From: Howard C. Berkowitz (hcb@gettcomm.com)
Date: Sat Feb 22 2003 - 18:56:17 GMT-3
At 10:35 AM -0500 2/22/03, Aaron Woody wrote:
In a later message, I believe you unsubscribed to this list. I'm not
necessarily going to say that was a bad idea, depending on your
expectations. Perhaps the issue is that we need to have additional
lists, with different emphases, and as much moderation and peer
pressure as is feasible to stay on-topic.
Some time ago, Paul Borghese designated the original Groupstudy list
as open to networking questions of all types. That does make for a
wide range of ideas.
In my mind, there is an unquestionable need for a list of
CCIE-lab-oriented material. There may very well be needs for other
lists. Let me elaborate.
Take almost any field that one is learning. Pro football players
still do one-on-one drills that will never happen that way in a game.
Ballet dancers, fencers, and many martial artists have very stylized
practice routines that again don't mirror what they do "in
production."
>
>
>I think the question about the OSPF 400+ was valid. There is a gray area
>around this design and I wanted others opinion who had actually used OSPF in
>this situation. This is a group of CCIE's and wannabes right?
I'd have to disagree with you, but perhaps suggest an alternative.
This is not a list of CCIE's and wannabes. It is a list of CCIE
candidates, and of passed CCIE and other individually approved
individuals (by Paul) who want to deal with the stylized drill that
is the CCIE lab.
The approximate equipment list of the CCIE lab is well known, as is
the approximate number of devices. It doesn't break NDA to say that
the CCIE lab does not and will not contain 400 routers. So, I'd have
to say that the OSPF 400+ is not relevant to the lab list.
Is it relevant to the main list? Yes and no. Yes, in that there are
no real constraints on that list, but no, if you want a focused
answer.
>However, the 1
>million replies was the issue. The design questions I had were answered
>within first 5 emails, but everyone continue to come up with answer not
>relevant to my original criterion.
Here I have to take you to task. If you want a quick, specific,
qualified answer to a real-world design question, perhaps you should
contract with a consultant to do an analysis and design. Personally,
and this is NOT a solicitation for business, I've designed networks
with literally thousands of OSPF routers, and have worked on both
OSPF standards and OSPF code in router development.
Now, should there be a design and/or protocol "theory[1]" list?
Perhaps, especially if its first focus is topics that might be
encountered in the CID or CCIE Written tests. Your question is not
inappropriate to CID, although CID actually tends to deal with
smaller networks.
>If you actually read the email I sent it
>indicated concerns beyond basic OSPF configuration. Also, look at the OSPF
>without area 0 thread...his original question was lost very quickly because
>everyone wants to come up with genius answers relevant or not.
What do you mean by a "genius" answer and who judges relevance?
>I am not so
>bent on asking stupid questions about the lab, I am concerned with operation
>of Protocols, etc. that will be in the lab. Most of this stuff can be
>answered definitively reading a Doyle book or CCO.
I really don't understand your point here. Why stop with Doyle and
CCO? Why not join the IETF lists and keep up with the Internet
Drafts, but, even more, the discussion on the list of why things work
and don't work--getting into true protocol theory, not protocol
principles of operation. There are good pure theoretical discussion
(leading to protocol design) on such things as millisecond-range
convergence time in OSPF, OSPF and traffic engineering, OSPF rapid
recovery, and use of OSPF with additional constraints as required in
optical routing (e.g., preservation of the wavelength continuity
property).
>
>I thought you may need some clarification.
>
>Aaron...the guy who asked the "cluttering-up" design caveat question about
>OSPF hub 400+
To quote Tom Larus, repeated answering of questions, NOT "just what
you need in the first 5 responses", CCIE lab preparation CAN include
answering questions for practice:
At 12:14 PM -0500 2/22/03, Tom Larus wrote:
>
>At the end of a long day (at work or in the lab), it is nice to come home
>and relax by reading this forum. Sometimes, someone comes up with a new
>twist on an old solution to an oft-asked question. It is nice to see, again
>and again, the same or similar answers to oft-asked questions, because it
>provides a good review and tends to confirm that the answer is indeed
>correct. If it was not correct, someone in the group would proudly offer a
>correction. Reading the list was a way to relax while I prepared for the
>Lab. I generally did not count it as Lab prep time in my lab notebook, yet
>I was learning even as I relaxed.
>
>Also, I like having the chance to take a stab at answering simple questions
>or questions which may have been answered before. When I post an answer,
>like I did the other day, I can then receive some correction about a point
>of syntax. ( I'm rebuilding my lab now so that I can check syntax at home).
>A trainer from a first-rate CCIE training outfit and I each made the same
>mistake in our response to a relatively easy question-- we each put "ip" in
>a standard ip access-list-- and our error was noted. All of this back and
>forth contribution and correction probably looked like a waste of time to
>someone who thought that the original question never should have been asked.
At 1:52 PM -0500 2/22/03, Chuck Church wrote:
>I agree. While the designated purpose of this list is to help people pass
>the CCIE lab, some of these off-topic discussions like the large OSPF design
>are good. Do you just want to pass the lab, or do you want to be a highly
>knowledgeable networking professional?
I think those are really two separate problems, and may need two
separate lists. As you point out, the techniques for passing the lab
may have very little to do with real-world best current practices.
I've designed some extremely complex networks, enterprise and ISP,
and it's been extremely rare to have the kind of weird mutual
redistribution problems apparently beloved of the lab authors.
Rather than mutual redistribution, the real world tends to use
hierarchical redistribution, BGP backbones-of-backbones, and static
and default routes.
>I prefer the latter, and I hope most
>of you do as well. The CCIE program does NOT teach you how to design a well
>thought-out network. Knowing how to implement BVIs and virtual links in a 5
>router network is good for the lab, but designing a production network like
>that will have the people in TAC making fun of you. I think to become a
>CCIE, you've got to have a desire to learn as much as you can about
>networking.
No argument, to be an employable and highly paid CCIE or equivalent.
>If you just want to learn the minimum, here's a better place to
>get certified: www.microsoft.com . The nice part about email is it has a
>subject. If I grow tired of a thread, I'll sort by subject and delete those
>right off the bat. Keep in mind, this list caters to many people with a
>wide range of skills and backgrounds. It's going to get off track a little
>at times. But would I ever unsubscribe? Never!
This archive was generated by hypermail 2.1.4 : Sat Mar 01 2003 - 11:06:29 GMT-3