers."

In point of fact Amazon's cloud services stood up to the difficulties much, much better than did cabled utilities and services in areas where more than four million homes and businesses are expected to be without power for most of this week and perhaps into next.

However, to listen to and read analyst comment you'd think Amazon Web Services had experienced total meltdown and that the US government and countless businesses are now on their knees as  a direct result. They are not.

What the storms revealed is, A) that Mother Nature maintains the power to make man look puny and helpless and every now and then proves herself to be the real boss and B) that individual companies have been saving costs on their cloud services by deliberately deciding nor to spend on the creation of geographically redundant links. In other words, its not the cloud companies such as Amazon that are at fault but their cheapskate users.

Many Amazon Web Services customers have decided that the cost involved to guarantee 100 per cent uptime is too much. They have taken a gamble and have lost and are now complaining that the service provider is at fault - which it isn't. You get what you pay for.