Geomajas database access pattern

Hi,

I've created a Geomajas 1.9.0 setup with hibernate plugin. I'm using mysql as database and have some dummy database for testing purpose. I'm using single vector layer and one osm layer (as present in the Roads example)

I wanted to know about the Geomajas database access pattern and observed following things:

• It makes around 50 select query to load and render layer.
• To add a point it makes aroud 20 select queries and 1 insert queries.
• To zoom to a feature (using search attribute option) it makes around 25 select queries.
• To select a polygon it makes around 35 select queries. 

It seems that same database interaction takes place even if we repeat some task (for example, select a polygon which we had selected previously). So I want to know if it is the expected behaviour. I'm planning to look at the hibernate caching option, but is there anything else we can do to enhance the performance. Is it only possible to enhance the performance by caching data at the server end or can we also do something at browser level where actual rendering takes place?

Thanks,

Pulkit