Hacker Newsnew | past | comments | ask | show | jobs | submitlogin

Good to see some more activity in embedded data space. Earlier used BerkleyDB [1], msql [2] (from where MySQL was built), Postgres 4.2 [3] and now settled on SQLite for all our embedded database needs. When need performance at present just use SQLite in memory db.

Will give Sled a try when it reaches 1.0.0 as at present SQLite serves our embedded requirements.

[1] https://launchpad.net/berkeley-db

[2] https://hughes.com.au/products/msql/

[3] https://dsf.berkeley.edu/postgres.html



Sled is not really a competitor to those.

Sled is an alternative to other on-disk key/value databases like RocksDB and LMDB.


Make sure to check out lmdb if you haven’t. It’s not relational and has an extremely anemic API but it makes a great low-level building block. Sled is similar I’m that regard, very different from the others you have named.


I've seen some startling performance claims about LMDB, but don't know anyone using it.

Why isn't LMDB used as an engine for e.g. MySQL? Why, for example, did Facebook go with LevelDB in MyRocks instead of LMDB?


My company uses LMDB extensively. LevelDB is a LSM database. LSM is generally better for write workloads over BTree DBs like LMDB and sled. LMDB also has a single writer restriction.


BIND9 can use LMDB for storing the configuration of dynamically added zones. https://www.isc.org/blogs/bind-release-911/




Consider applying for YC's Winter 2026 batch! Applications are open till Nov 10

Guidelines | FAQ | Lists | API | Security | Legal | Apply to YC | Contact

Search: