The offline-first technology landscape is dominated by CRDTs and logical replication. These technologies offer flexibility via techniques like fine-grained conflict resolution, schema evolution, and storage portability. So it's no surprise that physical replication has been left in the dust, forever destined to just be that thing databases do... or is it?

In this talk, I'll present my research on using physical replication to enable offline-first applications. I'll share some of the key benefits offered by physical replication, and how to work around some of its downsides. Finally, if the demo gods are willing, I'll showcase some things you can do with physical replication which are much harder to do with a logical solution.