|
|
|
|
|
by rubiquity
1976 days ago
|
|
You just described distributed databases -- which are overwhelmingly now deciding to use SQL as their interface of choice. You're completely hand waving over the fact that data is just a bunch of bits on disk grouped into pages. Everything above that fact is a tradeoff. |
|
Do you need a distributed database to read a .txt file with cat, Emacs, and Firefox? No. Why not? Because there's an I/O standard they all use. Does that .txt file have to live on a single filesystem, or disk? No. Why not? Because the storage mediums all have a standard virtual filesystem interface.
There is no reason databases cannot do exactly the same thing. It's just that nobody has made them do it yet. They've stuck with the exact same paradigm, and that then drives how all of us build our systems, with this archaic 40 year old model that requires heavy-lifting maintenance windows and a lot of praying.