the part that still bites is the retry loop one bad IP is recoverable one bad IP plus polling every second until it answers turns every outage into a flood