[comp.sys.dec] DECnet encapsulation in TCP-IP

JHL1@psuvm.psu.edu (11/10/90)

We're interested in encapsulating DECnet within a TCP-IP package.  We
are looking primarily for a software solution, although hardware
solutions would be considered.  If anyone knows of such a product.
please e-mail at the following address.

JEFF@PSUPEN.PSU.EDU

Thanks,
Jeffrey H. Lynn
Penn State College of Agriculture

grr@cbmvax.commodore.com (George Robbins) (11/10/90)

In article <90313.134117JHL1@psuvm.psu.edu> JHL1@psuvm.psu.edu writes:
> We're interested in encapsulating DECnet within a TCP-IP package.  We
> are looking primarily for a software solution...

I believe TGV multinet can handle this...

-- 
George Robbins - now working for,     uucp:   {uunet|pyramid|rutgers}!cbmvax!grr
but no way officially representing:   domain: grr@cbmvax.commodore.com
Commodore, Engineering Department     phone:  215-431-9349 (only by moonlite)

rbraun@polygen.uucp (Richard Braun) (11/20/90)

grr@cbmvax.commodore.com (George Robbins) writes:
>In article <90313.134117JHL1@psuvm.psu.edu> JHL1@psuvm.psu.edu writes:
>> We're interested in encapsulating DECnet within a TCP-IP package.  We
>> are looking primarily for a software solution...
>
>I believe TGV multinet can handle this...

MultiNet is, IMHO, far and away the best TCP/IP package available for
VAX/VMS.  (This after reviewing products ranging from Process Software
to Fusion to Wollongong to...)  It's a Stanford-bred product written
in C with some TOPS-20'isms in it.

But I fail to understand the original question:  "encapsulating DECnet
within TCP/IP".  That just doesn't mean anything.  It could mean
buying a binary license to DECnet and one to TCP/IP and writing an
application which calls both.  Or it could mean buying a MultiNet
source license for $megabucks plus a TCI "CommUnity" DECnet source
license for $megabucks and integrating the two.  Or it could mean
going to DEC or Equinox or some other company and buying a LAT/TCP
terminal server to hook onto your Ethernet.  As you can see, you
can come up with solutions that cost anywhere from $3,000 to
$3,000,000.

What's the need, anyway?  And was it on VMS-only or does it have to be
platform-independent?

-rich

fac2@dayton.saic.com (Earle Ake) (11/21/90)

In article <897@fred.UUCP>, rbraun@polygen.uucp (Richard Braun) writes:
> grr@cbmvax.commodore.com (George Robbins) writes:
>>In article <90313.134117JHL1@psuvm.psu.edu> JHL1@psuvm.psu.edu writes:
>>> We're interested in encapsulating DECnet within a TCP-IP package.  We
>>> are looking primarily for a software solution...
>>
>>I believe TGV multinet can handle this...
> 
> MultiNet is, IMHO, far and away the best TCP/IP package available for
> VAX/VMS.  (This after reviewing products ranging from Process Software
> to Fusion to Wollongong to...)  It's a Stanford-bred product written
> in C with some TOPS-20'isms in it.
> 
> But I fail to understand the original question:  "encapsulating DECnet
> within TCP/IP".  That just doesn't mean anything.

	I think what he means is sending DECnet packages inside of TCP/IP
packets.  I am using IP over DECnet right now.  My TCP/IP packets are
enclosed in DECnet packets, sent over our company DECnet and received by
a machine that 'unwraps' them and then sends them off to the Internet.

-- 
_____________________________________________________________________________
             ____ ____    ___
Earle Ake   /___ /___/ / /     Science Applications International Corporation
           ____//   / / /__                 Dayton, Ohio
-----------------------------------------------------------------------------
Internet: fac2@dayton.saic.com             uucp: dayvb!fac2

grr@cbmvax.commodore.com (George Robbins) (11/21/90)

In article <897@fred.UUCP> rbraun@fred.UUCP (Richard Braun) writes:
> grr@cbmvax.commodore.com (George Robbins) writes:
> >In article <90313.134117JHL1@psuvm.psu.edu> JHL1@psuvm.psu.edu writes:
> >> We're interested in encapsulating DECnet within a TCP-IP package.  We
> >> are looking primarily for a software solution...
> >
> >I believe TGV multinet can handle this...
> 
> But I fail to understand the original question:  "encapsulating DECnet
> within TCP/IP".  That just doesn't mean anything.  It could mean
> buying...

