4 ms·

Where it gets really tricky is caching. Say you do your intercontinental join but then want to cache HTML fragments for performance, well, is that classed as "t

• buro9 11y ago
  I think it will be reasonably argued that data in this context is the master record. The source of truth to which queries are sent. A cache is not that, and you need only look at something like the EU e-commerce directive to find exceptions for caches and networks on the basis of being a "mere conduit" for the communication. It is not as if the data is now toxic and cannot be cached or communicated outside of the EU, only that the data must be stored in the EU and should not be replicated to any database or storage outside of the EU that would prevent EU privacy law taking effect. That's important as EU privacy law already has enough exceptions to allow reasonable scenarios like caching to function. And if you are going to say "well I could just query my cache", then I'd suggest that if your cache is able to do much more than a single key|value lookup to retrieve the cached item then it is in fact a database you'd lose the protections of being a cache and you're back in the world of not storing EU data outside of the EU.
  • weddpros 11y ago
    That still means changing the whole architecture of Facebook, Tweeter and the likes... from a replicated system to a cached geo-centralized system. Also, it doesn't bar the NSA from consulting the information from the USA... which is supposed to be the goal of the whole thing. Actually the only way to surely block the NSA is probably to disconnect Europe from the Internet. LOL