Im just trying to take SQLites idea of "will persist data without a server", and add on top "will respond to write events to the SQLite file". All within a single OS - no distribution. I think maybe I confused things with the example - the "remote store" may just by another GUI that gets updated with whatever is in the SQLite file - like a live stats dashboard. I think there are only two ways to do this: 1. Modify writers to write diff events to the file. 2. Compute diff events in the reader by some type of comparison with an older snapshot. I do not want to do 1 because that means I cannot observe SQLite files from other software and respond to their writes. Bedrock is in category 1 above - each node writes its diff events to the file/system. Im asking if there is an alternative to 2 as it does not sound feasible for large databases.