Hello,

I'm looking to see what, if any, users of Hibernate are using it for an Enterprise-scale application. Not just using it *in* the enterprise, but using it to run mission-critical components of your business.

Hibernate looks to be really cool technology, and from the looks of it can really speed up application development. But, as with any technology, if used incorrectly things can go horribly wrong. Hibernate seems particuarly fraught with such "dangers"...OO modeling is notoriously tricky to get done right, and Hibernate allows massive amounts of code re-use (potentially making it very tricky to tune and optimize things in the DAL).

So I'm curious to learn about anyone who has deployed this in a production (or production-like) setting. A TPC-W would be quite sufficient! In the absence of that, anything that supports 1000+ concurrent users and uses a database of > 100 GB would be most interesting to me!

Thanks,
:-Phil

I'm also interested in hearing about people's exeriences with hibernate. I am in the process of getting hibernate approved for use in our company. As you may guess, the powers that be are very cautious about implementing new open source technology. Toplink has been used here, with varying degrees of success. One question that will certainly be asked is, "Given the problems we've had with TopLink, why would Hibernate be any better?"

Some other questions I would have:
What kind of problems have you had using toplink in an enterprise application?

Does the existance of legacy tables that we can't change cause undue hardships?

Does the fact that Hibernate uses run-time SQL generation and reflection make debugging difficult?

Any input would be appreciated.
Thanks

joe

The problem I had was that the Toplink doco absolutely refuses to go into detail about *how* anything works. So you're left with a whole bunch of ways of doing things, but no idea when to use which and what trade-offs you're making. This was three or four years ago, so maybe it's improved since.

The Toplink project I worked on just ended up being a big mess with all these poxy little workarounds (that absolutely killed Toplink's ability to do any optimisation) because nobody knew how to use it properly.

The Hibernate doco is, on the other hand, easily some of the best doco I've ever used. It's not perfect and it won't write your code for you, but it's pretty good all the same. And you can help to make it better if you want.

I'm a nuts & bolts kind of guy, so the ability to dig into the Hibernate source is of course a big help to me, but that might not be your thing. The Hibernate source is also quite well written and easy to read (just stay out of the QueryTranslator :).

The support on the Hibernate forums is pretty good. Not much hand-holding, but that's fine by me.

Keep in mind, Hibernate is fundamentally not "easier" than Toplink or any other O/R mapping tool. You'll still need a Hibernate "go to" guy. And half your team will still be completely SQL unaware - Hibernate won't help them do persistence-oriented guff any better than they do now.

O/R mapping is a complex and difficult thing and there is no silver bullet for it. The difference is that, with Hibernate, you can put in the hard yards, figure out how it works, and then implement something that works reasonably well - all without having to pay consultants or send your team on training.