From: Curtis Call (curtiscall@xxxxxxxx)
Date: Mon May 28 2001 - 17:10:35 GMT-3
I think this is already a settled issue but I wanted to throw a small lab
together anyway and this is what I came up with:
A B C D E
area2 | area 1 | area 0
vl<------------------------>vl
Wonderful drawing I know. A and B are in area 2. B, C, and D are in area
1, and D, E and B (via the virtual link) are in area 0. B has a virtual
link with D. In order to test this I sent a ping packet from A (in area 2)
to E in area 0. Following OSPF rules this packet should first go to B
since it is in area 0, and it should then always stay in area 0 since it is
destined for an intra-area destination. Now, in order to observe what we
are trying to see here I made a static route on router C (which once again
is entirely in area 1) that pointed any packet that is destined for router
E's loopback (10.0.0.5) to it's own loopback. Here is a snippet of the
debug traces:
01:37:26: IP: s=10.0.0.1 (FastEthernet0/0), d=10.0.0.5 (Loopback0),
g=10.0.0.5,
len 100, forward
So, you can see that a packet that was sent from Router A (10.0.0.1),
destined for Router E (10.0.0.5) was sent hop by hop over router C. I
think the question to ask yourself in order to fully understand the issues
here would be, if there was an actual tunnel running between B and D, would
C ever see an IP packet with a destination IP of router E, why or why not?
Just as a side comment. I used Juniper routers for both sides of the
virtual link, and Junipers require a special tunnel PIC in order to perform
tunneling. These two routers in particular are not equipped with that PIC
so if tunneling was actually used for virtual links, this would have never
worked without installing that PIC first.
**Please read:http://www.groupstudy.com/list/posting.html
This archive was generated by hypermail 2.1.4 : Thu Jun 13 2002 - 10:30:55 GMT-3