Due to differing data privacy regulations for different countries in Europe, we are looking into the possibility of separating user data based on specific geography. At the moment, we believe it is ok to have this data still hosted in the same physical / virtual database, but it may have to be separated into separate data model based on the geography. Has anyone done something like this dynamically, so that on the fly you could create a new 'geography' such as 'France', have a new data model created for this geography, and then have data of users from this geography stored/retrieved from that data model instead of the primary one? If anyone has thoughts on this, I would love to hear them :)
Thought: why not conform to the most restrictive set of regulations? The issue isn't separating data, the issue is (re)writing logic to deal with missing data (that is either removed or is unavailable in the first place), implementing different anonymizing methods, different reporting methods so you're being auditable for compliance to multiple sets of regulations.