From: Howard C. Berkowitz (hcb@gettcomm.com)
Date: Thu Nov 21 2002 - 17:24:49 GMT-3
>Howard,
>
>     True, the current R&S lab barely scratches the surface of BGP and it's
>current accepted techniques.  But with only 8 hours, it's tough to squeeze
>much in about it.  I'll admit I've never worked for an ISP, and would
>probably flounder my first couple weeks at an ISP position.  But I'm sure
>many others on this list that have never worked for an ISP are in the same
>position.  But other subjects on the lab are similar.  DLSW, IPX?  Those
>are/were pretty lightly covered in the labs, compared to what real-world
>companies do.  I guess if you want something in writing that says you can
>handle an ISP position, the Juniper certification is where it's at.  That's
>just what I've heard though...
>
>Thanks,
>
>Chuck Church
>CCIE #8776, MCNE, MCSE
Thanks, Chuck, and I'm really not trying to divert the thread too 
much off topic.  From my personal experience, the light didn't dawn 
for me on BGP until I started studying routing policy (RPSL, the 
routing arbiter), etc., and finally figured out the _purpose_ of most 
of those knobs.
Even more than DLSW, in CID, IBM protocols were always a joke, at 
least from my standpoint as having been an NCP/VTAM system 
programmer.  There is so much to know on the IBM side that I couldn't 
see just studying the Cisco part being terribly useful.
I'm still a believer in narrowing the scope and deepening the depth 
of the CCIE, using an awfully strange model -- the team specialties 
in the TAC.  Even if that's not done, knowing the context and 
motivation of why certain protocols were developed, I think, makes it 
much easier to understand what they are doing. Knowing something 
about queueing theory, at a not terribly mathematical level, gives a 
lot of insight into QoS and performance. At least scanning through a 
computer science textbook about memory management and interrupt 
handling helps a great deal with switching paths, although the Inside 
IOS Architecture is superb.
This archive was generated by hypermail 2.1.4 : Tue Dec 03 2002 - 07:23:08 GMT-3