NetNews Usenet Archive 1993 #3/NN_1993_3.iso/spool/bit/listserv/nodmgtl/860

Text File | 1993-01-21 | 9.3 KB | 179 lines

Comments: Gated by [e-mail]
Path: sparky!uunet!paladin.american.edu!auvm!JHUVM.BITNET!JIMJ
Message-ID: <[e-mail]>
Newsgroups: bit.listserv.nodmgt-l
Date:         Thu, 21 Jan 1993 13:27:39 EST
Reply-To:     [e-mail]
Sender:       Node Management <[e-mail]>
From:         Jim Jones <[e-mail]>
Subject:      Re: _possible_ death of BITNET
Comments: cc: [e-mail]
In-Reply-To:  Message of Thu,
              21 Jan 1993 09:51:49 -0500 from <[e-mail]>
Lines: 164

 Before I even get started, let me say that I've CC'd a copy of this
 note to FUTURE-L@UGA, and directed replies to that list.  If you want
 your responses to appear on NODMGT-L, you'll probably have to manually
 twiddle the destination.

On Thu, 21 Jan 1993 09:51:49 -0500 <[e-mail]> said:
>In article <[e-mail]>, Jim Jones
><[e-mail]> writes:

 My argument is essentially that connectivity issues are not the only
 important measure of a "computer network".  And that if Bitnet does
 deliver useful services it can survive even if it does rely heavily
 on some other network for transport services.

 In that context, I'll respond to your observations.

>1.      There is no such person as "BITNET", hence he cannot provide services.
>        There are quite a number of independent entities that provide services
>        over the NJE catenet and, in many cases, the Internet as well.
>        BITNET is the subset of those entities that (a) operate a cooperative
>        NJE network which (b) they have hired CREN to administer and
>        coordinate.

 I do use the term "Bitnet" somewhat ambiguously.  Sometimes I mean CREN,
 and sometimes I mean the cooperative efforts of Bitnet members.  Both
 provide valuable services.  Eric has mentioned LISTSERV, which is the
 most visible example of a service offered by the later group.  Many
 sites contribute resources (machine and human) to maintaining a working
 LISTSERV structure that delivers first rate services.  And while users
 on the Internet can participate, it's by definition a Bitnet service.
 The INTERBIT services provided by the members of Bitnet is another
 example.  CREN also provides services, like a coordinated respository
 of information about the members.  That is also a valuable service.

>2.      Some people confuse the terms BITNET and CREN.  CREN does provide
>        services.  Those services today are almost entirely concerned with the
>        operation of the NJE network.

 Both CREN and the Bitnet members themselves offer Bitnet-based services.
 My point is that people that feel it's useful to preserve Bitnet are
 charged with defining the services that make them feel that way, and
 ensuring that those services are sufficient to sustain the network.
 Which brings me back to my original point.  If Bitnet doesn't survive,
 it will be because CREN and the Bitnet members that would like Bitnet
 to continue weren't able to deliver a suite of services convincingly.
 Connectivity is not the primary issue as was suggested in the original
 posting that started this thread.  Does it mean the CREN/"Bitnet folks"
 have to offer new services and expand existing ones?  Yes I think it
 does.  But the idea that Bitnet becomes completely irrelevant the moment
 the last PC is connected to the Internet is a gross oversimplification.

>3.      The NJE network provides two facilities that the Internet traditionally
>        has not: Nodal Message Records and unsolicited file transfer.  There
>        are at least two NMR-over-IP schemes in some stage of development.
>        Unsolicited file transfer is more difficult, but not impossible, and
>        I'd be surprised to find that nobody was working on it.  (I'll probably
>        tackle it eventually if nobody else does.)

 The Internet Message Send Protocol as I've seen it defined isn't a
 true alternative to NJE "messages".  There are authenticity problems
 in many TCP/IP based services, including MSP.  That is, while on an
 NJE network you can pretty much be sure that an interactive message
 you receive from someone was actually sent by the user/node that
 it claims to be from.  Unless MSP is extended, it's no more reliable
 then SMTP based mail in that regard.  Yes, it's a minor point, but
 I think it's worth mentioning.  And in anticipation, yes, since many
 Bitnet links are TCP/IP based now, it would be possible to spoof NJE
 messages on one of those links.  But the logistics of doing so seem
 nightmarish in comparison to spoofing an MSP message.  And it's ironic
 to note that it's precisely because the transport isn't over an
 "old fashioned" NJE line that the potential security exposure exists.

