>"Howard C. Berkowitz" wrote:
>>
>> The topic that will never die... :-)
>
>Right because I am getting more and more confused by the thread I
>(inadvertently) started ;-( The more I wanted to get away from OSI
>reference model, the closer we got.
>
>I thought we could do without (anyway TCP/IP is older than the '84
>endorsed OSI model standard) and could concentrate on TCP/IP layering
>concepts. Not that I do not like OSI, the contrary. On one side this
>shows the usefulness of (and reason why we all need to learn and
>understand properly) OSI RM.
Because TCP/IP was developed without serious consideration of
layering. People designed protocols (and not particularly services)
with respect to one another, to specific requirements, rather than to
an overall model. You'll find no discussion of layering in early
(and many current) RFCs.
Once ISO 7498 became widespread, there was an industry urge to apply
it to everything. I remember receiving a proposal from IBM to the
Corporation for Open Systems, the OSI research lab, which started,
"OSI and SNA are very similar. They both have seven layers. Let us
tell you about SNA."
Like it or not, formal layering is an OSI concept. It is an OSI
concept that gets far deeper into theory of data structures and
interprocess communications than most presentations dare.
First, I don't think it's necessarily a good idea for anyone to keep
asking what OSI layer a protocol fits into. Second, if you are going
to do that, than you certainly have to go beyond the definitions of
ISO 7498 without appendices. Third, you have to be willing to deal
with the original OSIRM distinctions between services and protocols,
and be clear what you mean in a given discussion. Fourth, you should
consider at least the management annex to 7498 (distinguishing
between layer and system management), and seriously consider using
the U/C/M plane concept introduced in the B-ISDN reference model.
>
>So now I am confused on both fronts:
>
>"Howard C. Berkowitz" wrote:
>
>>
>> Again, we are getting into a situation where there is a desire to
>> coerce things into a simplified version of the OSI model. Real OSI
>> documents are very careful about the protocol versus service
>> definition, and indeed you will find separate documents, say, for the
>> transport service and the (several) transport services. Doing things
>> this way completely sidesteps the "what layer is this" problem.
>>
>
>I always thought the the layer is defined by its service(s) provided to
>its upper layer. These services are reflected in the functions needed to
>be performed within the layer. And protocols residing (operating) at the
>layer should "fulfill" the function(s).
>
>Overall, where the number of protocols operate at a certain layer, they
>altogether should cover all the functions at the layer (with some
>potential "redundancy" - where alternative protocols exist, well, not
>quite, the alternative protocols actually seem to perform subset or
>superset of layer functions - like UDP or TCP, or TP0-TP4 of OSI
>architecture).
>
>What is wrong in my understanding (and need to see the structure) which
>somehow tortures me to fit ARP a protocol in a particular layer of the
>protocol architecture it was developed for?
>
>Rita
>
>P.S. Sorry for those feeling this discussion redundant. My horoscope for
>today says it all:
>"You may not be the star performer at the moment, but there's great
>delight in playing second fiddle. Without your part, the arrangement
>would sound empty. Small contributions yield big rewards. " ?-)
Message Posted at:
http://www.groupstudy.com/form/read.php?f=7&i=8701&t=8335
--------------------------------------------------
FAQ, list archives, and subscription info: http://www.groupstudy.com/list/cisco.html
Report misconduct and Nondisclosure violations to [EMAIL PROTECTED]
