From: Howard C. Berkowitz (hcb@gettcomm.com)
Date: Wed Jan 01 2003 - 15:50:06 GMT-3
At 5:15 AM +0000 1/1/03, Sage Vadi wrote:
>Hi all,
>
>Debug dialer events/packets:
>
>BR0 DDR: ip (s=133.10.33.2, d=224.0.0.5)
>
>Q) If the IOS that you are using supports OSPF demand
>circuits, should this not neutralise the router from
>sending out OSPF multicasts periodically? Is this
>normal behaviour?
Any OSPF router should respond to 224.0.0.5. Yes, that's a multicast
address, but it simplifies OSPF's implementation not to have to track
the specific address at the other end. At present, the only OSPF
packets with unicast destinations are used in retransmission.
RFC1793, defining OSPF demand circuit extensions, makes no mention of
changes in destination address.
If I understand your question, OSPF should only send out updates on a
demand circuit when there is a topology change. If it didn't, the
remote router would lose database synchronization. So, there should
be _some_ OSPF activity.
Demand circuits do not get routine hellos. Point-to-point do.
There is an Internet Draft, draft-ietf-ospf-dc-05.txt, that proposes
a neighbor probe mechanism to check that the OSPF speaker at the
other end is alive, but it specifically will not bring up the demand
circuit to check this. While there's a Cisco coauthor, I don't know
the status of Cisco implementation (if any).
Multicast addresses only have a different effect on real or simulated
broadcast networks.
>
>As a workaround I have setup an ACL that only permits
>ICMP, thus the OSPF multicasts would never trigger the
>link anyway.
.
This archive was generated by hypermail 2.1.4 : Sat Feb 01 2003 - 07:33:38 GMT-3