A newbie-question about DATABASE (tables) LOAD BALANCING

[Davide C.]

Hi, I'm new to mysql and load balancing strategies, so please forgive me if I ask a stupid question.

I'm developing a dating site and a crucial part of this project is the search. I mean, users can search, say, a girl who's 26 years old, that comes from Germany, who has blue eyes and so on...the problem is that I really don't know how to balance the load of the queries (I use PHP). I think that you all know the problem : when 5000-6000 people are searching on the same time quering the same tables, the time to wait for a response gets longer and longer.

Someone has told me to use Memcached in order to store query results in a object to avoid repetitive database query. This is good, of course but I'm sure there are other strategies.
Another friend suggested me to use Memcached as a semaphore. To replicate the same "search" table 8 times setting a memcached-value to 1 indicating which tables are currently locked (so other queries use other tables, if there are free) and after the query setting this value to 0.
I made some test that revealed that this is useless. MYSQL handles the query queues better and FASTER than this system. Using only 1 table and letting mysql handle the queues leads to (almost) better results!

So, if anybody has some suggestion...

Thanks in advance

Davide (from Switzerland)