This incomplete tutorial is about how to write a sqlite clone, with particular emphasis on how to implement the underlying b-tree: https://cstack.github.io/db_tutorial/

I just implemented one in go, working on the repl now. I plan to eventually write it in C++ for a real database, but I’m doing this for my team to teach them the difference between LSM trees and Btrees. I’m hoping to show that a specialized concurrent btree, though harder to write, performs better than LSM trees for the same data. Maybe I’ll make a blog post about it.

EDIT: For more context, this is for time series metrics, so super specialized data.