Could we store our personal application data in the cloud?
August 18, 2009
Cloud based computing and data storage is something that I find fascinating. The whole concept of moving away from rigid hosting centres to a distributed system, where data just flows around the web and gets requested and added to by apps and services also in the cloud opens up so many possibilities. Where cloud based computing is going is anyones guess but this great article by Carl Hewitt on O’Reilly Radar gives makes a good assumption at where the storage of personal data could be going and identifies a lot of the issues that will arise around privacy and regulation (good read).
The interesting stuff on privacy etc aside, it was Carl’s description of personal ‘client’ clouds in particular that drew me to write about it. This is something I’ve been thinking of for a while; particularly with reference to the data that services like Facebook, Twitter, Delicio.us and Flickr store on my behalf. At the moment that data is hosted by each service provider in their choice of data centre and data from other services can only interact with it with the help of OpenAuth, Facebook Connect and other ID services I can subscribe to. What I’d really like to do is hold my data myself, perhaps in a client cloud, and make it accessible to the services I choose.
This led me to think; could services like Facebook and Twitter work on a cloud data model whereby they don’t host anything for you apart from basic identity mapping to enable you to log in? Once logged in your data would be accessed by the application (say Twitter for example) from your client cloud storage solution using an identity key so the experience wouldn’t change. As broadband speeds increase, storage becomes cheaper still and processing speeds go up this should be possible if the privacy issues could be overcome. Imagine being able to take real control over what data you leave behind and actually have the option to disclose data to applications to make them usable but then update that and keep it with you in your storage repository. Every user would have an API key for each web app allowing the services to continue to work when you’re not ‘online’ (although in this scenario you’re data is always online) so functionality wouldn’t be affected.
Of course Facebook and co. want your data to enable them to target their ads better. If we owned our data then we’d be able to allow them access to whatever pieces we wanted to release for advertising purposes. Happy with them being able to target you based on factors such as what sex you are and location but don’t want them to target you based on the last product you recommended? If you owned your data then that would be possible.
The ability to have all your data available to then share with other apps and mashups make this a very attractive proposition to me. I think we’ll begin to see web apps which leave the data on your hard drive soon (in fact there probably are some already), so how much longer till we take that a step further and put all our data in the cloud?