I am currently forumulating a strategy for our MySQL servers to provide performance, redundancy, backups and high availability.

I have been reading a lot about MySQL's replication capabilities. Several places recommend routing all write queries (insert, update, etc.) to the master server, and routing all read queries(select) to the slave servers.

I am wondering how this is accomplished. Is this something that can be set up in MySQL to do this, or is it handled via application logic? Could someone possibly provide an example of this?

Thanks in advance for any advice!

Aaron

I'm also quite interested in such an architecture. If you find a solution somewhere, I would appreciate if you briefly described it here.