I've been busy building out a fairly realistic test application for the past couple of days so I can really get a handle on Hibernate's performance and scalability. The application is a book store e-commerce site very loosely based on Amazon. I'll make the code available once it's complete.

I've defined a stateless session bean interface with the business methods e.g. login, findBooks, createOrder, addBookToOrder, and so on. I currently have two implementations of the bean - one that uses JDBC DAOs and one that uses Hibernate. I'm considering implementing another version that uses Spring JDBC as well.

I've developed a tool that can populate the database with configurable amounts of random data so that the tests are representative of real-world data volumes.

I then have one standard multi-threaded test client that will work with any of these bean implementations. I'm hosting the beans in JBoss on a dedicated server so that each implementation will be using the same connection pool and transacation management to try and make the tests as fair as possible.

The initial results are showing poor scalability with Hibernate and I have a thread open on the Hibernate forum where I'm looking for guidance on my configuration and approach in case I'm not using Hibernate correctly. At this stage it's also possible I have bugs in my code so I'm not going to post any figures until I'm confident that my tests are correct.

I'm starting a new contract tomorrow so I'll have to put this work on hold for a couple of days but hopefully I'll be able to post some stats at the weekend.

Labels: java, persistence