We're back to the old problem of having to capacity plan for a 99th percentile event.
Yes, it would be great if the server was "snappy" when there are 1000+ concurrent users hammering on it. However, the costs associated with additional hardware, colocation, and bandwidth are prohibitive. What we have works very well and has tons of excess capacity 99% of the time.
Do we really want to have to raise a bunch of money and then double (at least) our monthly costs just so the server can perform well during "one off" events?
|