From: Howard C. Berkowitz (hcb@xxxxxxxxxxxx)
Date: Fri Apr 19 2002 - 18:53:27 GMT-3
At 4:42 PM -0500 4/19/02, Frank Jimenez wrote:
>Re: ISIS and curses...
>
>Not surprising. When you name a protocol
>after an Egyptian goddess, things like
>that can happen.
And you wonder why people have trouble with switches whose basic
command is Set?
>
> If you ever come up with a new routing
>protocol, my suggestion would be to name
>it OSIRIS, and you may dispel the curse :-)
My assistant looks up from his cat food to comment it should be BAST.
>
>
>My own learning style seems to work best
>with a combination of styles. I learn more
>doing the small 2 or 4 router labs and then
>modifying the initial conditions to create
>more interesting problems.
Interesting. As you may know, I've been working (along with Alvaro
Retana of Cisco and others) on BGP benchmarking. Our specific goal is
dealing with single devices, but those are the devices being measured
-- we need multiple routers or router simulators to generate and
receive forwarded updates.
There was a very interesting comment on the BMWG mailing list:
At 9:42 AM -0800 3/18/02, David Newman wrote:
>Agreed. In fact, it could be argued that four nodes is the minimum necessary
>to exercise a routing protocol:
>
>--one node is an island
>--two nodes and a cable cut is two nodes off the air
>--three nodes and a cable cut is a dumb, automatic form of routing, since
>there is only one alternate path
>--four nodes and a cable cut forces arbitration, so now we are beginning to
>do something like routing
There's been a lot of emphasis on six router labs (typically with
additional external routers) because that appears to be the lab
environment. I know several commercial scenario writers who believe
in even more routers to add complexity.
But there may be good reason for people to create 4 router labs for
technology learning, unless you are dealing with something like
hierarchical routing. Well-developed 4 router labs, whether personal
or virtual, certainly would be cheaper.
>I've found the
>labs at www.fatkid.com to be an good resource
>for this, along with the examples at
>www.cisco.com on the TAC pages.
>
>Then to complete the circle, I'd love to
>see a set of large scale labs that build
>everything together. Maybe a single large
>lab with a play-by-play on how to build it
>- then several others that the student
>would have to put together by themselves.
>
>Frank Jimenez, CCIE #5738
>franjime@cisco.com
>
>Disclaimer:
>These are my own personal opinions and
>not necessarily those of Cisco Systems.
Thanks much.
Howard
(I think my opinions are the opinions of Gettlabs, most of the time,
except when they cost too much money.)
This archive was generated by hypermail 2.1.4 : Thu Jun 13 2002 - 10:58:14 GMT-3