No one seems to have been hurt by Microsoft’s Windows Azure cock-up. Peter Judge thinks maybe that’s because no one is using it
Microsoft‘s many regrets about Windows Azure must surely include the name. A word used to describe the blue of s summer sky might have seemed a good one for a cloud service, but as the service failed again last week, it was also a gift for headline writers.
Today, the Internet is full of mentions of the Blue Sky of Death, the Azure Screen of Death and other references to the Blue Screen of Death which plagued Windows users every time the operating system failed. The BSOD is now thankfully a rare occurence, but it always carried with it a suspicion of incompetent programming. And that taint clearly afflicts windows Azure: this clumsy failure follows an equally inept foul-up last year, when the service failed to take the Leap Year into account.
It doesn’t help that Microsoft has been basing campaigns on the reliability of Windows Azure. It also doesn’t help that when spokesperson Adrienne Hall peeped above the parapet – just long enough to fob off customers with a couple of sentences of flannel – her job title turned out to be the general manager of Microsoft Trustworthy Computing.
It is early days, of course. Microsoft’s response to the Leap Year cock-up, when it eventually appeared, was exemplary for its thoroughness. This time round, it will have to be equally honest, as well as including some actual apologies to customers hit by the failure, as well as an explanation of what it will do differently, and why we should believe its promises of reliability in the future.
If nothing else, Microsoft will have to explain the fact, pointed out on The Register’s Sysadmin blog, that Microsoft’s Windows Server 2012 includes robust and powerful tools to automate the very task which Microsoft failed at. The operating system includes a feature called Centralised SSL Certificate Management (CSC), which can direct a “symphony of re-validation” on the most complex network of servers.
It is possible that Microsoft was not using its own tools here.
However, there is one striking fact about this failure – so far I have not seen any major customer complaining about it. Last year, some eager beavers had put part of the UK’s government G-cloud on Azure and complained as it was rendered unusable.
This time round, we have not heard any major users actually inconvenienced by live apps on Azure. This is despite Microsoft’s claims that many large organisations are running business-critical systems on Azure.
We rang EasyJet, a flagship Azure customer, and its spokespeople were unaware of any outages caused by the Azure failure. They have promised to get back to us if they can find any more details.
There are two obvious conclusions to draw from all this.
1. No one much is actually using Azure for real business critical workloads.
2. If you are, given Microsoft’s apparently accident-prone nature, you probably shouldn’t be.
How trustworthy is your Microsoft knowledge? Try our quiz!