We’ve started deploying HP 1910 8 port switches to isolated areas of our network, because they are reasonably priced. Our core devices are HP 28xx series switches and Cisco 29xx series switches.
We’ve noticed that the HP 1910 units aggressively broadcast ARP requests for everything in their ARP tables. We have about 30 of the 1910’s now, and they’re responsible for 90% of our internal network traffic—2,000-4,000 packets per second of nothing but ARP requests.
It does calm down every now and then—you might get 30 seconds while nothing is transmitting, then it starts up again. Watching an individual switch’s traffic, they seem to chatter every 2-3 minutes, which roughly matches the ARP ageing default. The problem is, for an ARP table with 100 entries, it might send 20k ARP requests!
Does anyone have any ideas on how to make these calm down, or at least make them act like our Cisco and 28xx series switches, that don’t have this kind of flooding issue?