Salted hashes don't seem to offer any real protection for user passwords by itself. So long as a hacker has a quick way to check if a password is right or wrong, it's trivial to brute force it to view the plain-text passwords. Especially if the attacker is using GPUs or an FPGA/ASIC solution.
From what I've gathered, I should be sending my hashed+salted passwords to a (hardened) dedicated server on my network whose sole task is validation (via secret keys). A hacker would have to gain access to both servers before plain-text passwords could be retrieved. I like this solution, but I don't like the cost of renting another server.
Are there any alternatives to using a dedicated validation server?