yeah, i remember in the old days, it was a must to put ip default-gateway in layer 2 switches so we can manage them. but how is possible the newer switches can operate, and be managed/pinged without adding a default-gateay?
im really puzzled.
While the response from NT gives the traditional explanation (you need ip default-gateway if operating a layer 2 switch and need a default route (static or dynamically learned) if operating at layer 3) I have discovered that the newer Catalyst switches do not behave in exactly that way. I have had the experience that I believe the original poster is asking about: I have had switches that do not have ip default-gateway and do not have a default route but are still able to access other network resources. I discovered that the explanation is that the switch would ARP for all IP addresses that you attempt to access (both for local subnet addresses (the expected behavior) and for addresses in remote subnets (not the expected behavior)) and if the connected layer 3 device has enabled proxy arp then the layer 2 switch has IP connectivity to the complete network.