[comp.sys.amiga] Networking

peter@sugar.hackercorp.com (Peter da Silva) (10/07/90)

I'm not sure what you're trying to say here, but OSI based and DDN based
networking are completely incompatible. They run on the same hardware (at
the same time, even) but cannot talk to each other at any level.

OSI is a connection based protocol developed in Europe. It is relatively
new, and there is damned little support for anything using the OSI protocol
stack. We use a network that runs the OSI stack on Ethernet... it's pretty
nice, but isolated. By dint of running two sets of networking software we
are able to communicate with the real world of...

DDN is a connectionless protocol developed by the DoD. It's usually referred
to as TCP/IP because the basic packet switching and reliable comminucation
protocol go by those acronyms. It was developed ad hoc over many years and
is now quite reliable and widely available. The Ethernet software available
for the Amiga talks DDN...

...which is why it would not be able to run on an OSI network. No great loss:
OSI is not terribly important these days.
-- 
Peter da Silva.   `-_-'
<peter@sugar.hackercorp.com>.

davidm@uunet.UU.NET (David S. Masterson) (10/08/90)

In article <6722@sugar.hackercorp.com> peter@sugar.hackercorp.com
(Peter da Silva) writes:

   ...which is why it would not be able to run on an OSI network. No great
   loss:  OSI is not terribly important these days.

But didn't DoD express interest in dropping TCP/IP in favor of OSI on the idea
that the TCP/IP protocol had gone just about as far as it could whereas the
OSI protocol would have more "extensive" features (not sure what, though)?

I also thought, but am not sure, that the lower levels of both protocols had
the ability to interconnect.

--
====================================================================
David Masterson					Consilium, Inc.
uunet!cimshop!davidm				Mtn. View, CA  94043
====================================================================
"If someone thinks they know what I said, then I didn't say it!"

joseph@valnet.UUCP (Joseph P. Hillenburg) (10/08/90)

What does DoD stand for?

-Joseph Hillenburg

UUCP: ...iuvax!valnet!joseph
ARPA: valnet!joseph@iuvax.cs.indiana.edu
INET: joseph@valnet.UUCP

new@ee.udel.edu (Darren New) (10/09/90)

In article <CIMSHOP!DAVIDM.90Oct7211919@uunet.UU.NET> cimshop!davidm@uunet.UU.NET (David S. Masterson) writes:
>I also thought, but am not sure, that the lower levels of both protocols had
>the ability to interconnect.

Well, you can run the higher layers of OSI over the lower layers of
TCP/IP, (and probably vica versa, but I haven't seen that since TCP/IP
is much more common in the States than OSI), but that doesn't make them
"talk" to each other.  Things like FTP and Telnet are just plain
different in OSI (much more flexible, but with the price of efficiency,
of course).         -- Darren
-- 
--- Darren New --- Grad Student --- CIS --- Univ. of Delaware ---
----- Network Protocols, Graphics, Programming Languages, 
      Formal Description Techniques (esp. Estelle), Coffee -----

UH2@psuvm.psu.edu (Lee Sailer) (10/09/90)

>But didn't DoD express interest in dropping TCP/IP in favor of OSI on the idea
                                    ^^^^^^^^^

Well, yes, except that "dropping" is far too strong a word.  Some speculate
that the DoD will spearhead the gradual conversion to OSI.  One interesting
wrinkle is that perhaps OSI is all ready too slow, and so will get passed
over for something faster, soon.

                                lee

jnmoyne@lbl.gov (Jean-Noel MOYNE) (10/09/90)

In article <6722@sugar.hackercorp.com> peter@sugar.hackercorp.com (Peter 
da Silva) writes:
> ...which is why it would not be able to run on an OSI network. No great 
loss:
> OSI is not terribly important these days.


     I'm sorry but I don't agree !  OSI is not dead ... OSI is starting, 
and if TCP/IP rules for the moment, wait a little, and you'll see. OSI is 
really the big deal now, for a few reasons: first, you now have 
implementations of OSI going out, and not lousy ones. Just take Digital's 
example. DECNET phase V will be released soon (6 months I belive) on both 
Ultrix and VMS. And DECNET phase V is OSI ! DEC invested a great deal in 
OSI networking, and so are a lot of big names, like cisco (and a lot of 
others). There is a another point why OSI is going to be great: GOSIP, in 
clear the governement people are now obligated to buy and to use only OSI 
stuff (this is the theorical part, the practical is: they are beginning to 
move to OSI now).

     TCP/IP is the most important suite for the moment, the Amiga has it 
now, that's great it can speak to all the other people. Now, the next move 
is toward OSI, and the Amiga should have at least one OSI implementation, 
you will soon see OSI networks popping up, and the Amiga shouldn't close 
the OSI door.

