>In the spirit of open debate, and just to be contrary, may I hold out an
>alternative viewpoint.
>
>We have all had our fun with test questions. Just recently we had an
>analysis of an OSPF related question, a very interesting thread, and one
>well worth checking out in the archives if you missed it.
>
>What can be interpreted as ambiguous and poor and obscure to a rank amateur
>may be crystal clear to an experienced professional.
If it's the same thread I'm thinking of, it's not insignificant that
Priscilla, Peter, Pamela and myself came up with slightly different
readings of the questions (I think there were other experienced
professionals answering, but I seem to remember only the P's).
Unfortunately, terminology can be rather vague. Unfortunately as
well, to be precise tends to be based in rather formal computer
science methodology. For example, to answer the question if a given
protocol belongs to some OSI layer numbered N, the precise answers
include:
If the (N)-protocol provides a (N)-service to a (N+1) client
layer entity and carries (N+1) service data units to a peer (N+1)
entity via (N)protocol data units, the protocol belongs to layer (N),
If the (N)-protocol does not provide any (N)-service to a layer
(N+1) but does affect the behavior of other (N)-entities by sending
(N)-PDU's with protocol control content, the (N)-protocol is a
layer(N) management protocol.
Let me merely say that one day, while shaving, I realized I
understood EXACTLY what the OSI definition of the presentation
service meant, and nearly cut my throat in shock.
It has been said that the physical universe is merely a metaphor for
that which can be defined in the full OSI reference model. It also
has been said that an elephant is a mouse designed for full OSI
conformance.
>Being a rank amateur, I
>can't come up with a good example. But over the course of time there have
>been a number of these discussions here. I read through these kinds of
>threads religiously, and in the end I learn two things - thought process,
>and protocol behaviour.
>
>Over time I have come to realize that Cisco isn't interested in turning out
>certified individuals who can plug a few routers together and get it to
>work. Cisco is interested in turning out certified individuals who
>thoroughly understand protocol behaviour, and thus can add value in any
>situation where they must do complex analysis. Protocol behaviour includes
>routing protocols, redistribution among protocols, behaviour at each so
>called OSI layer, protocol transmission responsibility, and so on.
>
>Yes, X.25 is pretty much gone from the U.S Landscape, but worldwide it is
>still alive and well. Yes there are new technologies such as cable modem and
>DSL, but the fact is that without a foundation in protocol behaviour over
>other WAN technologies, how can one understand the issues involved with
>newer technologies, which in general build upon these earlier ones.
Absolutely. X.25 remains quite appropriate for certain
high-error-rate, high-delay, or low-speed links. By X.25, I do
include the packet level even on point-to-point links, because, if
for no other reason, it can give you more fragmentation than can IP
over LAP-B, which will optimize throughput on some bad links. LAP-B
also is the basis for SSCOP, the high-reliability signaling link
protocol in ATM and SS#7.
_________________________________
FAQ, list archives, and subscription info: http://www.groupstudy.com/list/cisco.html
Report misconduct and Nondisclosure violations to [EMAIL PROTECTED]
