> There are a few blogs up about major battles where their servers are just being thrashed and how they were able to stay up. They also mention that their users warned them about this battle before hand which was really cool.

So, I know that I'm expecting too much, but remember that the only way CCP can handle large fights is to simulate them at 1/10th realtime, and get advance notice of said fights so that they can move the simulation for the target system to a single, beefy server ahead of time.

It's a goddamn shame that they don't have the resources available to figure out the -admittedly difficult- problem that is

* Reworking their simulation to be parallelizable across nodes

* Dynamically shifting slices of a battle within a system across physical nodes