jmr@motown.UUCP (John M. Ritter) (12/12/85)
[ *** ]
Many thanks to all who responded to my request for IBM <<--->> Communications
information. Several asked if I would post the results of my inquiry, and
here's what I found:
Most everyone mentioned kermit for file transfer. Next in line was UREP from
Penn State, followed by some kind of protocol converter. Breakdown is as
follows:
==============================================================================
Kermit seems to be available on the Arpanet from Columbia University cu20b by
anonymous ftp. The contact person I have listed is:
Frank da Cruz
Columbia University Center for Computing Activities
New York, New York 10027
Stolen from a similar request came this additional information, thanks to
Dennis Bednar, {decvax,ihnp4,harpo,allegra}!seismo!rlgvax!dennis
Kermit Protocol Manual, 5th Edition, April 5, 1985, by Frank Cruz.
A manual for people who must implement the kermit protocol,
not intended for the casual reader.
"Kermit", PC Tech Journal, January 1985, by Augie Hansen.
A nice, easy to read, discussion of the Kermit protocol.
Kermit is compared to the 7-layer OSI Reference Model. Im-
plementations and user experience are also briefly mentioned.
UNIX Kermit, DRAFT, March 1985, Frank da Cruz, et al.
A must, but somewhat painful, reading. It is a user's manual
for C-kermit, a version of kermit written in the C program-
ming language for UNIX systems. The command line options are
discussed in detail, and would only be useful for knowing how
to invoke remote kermits, assuming that it is the UNIX C-
kermit version. Also, the interactive commands are discussed
in detail. The interactive commands section is a must read-
ing, since it can be very useful for defining additional
characteristics of the physical interface, but which are not
settable from OFFICEPOWER.
"Kermit: A File-Transfer Protocol For Universities, Part 1: Design
Considerations and Specifications", Byte, June 1984, by Frank da
Cruz and Bill Catchings.
An excellent article that discusses the motivation for the
development of the protocol: the need for cheap but reliable
file transfers between mainframes and micros at a university
campus. An excellent case study of what kinds of things can
go wrong when you try to design a general purpose point-to-
point data link protocol that must work between a multitude
of different computers, each with their own incompatible nu-
ances.
"Kermit: A File-Transfer Protocol For Universities, Part 2: States
and Transitions, Heuristic Rules, and Examples", Byte, July 1984,
by Frank da Cruz and Bill Catchings.
The internals of packet formats, and the state transitions
used for reliable file transfers.
Kermit User Guide, 5th edition, March 2 1984, Frank da Cruz,
editor.
Sections 1-4 serve as an excellent tutorial on how to use
kermit, independent of the version. Section 9 is UNIX-
specific, but is currently out of date, since the options
flags are passed differently.
==============================================================================
UREP is Unix Rscs Emulation Program, written at Penn State University, written
by Dr. Robert Michael Owens, and is constantly being upgraded. UREP was written
for 4.2BSD and is available for $225 from:
The Pennsylvania State University
Computer Science Department
Whitmore Laboratory
University Park, PA 16802
(814) 865-9505
For those System V users, you (we!) haven't been left behind. Mike Hrybyk at
Johns Hopkins has written a vpm driver for UREP and ported it to SysVr2.
Naturally, you must purchase UREP from Penn State first. He can be reached at:
Mike Hrybyk
Sr. Systems Programmer
Johns Hopkins Homewood Computing Facilities
(301) 338-8096
ihnp4!whuxcc!jhunix!mwh or root
==============================================================================
Protocol converters were the next thing. I'm not too sure how they'd be for
file transfers. Black Box (of course) sells some which would allow a logon to
an IBM system. Also available is a program which, using a protocal converter,
makes data from the UNIX system look like a high-speed terminal entry. Perhaps
one of the quickest fixes. Contact:
G. David Butler
{motown,sys1}!ninja!gdb
for more information.
==============================================================================
Finally, Barry Shein, Boston University writes:
> Do you have anything resembling a simple serial line? (eg 3705 style
> half-duplex ASCII [ascii, 1/2 dup not important but that's how they
> usually come.])
>
> If you do I have some spooler software I used a while back to send
> to our 3081. It would require some work (esp on the IBM side as I
> used C on our 3081) but the protocol, esp on the IBM, is *very*
> simple. The transfers look a lot like a line printer spooler to the
> UNIX users (job gets spooled, daemon wakes up, they can get mail
> back when it completes, will keep trying to get it across silently
> in the background.) It was mostly for transferring *to* the IBM
> (started out as a hack to get to their 8700 laser printer and grew
> a few features for file transfer and remote batch as that turned out
> to be trivial, mostly just punch to different spools on the IBM end,
> maybe insert a little JCL for the user, minor.)
>
> At any rate, it would take a bit of work to pull it together for you
> so please feel free to get back to me if you need more info to figure
> out if it is really useful (it very well may not be, and definitely would
> require quite a bit of work.)
Barry may be reached at:
Barry Shein
Boston University
buit4!bu-cs!bzs
==============================================================================
Again, many sincere thanks for all who responded. I know this has helped me,
and I hope it may help others.
------------------------------------------------------------------------------
"I enjoy working with humans, and have John M. Ritter
stimulating relationships with them." Allied-Signal Corporation
- HAL 9000 Corporate Tax Department
{bellcore,harpo,ihnp4,infopro,princeton,sys1}!motown!jmr
------------------------------------------------------------------------------