Through a probably trivial date error, Microsoft’s Azure has screwed things up for all the other cloud providers, says Peter Judge
It had a certain inevitability about it. Microsoft promoted its Azure cloud service using the slogan “I laugh in the face of unpredictability”. What could go wrong?
Even if you didn’t know, you could guess what would happen next. Yesterday, Microsoft’s cloud reputation evaporated and every other cloud provider will suffer the consequences.
A leap in the dark
Microsoft’s developers, it seemed, didn’t think of that. The service has not been around for four years yet, and on its first leap year day, it collapsed. The service management part of the service failed.
You could tell this was serious. Initial problems propagated to different territories, and live customer-facing sites became unavailable, although frustratingly, their owners could see the virtual servers running perfectly well, but inaccessibly.
The UK’s Government’s G-cloud service went down. This was not a disaster, as the service is still under development, but it dealt a heavy blow to the idea that cloud services – or at least Microsoft Azure – could be a reliable replacement for other IT functions.
At TechWeekEurope, we started to hear from providers with a stake in other models. “While we certainly don’t relish these moments, the downtime can be significantly mitigated if organisations were to adopt a hybrid cloud strategy,” said Vineet Jain, chief executive of Egnyte, a company which – wouldn’t you know it – provides hybrid cloud storage.
We would tend to agree that any new service should be treated as potentially unreliable, even – no, especially – if it claims to be supremely reliable. If a service says it can thumb its nose at unpredictable events, elementary superstition, as well as an understanding of human nature, that service should be put under a very critical spotlight.
Microsoft goes quiet
Microsoft clammed up tight when its service celebrated leap year by leaping off a cliff. No one could get any comments out of the company as it struggled to understand, contain and repair the damage. And today’s blog post, the only official communication so far about the event, puts a very business-like gloss on something which on the inside must be simply gut-wrenching.
Azure had problems while it was in the test stage and in 2009, users of Microsoft’s Sidekick product lost data when systems failed. This failure has happened in Microsoft’s flagship cloud computing service. If it is still subject to this sort of trouble, who will trust it?
The date bug is embarassing enough, but the confusion and the proliferation of problems after it struck makes Microsoft look even worse as the problems propagated around the network. While Microsoft got US services back up, those in Europe and elsewhere had to be “squashed”, we have been told.
The increasing importance of the cloud lends ever-greater significance to cloud problems. When BlackBerry maker RIM suffered an extended outage in 2011 it was already in trouble, but the blackout started an escalation of shareholder disatisfaction, and contributed greatly to the once-mighty phone maker’s ongoing trouble.
Microsoft will have to start its cloud marketing from scratch, to rebuild a level of trust that has now crumbled, but so will every other cloud provider. Launched in 2006, Amazon’s cloud services have already passed unscathed through a leap year, and has a customer base that dwarfs that on Microsoft Azure.
But still, every potential customer for Amazon Web Services and plenty of the existing ones will be demanding reassurances that something similar can’t happen on the Amazon service. And every other web provider will have the same experience.
The results of Microsoft’s outage are all too predictable. And not remotely laughable.
How well do you know the cloud? Take our quiz