I'm reading through Lou Rossi's token ring paper and read something that
is ponderous. Over ethernet, an ARP request is broadcast but the reply
is unicast. In this paper (p. 4) there is a scenario where two hosts
are separated by two bridges and a ring. Host A wants to transmit to
Host B so it ARPs for B's MAC address.
Now, the paper mentions that after the ARP reply, A knows the MAC
address of B but not the location. My question is this: is an ARP
reply over token ring unicast or broadcast? If it's unicast then how
did B send the response to A? Wouldn't it have to send an explorer
packet first to find the path to A?
It seems to me that this is the process:
A wants to say hi to B
A sends an ARP request (broadcast) to B
B receives request and wants to send unicast response to A
B sends a local explorer for A (no response)
B sends an all-routes explorer for A (gets a response)
B sends a unicast ARP reply to A
Is that right? If that's the case, then here's what I gather happens
next:
A now has B's MAC address
A sends a local explorer for B (no response)
A sends an all-routes explorer for B (gets a response)
A proceeds to transmit data to B
Here's what doesn't make sense to me. If B had to know the path to A
to send an ARP reply, why doesn't A just take the data in the RIF from
that reply to figure out the reverse path back to B? Why waste time
with explorer packets when it was just given the path in the ARP
reply??
Or, do ARP requests/replies even have RIFs? If not, are they all
broadcast in token ring?
Can you tell I'm just starting out studying token ring and SRB? <g>
Thanks as usual!
John
_________________________________
FAQ, list archives, and subscription info: http://www.groupstudy.com/list/cisco.html
Report misconduct and Nondisclosure violations to [EMAIL PROTECTED]
