The Routing Over Large Clouds Mailing List Archive by date

[Juha Heinanen <Juha.Heinanen@datanet.tele.fi>]

i took a look at 1620 and i don't think the authors have made a fair
conclusion about the "extended arp" scheme when they say:

      However, solving the SM problem at the LL has several costs.  It
      requires a complete round-trip time before the first datagram can
      flow.  It requires a hop count in the ARP packet.  This seems like
      a tip-off that the link layer may not be the most appropriate
      place to solve the SM problem.

nbma arp as presented in the previous ietf meeting implements arp
servers in routers.  one can thus start sending ip packets to the next
hop router immediately and at the same time issue an nbma arp request.

what comes to the hop count in the arp packet, it is true that it is a
change in the original arp packet format.  but the change need to be
made only in the hosts that are connected to nbma networks and at least
in the case of atm those all all new hosts that need to be changed
anyhow.  remember that also atm arp introduces a change in the arp
packet format.

-- juha