JPA distributed cache vs JDBC

I have to make an开发者_StackOverflow application with multiple nodes connected to a single database in the backend. all nodes need to be in sync as any client can go through any node and should expect the same result. my question is if anywone has used some JPA distributed cache(something like eclipselink coorinated cache) for this and knows what the benefits of use JPA would be in this situation vs just using JDBC. I understand i would get the caching benefits but would there be any other reason ?

If your question is, what are the advantages of JPA distributed cacheing over simply turning on cacheing in your database/driver - recognize there is still work to marshal/unmarshall objects that the JPA implementation does after the data has been retrieved from/sent to the database. This effort can be eliminated if the distributed cacheing can simply share the latest version of those objects without hitting the database.