On the server I lease for BCF, I hired a professional server management team to "harden" the server to reduce (not eliminate) risk of hacking (Another reason I appreciate those of you who subscribe - to help pay for such things). One of the biggest issues I have, that is not really server-related, is the number of spammers who attempt to register here daily.
I have a three-step process to minimize the possibility of spammers / hackers from registering. First, I use "human verification" to minimize the possibility of bots registering. Second, for those that get past the human verification, I have a third party program that checks the user name, IP address and email address of all new registrations. It checks these against an off-site, third party database of known spammers. If any of these three items matches a known spammer in the database the registration is rejected and not even sent to the registration queue.
For those that do make it to the registration queue, I personally look at each registration before I approve (or delete) the new member account. If anything looks fishy, I will usually trace the IP address. For example, if they say they are located in California, but the IP traces to Russia, I delete that account with no notification. A lot of times, I'll get a registration and where it asks "What British Car Do You Own?" they might say "BMW or Mercedes" - delete. If everything looks good, I will approve it. These steps are about 99% effective, but every once in a while a spammer will sneak past all my checks. Rare, but it happens.
Just to give you an idea of how significant the problem of spammers is, I've attached a screen shot from my Anti-Spam program's log file. All these (except a few) are spammers the system stopped cold! Keep in mind that this is just half a day's worth today (and this was a slow day). (Note, for security reasons I've blurred out partial IP addresses of spammers and also blurred the emails of any legitimate members (for privacy reasons).