Quote Originally Posted by naardejood View Post
My issue with the whole "server meltdown" idea is that Squeenix knows exactly what login numbers look like (every service on the internet has metrics for this sort of thing) and they know exactly what their servers can handle. It's clearly indicated in the hardware's specs. You can't really "overload" the hardware by accident without intentionally telling the hardware to accept more incoming connections than it's programmed to handle. You can only push so much data down a wire and the hardware on the other end of that wire is engineered specifically to handle that quantity of date. If more people than the servers can physically handle attempt to log in then they'll just get turned down by the routers.

I had a contract with an MMO publisher of note for four months, and their policy was to launch with a good deal more server capacity than they expected to need to prevent these sorts of issues. From my understanding, a lot of these large hardware contracts have deals with their server and network hardware providers that allow them to return underutilized hardware within the first three months for a partial, if not complete, refund, allowing them to OVER-provide without worrying that much about the cost. I'm sure Squeenix has the same sort of setup with whoever's providing their hardware. It's all also under warranty, so there shouldn't be any real concern for testing and whatnot beyond what the hardware specs indicate. All that testing is already done for them by the manufacturer.

Even if they end up with more worlds than they need, that's what server merges are for.

Ahh the SW:TOR way of launching.

Worked well for them...until the free playtime was over and their playerbase vanished.

Server merges are useful, but it's hardly a shining endorsement when you have to merge worlds in the first 90 days (using your contract suggestion).

We're still inside the first 30 days.