Pear DB on mysql replication cluster

does anyone ever tried to use Pear DB to access on Mysql replication cluster? i mean... while one can update/select on mysql master server, you can just do selects on the slaves. it would be useful to implement such as a dsn-array using to connect to a master when you have a manipulation query (i.e. using DB::isManip()) or to the slaves pool when the query is not. one can also think to build something like a round-robin on the slave choice. does someone would give me a hint on how to implement it? bye, alberto. -- "Imagination is more important than knowledge"