With the recent T-Mobile and Microsoft Sidekick data loss incident, it is clear Cloud Computing is often mischaracterized as a safe means of data access, storage and application access. Take this quote from this report on the Sidekick issue.
“The Sidekick service, run by Microsoft unit Danger, is supposed to be more secure in storing data because it is kept in the "cloud," which involves storing information on the Internet and not one physically vulnerable location, making the temporary loss of data striking.”
What strikes me is the misconception in the statement. The data loss is striking because it is believed that cloud computing is secure and safe. Sadly, just because data is available on the internet does not make the data safe or secure. Just because an application is on the internet does not guarantee its availability, longevity or protability.
Take for example a service like Shutterfly, or Costco’s online print services. Karen, my wife, has made some fantastic albums using these services. The costs are reasonable, the quality is great. I have no complaint with the outcome (with one exception with Shutterfly sending invitations which had no printed text, but this ultimately had a good outcome).
Once I upload data, pictures and text, to a cloud application, there is no magical invincibility cloak surrounding my data. The service can be hacked, suffer from espionage, have a server fail, go out of business, and so on. Additionally, while the raw data may be mine, the formatted data that the service uses to provide its output is owned by the service itself. I may have my own personal copies of the picture and text, even the printed album in my hands, but the data to generate the album, calendar or greeting card I purchase is not transferable to another service. I cannot, in most cases, get the output of the first service in a generic, openly transferable bit of data to use in another service. In this sense, I do not own all the data that I may have purchased.
This is a common open source argument with proprietary data formats. If I do not have a copy of Wiget’s WordSmithe 1.0, I may not be able to read a document written with that tool. Cloud computing does not make this any less true and carries the additional risk of possessing the raw data to boot.
I will keep my data on my own storage and backup systems. I will even continue to use services like Shutterfly. I will always do so with a bit of caution in mind, asking myself, “if this service is gone in a year, how will I reproduce my work”?