> This sounds like a tall order to me Yes it may be too difficult but there are other things SQLite could do which would not be as invasive. For example, it could do priming reads for a range of pages it will need in the future and let the VFS asynchronously fetch/cache those reads for a future blocking read. That would neatly avoid any issues with asynchronous completions but I don't know how hard it would be to adapt SQLite for that model. All I do know is that trying to implement some kind of readahead in the VFS would be unlikely to yield the benefits we would like. I'm not that knowledgeable with SQLite internals to help with this project (sorry). This will probably be as deep as I go. :)