Bad news for Cloud Computing boosters. A massive screw-up on the part of the Microsoft subsidiary that runs the Sidekick service has resulted in all data stored on the service’s servers being lost. Data stored locally on the Sidekick devices is still intact, and T-Mobile will be figuring out a way to facilitate the upload of that local data onto the servers at some point. It’s a hard lesson to learn when we rely on someone else to safeguard our precious data. I’m afraid this will be the nail in the coffin for “Pink,” the Microsoft mobile device hardware project reported to be based on Sidekick tech.
Microsoft (now owner of Danger) did not have _any_backups_?? WTF? HellO??
IMHO that is also the death of all of MS’ cloud aspirations (although this wasn’t really a cloud, it was “just” a badly managed data center.)
That seems strange, anyway Microsoft already manage to sell horrible operating systems for ages now and they still continue. I won’t say they are going to stop their cloud aspiration.
I prefer to keep my data to myself, no cloud for me thank you. This way when I loose it all to a disk crash, I can only blame myself!
Same here. It seems that all the proponents of “cloud computing” seem to gloss over this aspect of the whole thing: your data is not in your hands once in the cloud, period. You’re at the mercy of the servers on which it is stored, and if anything happens and there are idiots in charge… boom!
Fully agreed. I would only allow backups to be stored in the cloud, and only when it is non-confidential data. It is far better to place a NAS / small server in a friend’s house and make backups to that other location.
lesson to be learned: The cloud is only a good idea if your backing it up.
how anyone, especialy a large company with many users!, can have let this happen is beyond me.
Running your own apps is only a good idea if you are keeping one more backup copy than you would if you were running in the cloud. Think about it.
When Francine Fishpaw loses her data because she didn’t back it up, it doesn’t make the news. And yet it happens to people like her every day. All the Earl Peterson’s, Dawn Davenport’s, and Donald and Donna Dasher’s of the world who lose their Microsoft Works data go unnoticed.
Agreed. The fact that this is such a notable exception should tell us something.
There are arguments against putting your data in the cloud. But this is not evidence for any of them.
Seriously , though, regardless of whether the data was stored online or at home, screw ups happen. Better to have backups in multiple places, and if I was running a large network, I’d have 2 locations online. One for backups and a separate one for the snapshots/diffs.
I don’t know if this is feasible for medium/large companies, but it would be well worth my time for a small business. my2cents
The problem is simple really…
People mistake a high-availability solution as a backup solution.
RAID/Cluster/Cloud != backup.
The problem arises when some software bug deletes the data – your so-called backup is immediately deleted too.
Backups should be taken frequently and the most distinguishing feature of a backup: It is offline storage.
A lot of people say that but’s that’s not really the case. The more frequently you perform your backups, the more it approaches RAID, especially if you want to be safe and backup all changes. After all, losing a single days’ work or even a couple hours of work can still be just as devastating.
Suppose your backup strategy is a simple sync. Well, a frequent backup will also be affected by whatever destroyed your data in the original. Suppose you say you won’t delete files. Ok. But your originals can still be destroyed by zeroing out files or by user/app level corruption (intentional or not like a bug). And again, your backup will be affected especially if it’s a frequent backup.
I suppose you can mitigate this by having multiple/rotated backup copies or a versioned backup scheme, or snapshots but that adds a large strain on storage costs and space.
Edited 2009-10-13 21:06 UTC
Microsoft uses Snow Leopard as their server OS, and someone logged into the guest account on their servers.
That’s gotta be the best explanation.
RAID, etc. just help you keep things running. It’s for HA, High Availabilty. Backup is for backup.
For when something goes wrong which isn’t just hardware failing: accidental deletion, software/configuration failure.