[comp.org.decus] VUnet

rshuford@well.UUCP (09/19/87)

Well, my second posting about the VMS Users Network brought a new set of
responses, so here's some more fuel for the fire. 

In the few months that I have been reading the newsgroup "comp.os.vms", I've
noticed that about every two weeks at least one person posts or relays a
message containing the question, "How does one connect a VAX computer
running VMS to the Usenet for mail and news?" For a long time I was puzzled
that nobody posted an answer.  My unspoken assumption went like this:

   "VAXes are famous for networking.  It seems like half of the machines
   on the networks are VAXes.  Surely it must be a simple matter to connect
   a VAX running its native operating system to the fundamentally simple
   Usenet."

That is, until I started seriously asking the same question myself.  I
slowly came to realize that there was a hitherto unsuspected gap in the
connectivity of the VAX.  Although DECnet running on a VAX is an indisputably
efficient style of networking for providing information transfer within a
single organization, there also exists in many cases a need for informal,
inter-organzational communication that is not easily met by the formal
mechanisms of DECnet.  For a long time, there has been no good answer to the
often-asked question of Usenet connectivity. 

Obviously, some people have found means to work around the problem.  Most such
people seem fortunate--they work in technically sophisticated sites with
extensive local networks, have access to computers running both Unix and VMS,
have access to Unix source code for fitting crucial pieces together, and know
enough to do it.  Or they have enough money to buy one of the commercial
Unix-flavor VMS enhancement packages that can provide the necessary software
base. 

There are, however, a lot of VMS sites that are not so blessed.  There has
been no way for an isolated, unsophisticated, or poor VMS site to connect
via UUCP for mail or Usenet news. 

Fortunately, other people had noticed the connectivity gap a lot earlier. 
Jamie Hanrahan was one of them, and he was sufficiently industrious that he
organized an effort within the Digital Equipment Computer Users Society to
try to plug the gap.  At the Spring 1987 DECUS Symposium in Nashville, some
of the results of his and others' effort began to fall into place. 

My previous postings (3856@well.uucp and 3887@well.uucp) described as best I
could the status of what we are tentatively calling "VUnet".  (If somebody
suggests a better name, it would receive consideration.)  VUnet will allow
VAX/VMS sites to create informal, inter-organizational networking links
similar to those that are typical of the UUCP/Usenet system.

Designing and building any substantial network is an engineering problem. 
There are probably 100 ways you could put a VMS network together.  Some
compromises must be made, so perhaps the framework I've sketched doesn't
contain 100% of everybody's favorite network traits.  But I hope that
readers of these postings will understand the limitations of a volunteer
effort and cooperate with the members of the DECUS working group who are
giving up their spare time to work on putting the VUnet tools in place. 

HOW TO CONNECT TO THE WORLD
--- -- ------- -- --- -----

The DECUS working group emphatically does want VUnet to be connected to the
larger networking world, either by transparent interconnection or by
gateways.  Connecting to Unix machines via the UUCP method was considered
obviously desirable, but only in the past few months has non-AT&T code to do
this been available.  Using PMDF/mmdf links alone would have meant that only
certain 4.3BSD Unix sites could serve as gateways between Usenet and VUnet,
but it seemed the best that could be done in the short term.  The mmdf
program is quite widely used in the PhoneNet portion of CS-Net, so possibly
had I posted to a CS-Net audience the catcalls would have been fewer. 

At the time of my earlier posting, I had no certain information about DCP,
one of the public-domain programs for UUCP.  I have now been informed by
Jamie that he has been partially successful in using DCP under VMS.  After
some debugging it should be possible to use the DCP code to operate a
"uucico" channel within PMDF, for VUnet sites to talk to Usenet sites
running UUCP. 

