JDBM is key value embedded java database. It provides java collections backed-up by disk storage. JDBM is also very fast (million inserts per second). It also has instance cache, transactions and other goodies. More details at .
I would like to restart 3 years old discussion  about making it ApacheDB subproject.
I believe that JDBM in last 12 years proved itself as viable project. And new JDBM3 with its speed and agility has chances for massive adoption. So I think it is good moment to stabilize development a bit and get strong partner.
Personally I am committed to maintaining JDBM as I did for last three years. Our community is not big, I am only regular dev. But JDBM is very 'hackable' and we often get patches from strangers.
I think JDBM3 is ready to become Apache sub project:
- well traced history of contributions, only 3 main devs
- already under Apache 2 licence
- changing package name is not problem (JDBM3 has completely new API and zero adoption)
- uses Maven2 including website
- good unit test coverage
JDBM3 is currently in alpha stage. It should reach beta in about three months. Stable probably in a year or two.
After JDBM3 is stable I will make it rock stable and concentrate on tools around (spring and JTA integration, .NET port etc...)