JE vs Derby Whitepaper
Over the past few months I've been working on comparing Berkeley DB Java Edition to Apache Derby. The result of this effort is a whitepaper that compares the two systems in "CRUD" tests. I embarked on this benchmark because I was curious about Derby and its performance characteristics. I certainly do not intend to launch into a benchmark war against Derby -- there are some things that Derby does better than JE (like ad hoc queries) and some things that JE does better than Derby. It all depends on your application's needs.
From the text:
This paper compares the performance of Oracle Berkeley
DB Java Edition to Apachei??s pure Java RDBMS, Derby. In some areas, this comparison is
apples-to-apples, and in others it is apples-to-oranges. For instance, both JE and Derby
are embeddable, transactional, pure Java storage engines
(apples-to-apples). On the other hand,
JE provides schema and a higher level data abstraction without SQL using a
faster direct access API, features Derby provides
with SQL (apples-to-oranges). The reader
should evaluate whether the additional features provided by a relational
database system (RDBMS) are required, at the expense of reduced
performance. In all cases comparisons
are functionally equivalent despite different underlying implementations.
In the tests described in this white paper, Berkeley
DB Java Edition performance exceeds Derby performance by a factor of 3 to 10.
