Is it feasible to use hibernate for developing portal applications? Will the system become slow due to the existence of an additional data management layer when a number of concurrent users access the portal through the internet?

Yes (it's feasible). No (it won't become slow).

Hibernate is used for exactly this purpose. It works, it's fast, it makes life so much easier once you've finished initial development. Maintenance and enhancements become very easy once you've successfully mapped your data model with hibernate.

When you think logically, the existence of one more layer will naturally slow down the system. I feel that writing database specific code will work faster than writing generic code. I also feel that it will have a bad effect more on web based applications where concurrent hits can slow down the system. I agree to the fact that maintenance will be quite easy and I love the way hibernate has developed the persistence model. But could you please give some proof on how hibernate will be efficient for portals in order to clear my doubts?

Experience > proof. There are many such apps in existence using hibernate, they don't worry about it.

The few extra microsecnds spent converting result sets into objects will be offset by the fact that the generated queries are better than the ones you'd write yourself, and by the fact that hibernate won't fetch objects that you don't need.

When writing apps that access the DB, don't forget that optimization always comes after you've found a bottleneck, not before. Don't worry about it, it'll be faster than you think, and likely faster than if you did it youself (unless you're as good at designing queries as the hibernate team are).

Thanks for your quick reply. I am taking your word for granted.