The Routing Over Large Clouds Mailing List Archive by date

Cell Relay Retreat>List Archive>month:1994-Dec> msg00016



[Date Prev][Date Next][Thread Prev][Thread Next]  
  [Date Index][Thread Index][Author Index][Subject Index]

From ARP to NHRP.... Re: From ARP to NHRP....

  • From: Grenville Armitage <gja@thumper.bellcore.com>
  • From: Grenville Armitage <gja@thumper.bellcore.com>
  • Date: Sat, 10 Dec 1994 14:57:30 -0500
  • cc: gja@thumper.bellcore.com, rolc@maelstrom.timeplex.com



>>Grenville:
>>
>>RFC 1577 assumes that when the address to be resolved
>>is not in the same subnet as the host issuing the ARP,
>>is not in the same subnet as the host issuing the ARP,
>>then the route should go through a router in the subnet
>>acting as a gateway to other LIS's. As such, 
>>acting as a gateway to other LIS's. As such, 
>>when the host receives a reply
>>through the NHRP, it has no idea that it came from
>>NHRP and would cache the address as the one of its
>>gateway router.

On face value I think your concern is based on a different
understanding of rfc1577 host-behaviour than mine. Currently
an rfc1577 host would _never_ query the ARP Server for an IP
an rfc1577 host would _never_ query the ARP Server for an IP
address outside the LIS, because the host's own local routing
table would indicate that a LIS attached router's address is
to be resolved instead. I parse your comment as saying that
the ARP Server itself locates a LIS attached router when hosts need
the ARP Server itself locates a LIS attached router when hosts need
to get outside the LIS, which is wrong (apologies if I've
misunderstood you). Which removes the basis for your concern
(I think).

gja