In the last few days, a cautionary tale has been unfolding in the world of cloud computing. There’s a Smartphone offered by T-Mobile, called ‘Sidekick’. It does the usual stuff – Internet access, email, diary, camera, contacts database, etc.
One of the key features it espoused was its close integration with ‘the cloud’ via a central datacentre for many of its functions. It was ahead of the industry in many respects, and gained a loyal following.
The product itself was created by a company called ‘Danger Inc’, which was bought by Microsoft last year.
On 1st October, ALL Sidekick users, worldwide, started to experience problems – and some completely lost their data-connections – which meant that a lot of users suddenly couldn’t do basic things, like accessing their contact databases.
After several days of erratic behaviour, T-Mobile have announced that, following some problems at their datacentre, many users may have permanently lost all of the data that was being stored, and advises current users not to reset their devices at any cost, because of a risk that they will also be wiped.
While the full details are yet to come out, it apppears that, during an upgrade of the servers at the Danger datacentre, something went wrong, and it then transpired that NO BACKUPS had been taken before the upgrade got underway.
To coin a phrase…..”Doh!”
Let’s be clear – this isn’t some startup company running on a wing & a prayer – it’s a subsidiary of Microsoft. Early reports are that the problem is being blamed on a sub-contractor who was hired to perform the upgrade, so it looks like one of those classc cases where everybody thought someone else had a certain responsibilty, only it turned out nobody did.
What to take from this?
Microsoft are not novices at the Datacentre game – they are putting HUGE resource into it, and are recognised as one of the world leaders in terms of technical design and development of best practices in the field. Yet despite this – something fell through the cracks, Sod’s Law kicked in, and hundreds of thousands of users were left high and dry, with very little chance of much more compensation than a refund of this months’ subscription fees.
If cloud computing is to succeed, then it is critical that suppliers can demonstrate they can be trusted with our data, and that they have processes and technologies in place to deliver on that. They may also be forced to offer Service Level Agreements that provide better compensation levels than are currently offered before they will ever stand a chance of gaining contracts with larger organisations. If Facebook goes down and loses your Mafia Wars score (or whatever), it would be annoying but not fatal. If your ERP system suddenly vanishes….