Layer 2 broadcast storm (Need clarification please)

Thanks for the reply.
That makes sense, but it still leaves me a little perplexed as
to how a switch can even ping out when its a device that operates at layer 2...
Regards,
Shamrock

Switches are layer 2 devices...But they can be assigned an IP address to be accessed via layer 3. The switch however won't "DO" anything on layer 3 (routing, filtering), it simply acts as a client, it still only dose "work" on layer 2. The ip address is simply to give you a way to manage the device besides plugging into the console port.

makes sense. It was simply something that had me confused


Thanks and regards.
Shamrock

Thanks for that, good link!
One question left:
Do broadcast storms run indefinitely around a
Layer 2 network?Is there a point where a broadcast storm
dies, such as some sort of Time To Live with a frame?
thanks!!!
Regards from Ireland,
Shamrock

Not sure if you've gotten your answer yet, here's what I know about layer 2
broadcast storms or loops in a redundant network. STP must be run as another
poster pointed out IF there are redundant links. RSTP converges faster than STP
which uses timers to allow up to 50 seconds for the network to converge.

Layer 2 broadcasts have no ttl field and can run until a device is unhooked.
I've seen this before on RSTP rings and its UGLY. Cisco phones run layer 2 and
can cause the same thing. My experience was with an industrial NW with some
crappy endpoint Nics designed by retards at a California state agency which I
will leave un-named.

A Cisco switch can become your Root bridge which makes trouble shooting much
easier and I'm told is much more reliable. Layer 3 routing loops will eventually
subside because of the limited hop count (time to live) in routers. RIP uses 15
as the max hop count and works surprisingly well with that. You force it by
lowering its cost on the network to be root. Other routing protocols have ways
of negating routing loops.

Layer 3 broadcasts are sent out at Layer 2 with a 48 Bit FF:FF:FF:FF:FF:FF
address but have to traverse up to the IP layer which has a TTL time to live
field. That is why Layer 3 loops are relatively harmless compared to STP rings
using bridges. Bridges create collision domains but , pass Arp broadcasts
creating one huge broadcast domain.

Just in case you're visual...
You'll notice that there is no sort of TTL, or anything that would keep it from running forever (as mentioned by ir0ck).

L3 broadcast storms should also get stopped by split horizon, route poisoning and reverse route poisoning, and some other stuff to make sure that this never happens, not just TTL's

In the lab I have created broadcast storms (heh! ) deliberately to see what happens. In some cases interfaces in L2 switches (Cisco 2950's) have gone into ERR_DISABLE mode all by themselves, with no intervention from Spanning Tree at all. Not sure why they do that, but it does at least stop the storm until ERR_DISABLE times out, which is set to 5 mins by default.

I *love* Spanning Tree :oops: