EJDB2 Roadmap

  • Get rid of the Tokyocabinet. TC is an awesome software but nowadays it has many hard to fix limitations.
  • Use highly concurrent B+Tree index structure with support of MVCC. I have thoughts how to implement it with some ideas from LMDB copy-on-write btree and B-link trees.
  • Review usage of BSON data format as the internal representation of documents. Considering more compact formats: MessagePack, Smile, UBJSON, CBOR