Path: utzoo!attcan!utgpu!jarvis.csri.toronto.edu!rutgers!tut.cis.ohio-state.edu!ucbvax!SATURN.ACC.COM!cam
From: cam@SATURN.ACC.COM (Chris Markle acc_gnsc)
Newsgroups: comp.protocols.tcp-ip.ibmpc
Subject: SNA over Ethernet via NDIS in OS/2 EE environment
Message-ID: <8905301958.AA27545@saturn.acc.com>
Date: 30 May 89 19:58:39 GMT
Sender: daemon@ucbvax.BERKELEY.EDU
Organization: The Internet
Lines: 32

Folks,

(Apologies to those of you who had to read this twice; 1st shot went to 
"ibmnets" instead of the correct "ibm-nets".)

A recent article in MIS Week describes an "agreement" between IBM and
Ungermann-Bass (UB), 3Com, and Western Digital (WD) to have the three
vendors comply with the NDIS (Network Driver Interface Specification) 
such that OS/2 EE Version 1.2 will support Ethernet adapters from those
vendors. The article goes on to mention how the OS/2 EE version of NDIS
allows SNA and NetBIOS to operate over an Ethernet LAN.

I was curious as to what "allows SNA to operate over an Ethernet LAN"
really means. Any thoughts?

One thing that bothers me is that SNA presumes that the data link is 
"reliable" (ie. SDLC, token ring using LLC type 2 operation, and the
370 channel). The "LLC" used with Ethernet is usually type 1 with no
link-level acknowledgments, flow control, or error recovery. For SNA
to operate over such a "LLC" would require the equivalent of a transport
protocol to be used between SNA and the Ethernet link-level. A type 2
"LLC" with reliable characteristics could be used over Ethernet and
then SNA run over that. Is this what people are doing or intend to do?

(I use "LLC" to represent "an" LLC versus "the" IEEE 802.2 LLC.)

Another question is how address resolution is (or will be) accomplished?

Send replies to me or to everywhere; it doesn't matter as I read both and
will summarize if I get anything worth summarizing.

Chris Markle - ACC