Speed up database access

We have a Sybase database which we regularly connect. Recently we started to build multithreaded applications, and the connections to the database started to increase. There are several applications connect to same database some c++ executables through RougeWave libraries, and many Matlab toolbox connections, and perl, ruby scripts.

I am thinking if there is another layer of interface that is used to receive data (maybe even control the writes to DB), data access might be faster. This interface code should make the connections to database and to the applications, but applications should not directly query the DB. This interface code should cache the data queries that are regularly requested, and might need to run multithreaded to respond multiple requests at the same time, but use a single cache.

I have actually heard this kind of caching is common for high volume database connections. Are there any open source examples??

Hibernate is a Java ORM interface to databases, which leverages quite a bit of caching to minimize database access. You might glean some ideas from analyzing how Hibernate facilitates this.

Nice information.