An EJB3 stateless session bean exposing CRUD operations for an entity bean will contain read / write methods which go against a single connection pool. If you would like to have your read & write operations go against different databases for scalability purposes, how do you accomplish this at the code level. Any pointers to appropriate design patterns here?
Predominantly most of the sites, push read (i.e. select operations) only content to point to different database servers (i.e. by replicating the data from the master database to the slave database).
How do you acheive code separation / UI separation to exploit the same, since all persistence relevant code (i.e. insert / update /delete) needs to talk to the master database server and all read operations go to the slave database servers. Appreciate any inputs on this topic.