Caching DB information

hello,

I'm looking for information in relation to caching relational database information in a C/C++ UNIX environment.

Are there any existing/useful libraries/interface available from both a C and C++ perspective on the UNIX platform.
Or is ODBC library the defacto standard interface.

Basically, with typical UNIX processes (written in C and C++), I'm looking at a scheme where data can be queried and cached at process initialisation

what are the best/proven techniques/methodologies for modelling data at the application level?
ie. how to best model and store the data returned from queries?

Any insight/suggestions appreciated.

Thanks.