• Benchmark
• Results
• Comparison

JPA Performance Benchmark (JPAB)

Comparison of EclipseLink with H2 embedded vs DataNucleus with Derby server

Each of the following tables focuses on a specific database operation, where the last table presents average results comparison.

Speed comparison of JPA database persistence operations (normalized score, higher is better)

Transaction Size =>	Few Entities		Many Entities		Average Score
	EclipseLink H2 embedded	DataNucleus Derby server	EclipseLink H2 embedded	DataNucleus Derby server	EclipseLink H2 embedded	DataNucleus Derby server
Basic Person Test	25.1	2.8	8.9	2.3	17.0	2.6
Element Collection Test	20.4	0.71	8.0	0.35	14.2	0.53
Inheritance Test	28.7	2.7	8.7	2.0	18.7	2.3
Indexing Test	37.6	4.5	14.0	4.0	25.8	4.3
Graph (Binary Tree) Test	17.7	0.62	14.7	0.53	16.2	0.57
Multithreading Test	30.3	4.9	11.4	3.4	20.8	4.1
All Tests	26.6	2.7	10.9	2.1	18.8	2.4

The results above show that in general EclipseLink with H2 embedded is much more efficient than DataNucleus with Derby server in persisting JPA entity objects to the database. Comparing the normalized speed of DataNucleus with Derby database server (2.4) to the normalized speed of EclipseLink with H2 embedded database (18.8) reveals that in these tests, EclipseLink with H2 embedded is 7.8 times faster than DataNucleus with Derby server.

A huge performance gap has been detected when using JPA element collections with small transaction size. Comparing the normalized speed of DataNucleus with Derby database server (0.71) to the normalized speed of EclipseLink with H2 embedded database (20.4) reveals that in that case, EclipseLink with H2 embedded is 28.7 times faster than DataNucleus with Derby server.

Speed comparison of JPA database retrieval operations (normalized score, higher is better)

Retrieval Size =>	Few Entities		Many Entities		Average Score
	EclipseLink H2 embedded	DataNucleus Derby server	EclipseLink H2 embedded	DataNucleus Derby server	EclipseLink H2 embedded	DataNucleus Derby server
Basic Person Test	30.4	0.56	29.1	6.3	29.7	3.4
Element Collection Test	19.6	0.36	27.6	0.93	23.6	0.64
Inheritance Test	33.8	0.45	39.8	6.8	36.8	3.6
Indexing Test	22.2	0.50	41.3	7.1	31.8	3.8
Graph (Binary Tree) Test	16.0	0.22	20.0	6.6	18.0	3.4
Multithreading Test	33.7	1.2	29.0	10.0	31.4	5.6
All Tests	25.9	0.55	31.1	6.3	28.5	3.4

The results above show that in general EclipseLink with H2 embedded is much more efficient than DataNucleus with Derby server in retrieving JPA entity objects from the database. Comparing the normalized speed of DataNucleus with Derby database server (3.4) to the normalized speed of EclipseLink with H2 embedded database (28.5) reveals that in these tests, EclipseLink with H2 embedded is 8.4 times faster than DataNucleus with Derby server.

A huge performance gap has been detected when using class inheritance in the object model with small retrieval size. Comparing the normalized speed of DataNucleus with Derby database server (0.45) to the normalized speed of EclipseLink with H2 embedded database (33.8) reveals that in that case, EclipseLink with H2 embedded is 75.1 times faster than DataNucleus with Derby server.

Speed comparison of JPA database query operations (normalized score, higher is better)

Retrieval Size =>	Few Entities		Many Entities		Average Score
	EclipseLink H2 embedded	DataNucleus Derby server	EclipseLink H2 embedded	DataNucleus Derby server	EclipseLink H2 embedded	DataNucleus Derby server
Basic Person Test	4.4	37.7	1.4	3.6	2.9	20.7
Element Collection Test	4.1	25.9	2.4	1.4	3.2	13.6
Inheritance Test	3.4	6.5	2.0	1.8	2.7	4.2
Indexing Test	13.0	0.030	22.8	4.9	17.9	2.5
Multithreading Test	1.0	failed	0.67	failed	0.85	failed
All Tests	5.2	17.5	5.9	2.9	5.5	10.2

DataNucleus with Derby server has failed in 2 tests (see exceptions).

The results above show that in general DataNucleus with Derby server is more efficient than EclipseLink with H2 embedded in executing the tested JPA queries.

A large performance gap has been detected when using simple basic entities with small retrieval size. Comparing the normalized speed of EclipseLink with H2 embedded database (4.4) to the normalized speed of DataNucleus with Derby database server (37.7) reveals that in that case, DataNucleus with Derby server is 8.6 times faster than EclipseLink with H2 embedded.

On the other hand, DataNucleus with Derby server is slower, for instance, when using database indexes with small retrieval size. Comparing the normalized speed of DataNucleus with Derby database server (0.030) to the normalized speed of EclipseLink with H2 embedded database (13.0) reveals that in that case, DataNucleus with Derby server is 433 times slower than EclipseLink with H2 embedded.

