http://arstechnica.com/business/news/2009/10/t-mobile-microsoftdanger-data-loss-is-bad-for-the-cloud.ars
QuoteIf youre a Sidekick user who has been waiting out the past weeks data outage in hopes of a happy ending, youre not going to like the joint press release that Microsoft and T-Mobile put out on the matter this past Saturday. The release contains a line that no service provider ever wants to see in print with their name attached to it: "Regrettably, based on Microsoft/Dangers latest recovery assessment of their systems, we must now inform you that personal information stored on your device—such as contacts, calendar entries, to-do lists or photos—that is no longer on your Sidekick almost certainly has been lost as a result of a server failure at Microsoft/Danger." Ouch.
Any of your data thats on Microsofts servers is just gone, and the only possible backup of it is whatevers cached in your devices local memory—so do not allow your device to lose power, T-Mobile and Microsoft/Danger warn in the press releases headline, or youll never see your data again.
As a make-good for the inconvenience of denying you service for seven days and then losing your data due to a failure to back it up properly, T-Mobile is offering customers one month of free data usage. If the irate response on various bulletin boards is any indication, T-Mobile couldve just given customers the finger and it wouldve had the same effect. Of course, its not clear what could possibly be done as a make-good for a blunder of this magnitude, since what customers really want is not free data but their data, restored back to their accounts.
This outage and subsequent data loss is just the latest, and probably the most dramatic, black eye yet for "cloud computing," since users falsely assumed that if data was "in the cloud" then it was stored with enough redundancy and fault tolerance to render such a massive data loss impossible.
Of course, most of the affected Sidekick users will probably fault the responsible service providers, and not "the cloud," but Sidekick users arent the demographic with whom the clouds reputation needs the most help. What probably has execs at HP, Sun, Intel, IBM, Rackspace, Amazon, Google, and the rest of the growing list of cloud infrastructure and service providers slapping their foreheads in frustration this weekend is the fact that cloud services already labor under a stigma of unreliability with potential enterprise customers.
In a number of cloud computing threads in the Ars Server Room, and in the handful of webinars that weve done on the topic in recent months, reliability is the number one knock that IT pros have against anything "cloud"-related. Part of the issue is that the IT function has a natural distrust of any systems that it doesnt own and control, especially when those systems present themselves as a black box on which IT will build something for internal use. But, anecdotally, these periodic, high-profile service disruptions seem to loom fairly large in enterprise perceptions of the clouds reliability, which is why this latest data loss is likely to translate into a revenue loss for more than just T-Mobile and Microsoft/Danger.
Its also likely that having Microsofts name attached to this blow-up exacerbates the damage. Theres a fairly standard tradeoff between agility and reliability thats presumed when evaluating cloud services. If your developers opt to live on the bleeding edge by using new cloud services from unproven providers, theyll have to be able to tolerate some downtime; it just goes with the territory. But Microsofts is not a new name in computing, and basic contact, calendar, and messaging arent cutting-edge cloud services. Data loss of this scope just wasnt supposed to be possible. So its easy to imagine that other blue-chip names will have to work extra hard in the coming months when pitching their cloud services to skeptical IT decision makers.
I know it wont effect many/any people here, but Its a serious F*ck You to anyone who was trusting them to keep the data safe.
the press releaser has been updated, theyre going to give out a 100 dollar customer appreciation card or sommat like that. spend 100 doillars with us basically.
question that remains to be answered is how in gods name did they fail to back it up?
Matt
its possible they were backing it up but the issue affected the backup, it does happen, personally i cant believe people dont backup their own stuff rather than relying on some company to do it if its that important.
the thing with this is that its pretty much the worst thing that had happened with Cloud Computing, and at least if people ask about the reliability you can now point at that and say its not 100%. Most of the IT Admins I know dont like the loss of control, but in some cases they are being strong-armed by the beancounters to use if for cheapness.
Back up wise, is it even possible for the users to make their own backup files? Beyond say, forwarding your emails to a private account?
Id had my hotmail account since my family got net access, and Ive never once considered the fact that the only copy of those emails is stored elsewhere and could go at any time. Given the fact they havent gone, I would have thought it was a logical assumption that other services stored elsewhere would have the same level of security, and I guess many users would feel the same. Certainly now itll make me stop and think twice before relying on any cloud computing stuff in the future ><
my problem with the cloud idea is simply this
if you want to be sure its backed up right, you have to do it yourself
therefore you have to keep your own backup on your own hardware
therefore you have to have hardware capable of doing the above
therefore why bother with the cloud in the first place?
I guess its a multiple redundancy thing
like wearing 2 condoms instead of 1
A proper backup routine should always include an off-site secure backup. Its like relying on RAID as a backup when technically its not, its just redundancy.
All about what risk youre trying to mitigate tbh.
Its just reminded me that I have yet to backup the last 6 months of work ><
http://www.engadget.com/2009/10/15/microsoft-recovers-most-if-not-all-sidekick-customer-customer/
Im guessing they just got a copy of easy recovery pro or something then..
They listened to Tekforums and had an Acronis TrueImage on standby.