[e2e] evolution of bandwidth as a term
cannara at attglobal.net
Fri Oct 3 10:16:15 PDT 2003
There's a difference between one word meaning different things in different
contexts (where there is no alternate), and using a word from another context
instead of word(s)already available for discourse. The latter is unnecessary
at the very least.
In fact, the example of "transport" is peculiar to no particular networking
level. We may think of the OSI model's definition of Transport Layer, and if
we use "Transport", with a big T, then maybe it's clear we're talking of that
layer's activities, in any of several protocols. But discussions using
"transport" in many specific communications paths may be legitimate, because
they perform the functions of a "transport" -- providing reliability, etc.
Terms like "link" are even more general, from common language, unlike
"bandwidth", so their multiple, sloppy uses are even less relevant to the
basic question of allowing misuse of well-defined scientific terms.
"James P.G. Sterbenz" wrote:
> At 08:02 PM 10/2/2003 -0700, Joe Touch wrote:
> >Just in case the notion of context is lost, this is the end2end-interest
> >list, where bandwidth isn't liable to be interpreted as frequency deltas.
> >Here are a few other terms with similar levels of confusion elsewhere, but
> >which appear to have commonly-accepted definitions on this list:
> > transport on this list means an end-to-end
> > protocol adding value beyond endpoint
> > delivery,
> > rather than what ATP helps electrons do,
> > or what glass does with photons
> And, in particular, there are a *lot* of folk to which transport is layer
> 2 or 3 (OTN = optical transport network); I'm currently working on something
> where I have to be very carful every time I use the word transport because
> sometimes it is end-to-end (L4), sometimes edge-to-edge (L3), and sometimes
> a lightpath (L2, sort of).
> > link on this list means a physical connection
> > between one or more hosts or routers,
> And in the same context a link between IP routers (switched lightpath) is not
> necessarily the same as a glass fiber link between photonic cross-connects.
More information about the end2end-interest