Speed comparison of JPA database update operations (normalized score, higher is better)

Transaction Size =>	Few Entities		Many Entities		Average Score
	EclipseLink H2 embedded	DataNucleus Derby server	EclipseLink H2 embedded	DataNucleus Derby server	EclipseLink H2 embedded	DataNucleus Derby server
Basic Person Test	16.6	1.1	7.6	0.85	12.1	0.99
Element Collection Test	15.3	0.81	10.5	0.62	12.9	0.71
Inheritance Test	15.7	1.2	9.8	1.3	12.7	1.2
Indexing Test	12.1	1.2	9.9	1.5	11.0	1.3
Graph (Binary Tree) Test	13.7	0.29	7.0	0.41	10.3	0.35
Multithreading Test	20.9	failed	4.2	failed	12.6	failed
All Tests	15.7	0.93	8.2	0.92	12.0	0.92

DataNucleus with Derby server has failed in 2 tests (see exceptions).

The results above show that in general EclipseLink with H2 embedded is much more efficient than DataNucleus with Derby server in updating JPA entity objects in the database. Comparing the normalized speed of DataNucleus with Derby database server (0.92) to the normalized speed of EclipseLink with H2 embedded database (12.0) reveals that in these tests, EclipseLink with H2 embedded is 13.0 times faster than DataNucleus with Derby server.

A huge performance gap has been detected when using graphs of objects with small transaction size. Comparing the normalized speed of DataNucleus with Derby database server (0.29) to the normalized speed of EclipseLink with H2 embedded database (13.7) reveals that in that case, EclipseLink with H2 embedded is 47.2 times faster than DataNucleus with Derby server.

Speed comparison of JPA database removal operations (normalized score, higher is better)

Transaction Size =>	Few Entities		Many Entities		Average Score
	EclipseLink H2 embedded	DataNucleus Derby server	EclipseLink H2 embedded	DataNucleus Derby server	EclipseLink H2 embedded	DataNucleus Derby server
Basic Person Test	35.1	1.3	19.6	1.3	27.4	1.3
Element Collection Test	19.3	0.61	8.5	0.33	13.9	0.47
Inheritance Test	36.0	1.0	14.0	1.3	25.0	1.2
Indexing Test	42.5	2.3	19.2	1.9	30.9	2.1
Graph (Binary Tree) Test	12.2	0.26	14.4	0.46	13.3	0.36
Multithreading Test	54.4	failed	25.8	failed	40.1	failed
All Tests	33.3	1.1	16.9	1.1	25.1	1.1

DataNucleus with Derby server has failed in 2 tests (see exceptions).

The results above show that in general EclipseLink with H2 embedded is much more efficient than DataNucleus with Derby server in deleting JPA entity objects from the database. Comparing the normalized speed of DataNucleus with Derby database server (1.1) to the normalized speed of EclipseLink with H2 embedded database (25.1) reveals that in these tests, EclipseLink with H2 embedded is 22.8 times faster than DataNucleus with Derby server.

A huge performance gap has been detected when using graphs of objects with small transaction size. Comparing the normalized speed of DataNucleus with Derby database server (0.26) to the normalized speed of EclipseLink with H2 embedded database (12.2) reveals that in that case, EclipseLink with H2 embedded is 46.9 times faster than DataNucleus with Derby server.

Comparison of JPA/Database speed - the averages (normalized score, higher is better)

Transaction/Retrieval Size	Few Entities		Many Entities		Average Score
	EclipseLink H2 embedded	DataNucleus Derby server	EclipseLink H2 embedded	DataNucleus Derby server	EclipseLink H2 embedded	DataNucleus Derby server
Basic Person Test	22.3	8.7	13.3	2.9	17.8	5.8
Element Collection Test	15.7	5.7	11.4	0.72	13.6	3.2
Inheritance Test	23.5	2.4	14.9	2.6	19.2	2.5
Indexing Test	25.5	1.7	21.5	3.9	23.5	2.8
Graph (Binary Tree) Test	14.9	0.35	14.0	2.0	14.5	1.2
Multithreading Test	28.1	3.0	14.2	6.7	21.1	4.9
All Tests	21.9	3.8	14.9	2.8	18.4	3.3

The results above show that in general EclipseLink with H2 embedded is much more efficient than DataNucleus with Derby server in performing JPA database operations. Comparing the normalized speed of DataNucleus with Derby database server (3.3) to the normalized speed of EclipseLink with H2 embedded database (18.4) reveals that in these tests, EclipseLink with H2 embedded is 5.6 times faster than DataNucleus with Derby server.

A huge performance gap has been detected when using graphs of objects with small transaction/retrieval size. Comparing the normalized speed of DataNucleus with Derby database server (0.35) to the normalized speed of EclipseLink with H2 embedded database (14.9) reveals that in that case, EclipseLink with H2 embedded is 42.6 times faster than DataNucleus with Derby server.

Other Head to Head DBMS/JPA Comparisons

JPA Performance Benchmark - Copyright © 2010, 2011, 2012 ObjectDB Software Ltd. - All rights reserved.
The benchmark program is released as open source under the GPL license.