RE: 3 OSPF non-zero areas

Thanks, Scott. One of my mentors once told me to "think like a router" and
it was the best advice anyone ever gave me with respect to the CCIE.

So, guys, I haven't seen any responses yet to my challenge question that
didn't involve changing the question (I didn't say you could add any more
tunnels). Back in the day (well, in the mid-late '90's when I was in high
school and writing under my 31337 identity - the good old days) it was
normal for highly technical engineers to post challenge questions to test
each other - in my opinion that's one of the reasons that that the old
"hackers" like Steve Wozniak got us to where we are now technologically.

If I get some activity out of this I'd like us to consider doing more
"thought experiments" here on the list - you'll be amazed how much better we
all become. Oh, I didn't say that I'd get the answer right either - so don't
be afraid to be wrong. J

Cheers,

Jared Scrivener CCIE3 #16983 (R&S, Security, SP), CISSP

Sr. Technical Instructor - IPexpert, Inc.

Telephone: +1.810.326.1444

Fax: +1.810.454.0130

Mailto: <mailto:jscrivener_at_ipexpert.com> jscrivener_at_ipexpert.com

From: Scott Morris [mailto:smorris_at_internetworkexpert.com]
Sent: Wednesday, 6 May 2009 3:49 AM
To: Jared Scrivener
Cc: Anthony Sequeira; Roy Waterman; Anantha Subramanian Natarajan; Cisco
certification
Subject: Re: 3 OSPF non-zero areas

Jared's actually right on target with this method of thinking... As a cool
way of studying (even though there's a plethora of variety between all the
vendors!) a candidate should be able to take any existing scenario and start
thinking about "what if....?".

Never fear to create your own modifications to things, or even your own labs
for that matter. For years, I've been telling people to work towards this.
The cool thing (as seen below) is that you'll end up creating some REALLY
ugly scenarios. But if you can think your way through them, you're doing
great on your comfort level with the technology!

Don't fear making mistakes! Every once and a while, you will create/dream
an impossible scenario. Something that simply cannot happen. By the time
you've worked at it, thought about it, labbed at it and tried to make it
work, you have at least been spending time thinking about the protocol. No
value is lost there.

Any complicated task before you can be broken down into a series of simple
tasks. That level of visualization though, requires a lot of work! :)

Scott Morris, CCIEx4 (R&S/ISP-Dial/Security/Service Provider) #4713,

JNCIE-M #153, JNCIS-ER, CISSP, et al.

JNCI-M, JNCI-ER

smorris_at_internetworkexpert.com

Internetwork Expert, Inc.

http://www.InternetworkExpert.com

Toll Free: 877-224-8987

Outside US: 775-826-4344

Knowledge is power.

Power corrupts.

Study hard and be Eeeeviiiil......

Jared Scrivener wrote:

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
<mailto:asequeira_at_internetworkexpert.com> <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#wp10029
10

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 <mailto:roy.waterman_at_gmail.com>
<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
<mailto:anantha.natarajan_at_gravitant.com> <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
Received on Wed May 06 2009 - 10:00:38 ART