tomsu@cromemco.UUCP (Tom Cumming) (07/14/90)
I am confused. What is the difference between TCP/IP network
"stacks", and the ISO network "stack". Ok, ok, I understand the
differences between TCP/IP and ISO it'self. But I am confused in the
implementation.
The environment that I am have in mind is say, a bunch of unix
machines all talking to each other though some type of network.
It seems to me (from experiance) that the TCP/IP network "stack"
(in above environment) is usually implemented as a module of some
sort that is linked into the kernel, i.e., Application-Datalink
layers are all "right there". For example, some applicaton running
under X connected through sockets to TCP/IP, sitting on a datalink to
ethernet hardware, are all in one "box".
On the other hand, the ISO "stack", although having functionality
almost identical to TCP/IP is almost always implemented differently.
I almost always see ISO Layers 1-3 (Hardware - Network) implemented
in something called X.25.1-3, and then maybe TCP/IP hacked on top of
that. For example, some application running under X connected through
sockets to TCP/IP, sitting on a datalink to a special modem that is
connected to an X.25 "cloud". You are not *really* running ISO because
"layers" are really inplemented on the far side of the modem!
Why don't I *ever* see something like this (for example): Some
application running under X, connected through the ISO session,
transport, network, layers, with a datalink to a ethernet controler.
Why is'nt there a top to bottom *ISO* stack floating around like the
BSD TCP network stack (The one I'm most familier with)?
Thanx, Tom C.
Please E-Mail me!
'fer sure e-addr: uunet.uu.net!pyramid!cromemco!tom