In response to Brian Field,

My replies to the list don't seem to be getting through, so Brian, if 
you don't see this response on the list, feel free to repost it.

Let me answer on two levels.  First, this isn't a Cisco restriction, 
but is the  way the protocol was designed.

Second, virtual links really aren't a great technology, so they 
haven't gotten much enhancement.

At a bit-fiddling level...

>Two questions regarding virtual links:
>
>1) Why is it that each end point for a virtual link
>must be the router's Router ID value?  Why is it not
>sufficient to terminate the virtual link on any of the
>target routers IP adresses or perhaps one of the router's
>IPs that are in the transit area?

Without asking John Moy exactly what he had in mind when designing 
virtual links, my impression would be that you need a stable router 
ID as the originating router for the LSA.  If you just used an 
arbitrary IP address, it would probably break the LSA aging 
mechanism, which is tied to advertising router.

>
>2) Is it possible to establish a virtual link through
>an area which is a stub area?  If not, why not?


No.

At the bit level, stub areas are distinguished by not having the 
transit-capable flag set.  Virtual links are defined to go only 
through transit areas.

Also, here's my post for your other question, if it also doesn't get through.

>In the ACRC book (page 200), the author shows the 10.64.0.1/24
>network being placed into an area 0 with the command:
>
>               network 10.64.0.1 0.0.0.0 area 0
>
>Why is the wild card mask 0.0.0.0 used and not a mask
>which matches the actual manner in which the network
>has been subneted?  Is there any difference to the operation
>of the router, area, OSPF, or area routing if the above was
>changed to:
>
>               network 10.64.0.1 0.0.0.255 area 0
>
>?
>
>If the above two forms result in identical operation,
>why is the wild card mask required?
>
>Thanks,
>Brian
>

By making the mask specific enough to match the router interface 
only, it makes troubleshooting and maintenance easier. While it is 
possible to have masks that meet multiple router interfaces, my 
experience tells me that configurations are most clear if there is 
1:1 correspondence between network statements and OSPF-speaking 
router interfaces.

___________________________________
UPDATED Posting Guidelines: http://www.groupstudy.com/list/guide.html
FAQ, list archives, and subscription info: http://www.groupstudy.com
Report misconduct and Nondisclosure violations to [EMAIL PROTECTED]

Reply via email to