I've built an event-sourcing/stream framework on top of Fjall v3, and it's been a fantastic fit. It's a great building block for many systems where a more high-level storage system is either overkill or a poor match for the access patterns you actually have (in my case, sequential access to prefix-filtered data gets you 95% of the way to an efficient persistent stream). I started trying to layer things on top of PostgreSQL and similar, and the impedance mismatch and resultant unpleasant code (and performance!) made me look elsewhere.
No comments yet.