[ih] Protocol layering

[Scott W Brim]

On Sat, Dec 03, 2005 10:52:10AM -0500, Phill Gross allegedly wrote:
> I've queried a number of experienced telecom folks and the consensus
> appears to be 1) the 3-plane model was not original intended to be
> applied to layered protocols, although it's been adopted later for
> that use, 2) they are not able to cite sources for standard
> definitions of how to apply it to layered protocols, but more
> interestingly, 3) they couldn't point to the original standard that
> defined it for telecom(!).  

Phill, I can't help on the history, but if you are looking for modern
discussion of planes-cum-layers, work in SG15 Question 12 might be
useful.  However, that's about the G.805/G.809 concept of layer
networks, not X.200 layers.  By the way, in SG13 some of us are
collaborating on a document about principles of interworking various
combinations of planes, layers networks and modes.  

> I've found a number of documents (eg, BISDN) that use some flavor of
> the 3-plane model, but most seemed derived and often modified (eg,
> adding extra planes).  I'm still tracing references.  If I find the
> headwaters, I'll report back.

The "plane" concept has been transmogrified many times, and often
conflates with "layer".  We now have a "services plane" and a number
of "planes" in security architecture that have nothing to do with
early concepts.

> Perhaps the 3-plane model is like that other widespread concept that
> defies clear definition, and like Potter Stewart, we just know it
> when we see it.  : )

I'll be very interested in what you find.

swb