If tunnel is the must-to-do then of course it works. However only one tunnel
is required, which is between R2 and R3. After tunnel interface is in area
0, no discontiguous area exist. Because R1 is connected to R2 and R4 is
connected to R3.
And, if there is a area router in the middle between R2 and R3, as this:
R1-R2-R5-R3-R4, there need to be two tunnels. one is between R2 and R5 ,the
other is betwee R5 and R3. (R2,R5,R3, are still in area 4). The only thing
left is to set the two tunnel interfaces in area 0.
2009/5/6 Jared Scrivener <jscrivener_at_ipexpert.com>
> Every router has a connected route between them. Forget the numbered areas
> - Would the 3 contiguous GRE tunnels create a contiguous Area 0? I can't see
> why not. From there it follows that every other router's numbered area
> interfaces could be advertised into OSPF without incident (as if they were
> loopbacks). This won't force the control plane path via the data plane path
> to build the control plane path (my description of tunnel recursion) as the
> numbered OSPF areas are going to be learned via OSPF with a Tunnel interface
> next hop (which is fine) and the GRE packets that carry the OSPF packets
> will be routed via the connected routes as each GRE packet is going over a
> single hop to a directly connected neighbor.
>
> A more difficult question (here's a gauntlet thrown down) is what would
> happen if there was an area router in the middle of each area (between R1
> and R2, R2 and R3 etc)? Please don't lab this up and paste the configs until
> everyone has a chance to visualize and discuss this (explain your reasoning
> from first principles). It's a great thought experiment so pick an answer
> THEN prove it on hardware (if you can visualize this, you can do anything -
> this kind if puzzle is as hard as it gets).
>
> Game on. :)
>
> Cheers,
>
> Jared
>
> PS If you see three numbered areas and can't create area zero you can
> redistribute each area into the adjacent one other by putting them in
> separate processes (ASBRs and no ABRs). No need for tunnels (the least
> preferable method). Redistributing between multiple processes should be your
> second preferred method (VLs between routers and adding a loopback in Area 0
> is method number one of course, if allowed).
>
>
>
> On May 5, 2009, at 10:28 PM, Anthony Sequeira <
> asequeira_at_internetworkexpert.com> wrote:
>
> I think Anantha is really going "outside the box" here as he has NO area 0
>> at all.
>>
>> It is really late for me, but I am thinking the only time we can go
>> without an Area 0 in OSPF is when we only have a single area.
>>
>> Warmest Regards,
>>
>> Anthony J. Sequeira, CCIE #15626
>> Senior CCIE Instructor
>> http://www.InternetworkExpert.com
>>
>> Internetwork Expert, Inc.
>> Toll Free: 877-224-8987
>> Outside US: 775-826-4344
>>
>> Test your Core Knowledge today!
>> Q: What is the BGP synchronization rule?
>> A: This rule states that a router will not use or advertise a route
>> learned by IBGP until a matching route has been learned from an IGP.
>> More Info:
>> www.cisco.com/en/US/docs/ios/12_2/ip/configuration/guide/1cfbgp.html#wp1002910
>>
>> On May 6, 2009, at 1:01 AM, Roy Waterman wrote:
>>
>> Sorry what i meant to say was:
>>>
>>> You can create a virtual link to extend area 0 to area 4, and then create
>>> a
>>> 2nd virtual link to extend area 0 (which was extended to area 4) to area
>>> 5.
>>>
>>> 2009/5/6 Roy Waterman <roy.waterman_at_gmail.com>
>>>
>>> Hi Anantha
>>>> You can create a virtual link to extend area 0 to area 4, and then
>>>> create a
>>>> 2nd virtual link to extend area4 to area 5.
>>>>
>>>>
>>>> 2009/5/6 Anantha Subramanian Natarajan <anantha.natarajan_at_gravitant.com
>>>> >
>>>>
>>>> Hi All,
>>>>
>>>>>
>>>>> Is it possible to run OSPF when there are three OSPF areas as shown
>>>>> below
>>>>> and none of them is not a area 0.
>>>>>
>>>>> R1 --ospf area3 --R2 ---ospf area4 ---R3 ----ospf area5 --R4
>>>>>
>>>>> My understanding is that, virtual-link would be only useful when to
>>>>> join a
>>>>> non-zero area to connect to area0 which is seperated by another
>>>>> non-zero area.Also the virtual links would be useful in case where area
>>>>> 0
>>>>> is
>>>>> broken in between a non-zero area.Assuming my above understanding is
>>>>> right,I
>>>>> am thinking the above scenario will not be able to use virtual links to
>>>>> propagate OSPF routes among R1,R2,R3,R4.
>>>>>
>>>>> In this case,if I would be able to build GRE tunnel between
>>>>> R1-R2,R2-R3,R3-R4 and run OSPF on area0 over them,In that case,will the
>>>>> OSPF
>>>>> routes would be able to get propagated ?
>>>>>
>>>>> Thanks for the help
>>>>>
>>>>> Regards
>>>>> Anantha Subramanian Natarajan
>>>>>
>>>>>
>>>>> Blogs and organic groups at http://www.ccie.net
>>>>>
>>>>> _______________________________________________________________________
>>>>>
>>>>
>
> Subscription information may be found at:
>>>>> http://www.groupstudy.com/list/CCIELab.html
>>>>>
>>>>>
>>>>>
>>>>>
>>>>>
>>>>>
>>>>>
>>>>>
>>>>>
>>>>
>>>> --
>>>> Regards
>>>> Roy
>>>>
>>>>
>>>
>>>
>>> --
>>> Regards
>>> Roy
>>>
>>>
>>> Blogs and organic groups at http://www.ccie.net
>>>
>>> _______________________________________________________________________
>>>
>>
>
> Subscription information may be found at:
>>> http://www.groupstudy.com/list/CCIELab.html
>>>
>>
>>
>> Blogs and organic groups at http://www.ccie.net
>>
>> _______________________________________________________________________
>>
>
>
> Subscription information may be found at:
>> http://www.groupstudy.com/list/CCIELab.html
>>
>
>
> Blogs and organic groups at http://www.ccie.net
>
> _______________________________________________________________________
> Subscription information may be found at:
> http://www.groupstudy.com/list/CCIELab.html
Blogs and organic groups at http://www.ccie.net
Received on Wed May 06 2009 - 16:23:20 ART
This archive was generated by hypermail 2.2.0 : Mon Jun 01 2009 - 07:04:42 ART