PMDF will still be the fundamental software used in VUnet.  Its virtues:

 -  low cost ($50 U.S.)
 -  clean interface to VMS mail
 -  smart address rewriter
 -  RFC-822 compatibility (for Internet-type addressing)
 -  programmable mail handler
 -  access to DECnet channels where they exist
 -  integral PhoneNet-style dialup connection to like-minded systems
 -  source code available in Pascal
 -  support for Excelan and Wollangong TCP/IP channels
 -  support for Jnet RSCS channels


NEWS-HANDLING SOFTWARE
---- -------- --------

Jamie now reports that he has heard from Geoff Huston of the Australian
National University; Mr. Huston says that a fully "netnews"-compatible,
VMS-based version of the DECUS Library program VAX-214 News will be
released in October 1987, with distribution possible through DECUS
at low cost to the user.  (I hope that it could be distributed by
Local User Groups--LUGs--in their economical fashion.)

Another possibility is a program developed at Brigham Young University:
BYU News.  The following information was sent to me by Jim Logan,
LOGANJ@BYUVAX.BITNET.

 > We have developed ... a message posting and display system for VAX/VMS 
 > systems.  We call it BYU News, and we use this system to display and
 > reply to the Usenet BBS as well as local BBS items....

 > We will be happy to provide a copy of the software [on tape or over
 > BITNET] for your use, but we must ask a service fee of $200 per site to
 > cover our costs.  The $200 is refundable if you are not satisfied.

 > Our limited measures indicate that this software is perhaps as much
 > as 4 times more efficient than the Unix ReadNews software.  We process
 > an entire day's worth of Usenet text in an average of less than 10
 > minutes of CPU time on a VAX 8600.

 > ...our software will work with any program(s) that can capture the
 > Usenet files in a VMS "variable length carriage return" format....

 > Another site is using our software with Carnegie-Mellon University's
 > TCP/IP to capture the Usenet batch files directly from a Unix system. 
 > After receiving the Usenet batch files, the BYU News program uses a DCL
 > procedure to route the batch files to other systems on campus (both VMS
 > and Unix) via DECnet or Jnet or TCP/IP. 

 > The human interface to our software uses the VAX TPU (text processing
 > utility) and provides a friendly "point and do" capability for reading,
 > posting, and managing the news....

The working group is also monitoring the status of the Unix-based news
programs of Usenet.  It may be possible that in the next few months the
rewritten News version C will come out, which may be easier to port to VMS
than the existing B release. 
 
If you would like to help, please contact

   Jamie Hanrahan
   c/o Simpact Associates
   9210 Sky Park Court
   San Diego, CA 92123 USA
   (phone:    619/565-1865)
   (UUCP:     {akgua | hplabs!hp-sdd | sdcsvax | nosc}!crash!pnet01!jeh)
   (ARPA:     crash!pnet01!jeh@nosc.mil)
   (Internet: jeh@pnet01.cts.com)

PMDF on a 9-track BACKUP magtape may be ordered by sending a check for $50 to:

   Ned Freed
   The PMDF Project
   Computing Services
   Harvey Mudd College
   Claremont, CA  91711

or inquire to    uunet!ymir.bitnet!ned  
           or  psuvax1!ymir.bitnet!ned

Those of you who dislike VMS VAXmail might want to investigate the "tools
mail" facility available from DECUS as part of the Software Tools package,
which Glenn Everhart says is available on the recent Languages & Tools SIG
tape or an old (Spring 1984?) RSX SIG tape. 

If you missed the previous newsgroup postings, you can look in the hardcopy
August 1987 issue of the DECUS Joint SIGs Newsletters (Vol. 2, Num. 12) in
the VAX Pageswapper section (Vol. 9, Num. 1, pp. VAX-19 through VAX-31),
for Jamie's own description of the effort.  Otherwise, although my formal
role in this is just "regional agitator", I'll respond to inquiries by
physical mail until 6 December A.D. 1987, the time of the Anaheim DECUS
Symposium.  Mail a stamped, self-addressed envelope to me at the address
below, and I'll send you a summary document. 

I misspelled "uucico" in my earlier postings.  Oops. 

.....Richard S. Shuford             UUCP: {hplabs | ptsfa} !well!rshuford
     42 Mountain View Dr. East      BIX:  richard
     Hickory, NC 28602-5431         19 September A.D. 1987