At 5:41 PM +0000 10/18/02, Priscilla Oppenheimer wrote:
>The Long and Winding Road wrote:
>>
>> OR - who / what is over rated?
>>
>> I revisited a couple of CCIE Lab preparation books last night.
>
>Are they really lab prep books? They may just be CCIE level because they
>cover advanced concepts not because they cover the nit-picky weirdness that
>you might run into in the lab.
>
>You may be one of the few people on the planet that understands all the
>things that could go wrong with redistribution! Could you write a white
>paper on it? Would you like to write a chapter for my next book? ;-)
>
>Priscilla
Part of the problem with redistribution is the terminology itself.
In the Routing Policy Specification Language (RPSL) for global
routing, there isn't a single "redistribute", but there are:
"import" and "export" statements, which can be filtered, for
transferring
information between routing processes in the same router or same AS
"accept" and "advertise" statements about what routes will be permitted
or denied by an AS, a router, or an interface. Again, lots of
filtering.
When I deal with a "redistribution" problem, even with IGPs, stating
the goals in RPSL makes things MUCH clearer. You can substitute IGP
tags for the community attributes used in exterior routing
specification.
"Redistribute" isn't even good English. To me, it's a transitive
verb, which implies that the routing process containing the
redistribute statement is distributing information to something else.
A better English version would be:
Redistribute|import FROM
INTO [UNDER constraints].
Unfortunately, redistribution deals with importing, not really with
exporting or advertising, and only incidentally to accepting.
I use this RPSL-based approach in my new Building Service Provider
Networks book, long before I talk about any specifics of BGP. I find
it clearer and more concise.
>
>
>> I sat down
>> with high hopes, after a frustrating afternoon at the rack. I
>> was looking
>> for some enlightenment.
>>
>> The first book I hit was one of the so called "CCIE Professional
>> Development" books. The chapter on PIM was a mere 28 pages, of
>> which a good
>> half were charts of various multicast headers. The chapter
>> mentions RP, but
>> never bothers to explain the concept of RP. CCIE level? More
>> like the "idle
>> curiousity" level.
>>
>> OK, so I go to one of the old standby's, one mentioned in every
>> "recommended
>> books" list, and get distracted by the chapter on
>> redistribution. As I read
>> through this chapter I am struck at how rudimentary it is.
>> There appears to
>> be no real insight here. Distribute-list in and out are glossed
>> over. Issues
>> with VLSM. to FLSM are covered in a a simplistic manner.
>>
>> So my question - for folks who have been to the CCIE Lab once
>> or twice, who
>> have put in the thousands of hours of reading and configuring,
>> what do you
>> think when you revisit the classics? Do you think maybe you've
>> outgrown
> > them?
Message Posted at:
http://www.groupstudy.com/form/read.php?f=7&i=55897&t=55897
--------------------------------------------------
FAQ, list archives, and subscription info: http://www.groupstudy.com/list/cisco.html
Report misconduct and Nondisclosure violations to [EMAIL PROTECTED]