Usually enscapulation means that you have a network path that supports
only one protocol, such as a point-to-point DECnet link, over which you
would like to have packets for some other protocol such as TCP/IP pass
transparently.  This is pretty much the opposite of what a gateway does.

Wollengong, Multinet and probably countless others handle the case of
getting TCP through a VMS DECnet link, I assume that this person was
interested in doing the opposite, trying to get a transparent DECnet
link over an internet or other "TCP/IP" path.

-- 
George Robbins - now working for,     uucp:   {uunet|pyramid|rutgers}!cbmvax!grr
but no way officially representing:   domain: grr@cbmvax.commodore.com
Commodore, Engineering Department     phone:  215-431-9349 (only by moonlite)

don@nic.the.net (Donald L. Nash) (11/21/90)

In article <897@fred.UUCP>, rbraun@polygen.uucp (Richard Braun) writes:
>From: rbraun@polygen.uucp (Richard Braun)
>Subject: Re: DECnet encapsulation in TCP-IP
>Date: 19 Nov 90 20:01:53 GMT
>Organization: Polygen Corporation, Waltham, MA
>
>grr@cbmvax.commodore.com (George Robbins) writes:
>>In article <90313.134117JHL1@psuvm.psu.edu> JHL1@psuvm.psu.edu writes:
>>> We're interested in encapsulating DECnet within a TCP-IP package.  We
>>> are looking primarily for a software solution...
>>
>>I believe TGV multinet can handle this...
>
>MultiNet is, IMHO, far and away the best TCP/IP package available for
>VAX/VMS.  (This after reviewing products ranging from Process Software
>to Fusion to Wollongong to...)  It's a Stanford-bred product written
>in C with some TOPS-20'isms in it.

I agree completely.  We use MultiNet extensively here, and I couldn't get along
without it.  TGV also does a good job with technical support as well.  But
anyway, back to the business at hand:

>But I fail to understand the original question:  "encapsulating DECnet
>within TCP/IP".  That just doesn't mean anything.

Perhaps a better way of saying this is "tunneling DECnet across an IP network". 
Here is an excerpt from the _MultiNet(tm)_System_Administrators'_Guide_ for
MultiNet Version 2.2.  This is the introductory paragraph for Chapter 5, found
on page 5-1:

A special DECnet device driver allows the MultiNet system manager to configure 
a DECnet line and circuit between two cooperating MultiNet systems across an
arbitrary IP network.  This special driver encapsulates DECnet packets in UDP
datagrams for transportation via the IP protocols, much like VAX/PSI
encapsulates DECnet packets in X.25 when doing Data Link Mapping.

This was quoted without permission, but I'm sure the nice folks at TGV won't
kill me for it.  Hopefully, they won't sue me for it either. :-)

I hope that clears the waters a bit.

				Donald L. Nash

				The University of Texas System
				Office of Telecommunication Services

Internet:  don@nic.the.net
THEnet:    THENIC::DON
BITNET:    DON@THENIC
PSI Mail:  311051200131::DON

rmilner@zia.aoc.nrao.edu (Ruth Milner) (11/21/90)

In article <897@fred.UUCP> rbraun@fred.UUCP (Richard Braun) writes:
>grr@cbmvax.commodore.com (George Robbins) writes:
>>In article <90313.134117JHL1@psuvm.psu.edu> JHL1@psuvm.psu.edu writes:
>>> We're interested in encapsulating DECnet within a TCP-IP package.  We
>>> are looking primarily for a software solution...
>>
>>I believe TGV multinet can handle this...
>
>MultiNet is, IMHO, far and away the best TCP/IP package available for
>VAX/VMS.  
>
>But I fail to understand the original question:  "encapsulating DECnet
>within TCP/IP".  That just doesn't mean anything.  
>
[...]
>What's the need, anyway?  