 The real point is, will Bitnet be desirable when MSP and SIFT are
 available to Internet users?  Like I said before, people that think
 so, probably have a few years to do something about it.  MSP and
 SIFT aren't going to be widely available on the Internet before then.

>4.      When these last two facilities are widely available in the Internet,
>        there is no more need for the NJE network.  The bother and expense of
>        tearing it down will prevent its removal for some amount of time,
>        but not forever.

 You're assuming that nobody is interested in offering new services or
 expanding the currently existing ones.  If you're right, then you're
 right.  But given that assumption, the question is rather trivial.
 It's like saying, "Since the patient has stopped breathing, he won't
 remain in this hospital bed much longer.  Only the bother of carting
 him to the morgue will delay his removal, but not forever.".  The
 patient still seems healthy to me.

>5.      When the NJE network is gone, BITNET will be the set of Internet users
>        who remember what BITNET used to be.

 Again, your conclusion is just a rewording of your assumption.

>        ...      CREN will be out of the transport business and into
>        application-layer stuff like mailing-list software and other
>        education/research end-user tools.
 (text deleted...)
>But *should* we continue to run such a network-within-a-network?  What does it
>buy us?

 Now here's a good question...  I've heard some people say that Bitnet's
 commercial restrictions make it a valuable organization.  That is, since
 Bitnet members are largely academic institutions, whatever services were
 envisioned and provided (by CREN or the members themselves) are more
 likely to have the their interests in mind.  If that's true (having no
 experience in a non-academic environment, I can't say), then Bitnet as
 a network-within-a-network could be useful.  CREN could expand it's
 role as a coordinator of centralized information, and by definition
 become a source of information about academic institutions.  If CREN
 can be used a research tool, or just a way to find information, then
 site may very well want to continue their membership.  The question is
 do Bitnet members as a subset of the Internet community have common
 needs that can be effectively address by some organization?  Or is it
 every university for itself along with every tom.burgerking.com,
 dick.nestle.com and harry.chrylser.com on the Internet? :)

 (text deleted...)
>I thought that the argument was, that the consumers all seem to be buying one
>of the choices anyway, and eventually will wonder why they need both.

 I hear the argument that "everyone should run U*NIX only and be on the
 Internet only just because things would be easier" all the time.  I
 don't agree with that argument for the reasons I stated.  My point was
 that as long as diversity is feasible, it's likely to be beneficial.
 As for the point you've raised.  I hope people decide whether or not
 to retain a Bitnet membership by considering the costs and resulting
 benefits, and not because the latest "hot trend in MIS circles" is
 to get rid of that nasty NJE network.

>>       Currently Bitnet offers services not available on the Internet and
>>  I don't see that changing in the new year or so.
                                     ^^^ "next", not "new" as typo'd

 I still maintain that the statement above is accurate, and that only
 sites that are truely bankrupt will be acting in their users best
 interest by dropping out of Bitnet.

>At that point BITNET will not be a technical network at all anymore; it will be
>a social network, and that in the process of being absorbed by a larger one.
>It will be a fading memory, something that old-timers talk about when they
>congregate.

 The NJE network isn't going away in the foreseeable future.  Bitnet is
 not going to turn into an eccentric bunch of drooling systems programers
 babbling about core memory and wire wrapped 3704 field updates. :)

>Okay, what are the advantages of the BITNET alternative, aside from unsolicited
>file transfer?

 I've listed some real, and some possible Bitnet-specific services in
 this (overly long) note already.  What I'm wondering about is whether
 or not CREN and Bitnet members at large are interested in making Bitnet
 into a viable alternative/addition to other networks.  If there is a
 commitment to do so, I think it will happen.  There's certainly enough
 talented people out there.

>--
>Mark H. Wood, Lead Analyst/Programmer    [phone]   [@disclaimer@]
>Internet:  [e-mail]     BITNET:  IMHW400@INDYVAX
>Celebrate freedom:  read a banned newsgroup.

 -jj