SQLite only works as a concept because it is not networked. Nobody truly understands the vast and unsolveable problem that is random shit going wrong within the

SQLite only works as a concept because it is not networked. Nobody truly understands the vast and unsolveable problem that is random shit going wrong within the communication of an application over vast distances. SQLite works great because it rejects the dogma that having one piece of software deal with all of that shit is in any way a good idea. Back your dinky microservice with SQLite, run multiple copies, have them talk to each other and fumble about trying to get consensus over the data they contain in a very loose way. That will be much, much less difficult than managing a distributed decentralized database (I speak from experience). It's good enough for 90% of cases. Remember P2P applications? That was basically the same thing. A single process running on thousands of computers with their own independent storage, shuffling around information about other nodes and advertising searches until two nodes "found each other" and shared their data (aw, love at first byte!). It's not great, but it works, and is a lot less trouble than a real distributed database.