     JNM

--
I write (badly) only my own opinions.
" Just make it ! "
Bo, in 'Bo knows UNIX'

david@twg.com (David S. Herron) (10/10/90)

In article <CIMSHOP!DAVIDM.90Oct7211919@uunet.UU.NET> cimshop!davidm@uunet.UU.NET (David S. Masterson) writes:
>In article <6722@sugar.hackercorp.com> peter@sugar.hackercorp.com
>(Peter da Silva) writes:
>
>   ...which is why it would not be able to run on an OSI network. No great
>   loss:  OSI is not terribly important these days.
>
>But didn't DoD express interest in dropping TCP/IP in favor of OSI on the idea
>that the TCP/IP protocol had gone just about as far as it could whereas the
>OSI protocol would have more "extensive" features (not sure what, though)?
>
>I also thought, but am not sure, that the lower levels of both protocols had
>the ability to interconnect.

Well.. since I happen to be in the group of people developing The Wollongong
Group's OSI software I suppose that means I know something about the
lay of this particular stretch of the swamps.

The DoD did at one time decide that "TP4" would replace "IP" at
some future time.  TP4 is one of the transport protocols in the ISO suite
and is the one which most closely matches IP (connectionless, able
to lose packets, sequenced, etc).

The US Government wants to have all their networking capability moved
over to the ISO suite.

This is what is providing the oomph to the ISO developments in the USA.
In Europe and other places, the phone companies normally tightly control
all telecommunications.  Part of that control is to specify that only
the ISO suite be used for networking, at least for long distance networking.

The lower levels of the protocols are quite different.  At the lowest
levels, however, most physical layers are able to accomodate many
protocol families at once.  Ethernet, for instance, has a "type" field
in the packet header as does PPP (the replacement for SLIP).

Some of the applications are able to be used with either TCP or OSI
transport protocols.  The mailer we sell, for instance, (WIN/MHS) can
do the X.400 MHS-MHS protocols using either TCP or OSI protocol stacks.


-- 
<- David Herron, an MMDF & WIN/MHS guy, <david@twg.com>
<- Formerly: David Herron -- NonResident E-Mail Hack <david@ms.uky.edu>
<-
<- Remember:  On System V it's "tar xovf", not "tar xvf"!

spear@locus.com (Brad Spear) (10/11/90)

In article (murfle) david@twg.com (David S. Herron) wrote:
>Well.. since I happen to be in the group of people developing The Wollongong
>Group's OSI software I suppose that means I know something about the
>lay of this particular stretch of the swamps.
>
>The DoD did at one time decide that "TP4" would replace "IP" at
>some future time.  TP4 is one of the transport protocols in the ISO suite
>and is the one which most closely matches IP (connectionless, able
>to lose packets, sequenced, etc).

Since you are working on OSI, I'm sure this is just a typo -- you meant to
say TP0, not TP4, didn't you?  TP4 is roughly equivalent to TCP, while TP0 is
roughly equivalent to UDP.  The intermediates, TP1 through TP3 provide
intermediate service levels that don't have direct counterparts.

Also, the OSI equivalent to IP is also called IP.  This layer deals with
network routing issues.  TP0 is just a very thin layer on top of this, much
as UDP is a very thin layer on top of the Internet IP.

Brad
-- 
Brad Spear				Locus Computing Corporation
spear@locus.com				Inglewood, CA, USA
Disclaimer: I speak for no one but myself.
This line intentionally left blank.

david@twg.com (David S. Herron) (10/13/90)

In article <18059@oolong.la.locus.com> spear@locus.com (Brad Spear) writes:
>In article (murfle) david@twg.com (David S. Herron) wrote:
>>The DoD did at one time decide that "TP4" would replace "IP" at
>>some future time.  TP4 is one of the transport protocols in the ISO suite
>>and is the one which most closely matches IP (connectionless, able
>>to lose packets, sequenced, etc).
>
>Since you are working on OSI, I'm sure this is just a typo -- you meant to
>say TP0, not TP4, didn't you?  

Ah.. more than likely -- I spend all of my time at a slightly higher
level than the transport protocols.  I'm (helping) doing the X.400
mailer -- especially user-interface stuff (my next project is the
User Agent).  Closest I get to TP's 0-4 is in telling the MHS to
use different transports to reach different places.  I knew (later)
that I shoulda pulled down my copy of _The_Open_Book_ and checked
my facts.  :-)

Now that I've reviewed my facts .. The DoD selected TP4 because it
most closely resembled TCP, not IP.  Big difference there!

There's a very good description in _The_Open_Book_ (M.T.Rose, author
of ISODE) on pages 97-100 of TP's 0-4.  The density of his description
is more from the subject, ISO protocol specs are *very* dense y'see,
than anything else.

sigh..
-- 
<- David Herron, an MMDF & WIN/MHS guy, <david@twg.com>
<- Formerly: David Herron -- NonResident E-Mail Hack <david@ms.uky.edu>
<-
<- Remember:  On System V it's "tar xovf", not "tar xvf"!