When the cloud goes bad

Bob Cringely wrote an interesting article about what to do if a web-based service (you may also hear of it described as cloud computing or SaaS) stops. What happens to your data? The article argues that once you licence most software, so long as you have hardware that can run it you are fine and that some companies still have mission-critical software running on Windows 98.

It then prompts questions about who owns your data according to the arguments that Cringely and others bring up. However I think that this is actually the same problem that computer systems have presented users for years. Obsolete file formats and data formats like super-sized floppy disks leave information isolated from its owner as surely as disappeared web service. This older demise of data put at risk historical data from the moon landings and artistic recordings to historic government communications.

The only way to counteract this is to:

  • Ensure that web services ensure an export service like Microsoft did with their dead video service Soapbox and Butterfield & Co. at flickr have done. Part of the reason for doing this is that many people don’t want to lose customers, let alone hold the door open to let them go if they want. However, it can also be a great incentive to keep you at the top of your game and that your web service keeps evolving and improving. Flickr even goes as far as to develop on the same APIs that they give to external developers
  • Secondly once the data is out, it needs to be interpreted or parsed back into information and this can only be done with documentation of the format. This lack of documentation is what kept Microsoft at the top of its game for so long and encouraged the European Union to go after them like a pitbull after a pork chop. If you have this documentation, then a reader for the files can be developed to parse the data correctly