Encapsulated DECnet is DECnet packets wrapped inside IP packets. To IP-only 
systems, the packets appear as normal TCP/IP and are routed accordingly. When 
they arrive at the destination (running MultiNet), the software strips off the
TCP/IP stuff, finds a DECnet packet inside, and hands it off to DECnet for
the real processing. As far as DECnet is concerned, a DECnet packet just came
off the net.

DECnet is very picky about nodes being "adjacent", i.e. directly reachable
through some line or another. Under normal circumstances you would have to 
have DECnet nodes every step of the way to reach a remote system.  With 
DECnet-over-IP, however, the IP-only node(s) in between are transparent
to DECnet, and even though the two nodes could be miles apart, they appear
to be adjacent. This can *enormously* simply configuring a wide-area DECnet,
especially if there are existing links which only run TCP/IP. It also reduces
the number of systems which must run DECnet; since most UNIX implementations
(at least as of a year ago) cannot do routing between networks, encapsulating 
DECnet within IP can save your hide if you find you need to talk to some 
(DECnet) system a few UNIX routers/gateways away.
-- 
Ruth Milner
Systems Manager                     NRAO/VLA                    Socorro NM
                            rmilner@zia.aoc.nrao.edu

hughes@copper.ucs.indiana.edu (larry hughes) (11/21/90)

In article <897@fred.UUCP> rbraun@fred.UUCP (Richard Braun) writes:
>grr@cbmvax.commodore.com (George Robbins) writes:
>>In article <90313.134117JHL1@psuvm.psu.edu> JHL1@psuvm.psu.edu writes:
>>> We're interested in encapsulating DECnet within a TCP-IP package.  We
>>> are looking primarily for a software solution...
>>
>>I believe TGV multinet can handle this...
>
>MultiNet is, IMHO, far and away the best TCP/IP package available for
>VAX/VMS.  (This after reviewing products ranging from Process Software
>to Fusion to Wollongong to...)  It's a Stanford-bred product written
>in C with some TOPS-20'isms in it.
>
>But I fail to understand the original question:  "encapsulating DECnet
>within TCP/IP".  That just doesn't mean anything.  It could mean
> [etc.]

DEC used to have a product called an "internet portal" which
enveloped TCP/IP within DECnet.  It was a hardware/software
solution consisting of an Ultrix workstation at both ends
(i.e. on two networks separated by a DECnet backbone) to handle
the enveloping/de-enveloping.

It's quite possible that they reversed this to envelope DECnet
withing TCP/IP...

 //=========================================================================\\
||       Larry J. Hughes, Jr.        ||        hughes@ucs.indiana.edu        ||
||        Indiana University         ||                                      ||
||   University Computing Services   ||   "The person who knows everything   ||
||    750 N. State Road 46 Bypass    ||      has a lot to learn."            ||
||      Bloomington, IN  47405       ||                                      ||
||         (812) 855-9255            ||   Disclaimer: Same as my quote...    ||
 \\==========================================================================//

mogul@wrl.dec.com (Jeffrey Mogul) (11/22/90)

In article <73306@iuvax.cs.indiana.edu> hughes@copper.ucs.indiana.edu (larry hughes) writes:
>DEC used to have a product called an "internet portal" which
>enveloped TCP/IP within DECnet.  It was a hardware/software
>solution consisting of an Ultrix workstation at both ends
>(i.e. on two networks separated by a DECnet backbone) to handle
>the enveloping/de-enveloping.

As far as I know, we still sell the Internet Portal.  Quite a few
are in use inside Digital.

>It's quite possible that they reversed this to envelope DECnet
>within TCP/IP...

Not that I know of.

-Jeff

warren@atmos.washington.edu (David Warren) (11/27/90)

Since someone already mentioned MULTINET, I'm running MULTINET's
encapsulated DECNET curently along with the rest of the U of W. It
works fine. We have links to northwestnet and other networks, all
using IP encapsulated DECNET packets. By the way, the new version
allows the use of TCP instead of Datagrams for better reliablility.
--
David Warren		INTERNET: warren@atmos.washington.edu
(206) 543-0945		UUCP:	  uw-beaver!atmos.washington.edu!warren
Dept of Atmospheric Sciences, AK-40
University of Washington