HN Explorer

Ask HN: Embedded Database Design Papers

1 points by gfs ↗ HN
In order to learn more about database design principles I thought it would be great to read papers on embedded databases. I have searched around for material on the popular open source databases like RocksDB, LMDB, or LevelDB but haven't found anything concrete. Aside from reading the source code, how can I teach myself about the data structures and abstractions that these provide? Eventually I would like to write a toy implementation after getting acquainted with a subset of this knowledge.

4 comments

[ 2.3 ms ] story [ 23.5 ms ] thread
SamReidHughes
RocksDB and LevelDB use a "log-structured merge tree" so you might want to use that as a search term. There is a section "Papers" at the bottom of the web page at http://leveldb.org/ .

LMDB uses a B-tree variant, according to Wikipedia. Maybe it's a B+ tree as Wikipedia says, but keep your thinking cap on -- it's probably not in exactly the form you see described on Wikipedia.

If you want to see a toy LSM-tree implementation, I have one at https://github.com/srh/nihdb -- it was written as a starter Rust project, not as a fast LSM-tree engine. It's almost the simplest plausible LSM-tree. But you'll miss out on design considerations relating to concurrency and performance.

philix001
Goetz Graefe [1] published interesting surveys on B-Trees, query optimization...

The Red Book can give you many ideas [2].

Mark Callaghan has been published a lot of stuff related to LSM recently [3].

[1] http://citeseerx.ist.psu.edu/viewdoc/download?doi=10.1.1.219...

[2] http://www.redbook.io/

[3] https://smalldatum.blogspot.com/?m=1