As the man who was one of the group who broguth JANET into existence,
and was its first Director, can I put a few glosses on to John
Hearns' note of last Wednesday on the development of JANET.

> Date:          Wed, 20 Nov 1996 10:38:53 GMT
> From:          "J.Hearns" <[log in to unmask]>
> To:            [log in to unmask]
> Subject:       Re: JANET

>
> I'd like to point out that the JANET deal is (as Ahmad says) open
> to students/faculty only.

Which is a pity really.  JANET is obliged to act in the way it does,
partly because of legal constraints, and partly because it is run as
a centrally underwritten service, with funds from the Higher
Education Funding Councils of Engalnd, Scotland and Wales (and
whatever the equivalent body for Northern Ireleand is now called).
I say it is a pity, because JANET has shown what can be achieved by
having a centrally funded network, with the service being free at the
point of delivery.  I would hate to see the NHS going for charging on
a usage related basis, but fear the worst.
>
> But JANET is a full part of the Internet anyway
> (there is no Internet. Just lots of networks, like JANET,
> connected by the Internet Protocol).
>
>

> I'd like to comment a little on the history of JANET,
> and a connection to NHSnet. Forgive me, as this may not be the
> place for my rantings.
> Please son't continue to read unless you are a real networks
> afficionado.
>
> At the recent MEDNET conference, Des Cline of the NHS IMG said
John ....  I think that must be Les Clyne

> that he had formerly been one of the JANET team.
> From it's beginnings, JANET ran on X.25 lines, just like NHSnet,
> and used proprietary protocols called the Coloured Books.
Not quite right;  the essence of the Coloured Books was that they
were NOT proprietary.  For non-guru's 'proprietary' in this context
means 'owned by a specific commercial group';  thus DECNET was
proprietary, the Internet protocols are not.  The coloured books were
created to ensure that University sites were not hog-tied to a
specific supplier as might (and did) happen where a site had a
substantial component from a single supplier.  Unfortunately, they
were not terribly good, or to be more precise, they were not well
implemented.  They were connection oriented, and carried a high
overhead;  but they did work, and they did inter-work among a wide
range of equipment.  However, because of the overheads, and the poor
implementation of many of them, they were never truly competitive
with the US implemented, Internet-related protocols.

> Thirteen or so years ago, as a nuclear physics student,
> I remember many happy hours of "TRANSFER /CODE=FAST" and
> coloured book email.
> The Coloured Books were incompatible with totally anything else,
> and our systems admins had to purchase them and install them
> on all our machines (in those days mainly mainframes and VAXen).
The Cololured Books wre NOT incompatible, in the sense that there
were protocol converting systems installed, whicgh worked well, and
allowed traffic to pass fairly freely between the Coloured Book and
the Internet Communities.  The really important thing is that such
inter-working was seen by the JANET management as an essential, and
failure to achieve it was just that, rather than an attempt to
prohibit interworking. The attitudes of the NHS are, from where I
sit, an intention to enforce restriction of access by the use of
protocols, which was never the aim of the JANET policy makers..


> Does this ring any bells with the NHSnet installation?
>
> Worse is to come - in order to communicate with CERN in Switzerland,
> the HEP commmunity had a leased line, and had to get CERN to install
> the Coloured Books on their mainframes.
> So we had a very 'closed' JANET, with mail relays off to BITNET.
> The Internet protocols came along, and were enthusiastically
> embraced by CERN and many other institutions worldwide.
> Of course, JANET was cut off from these.

Please see the remarks above about the provision of protocol
conversion systems;  and more important, the attitudes that informed
this provision.  I can categorically recall the use of JANET
protocols to send/receive mail and files from CERN to Leeds,
primarily because my daughter started there as a graduate student at
about this time.
> The first chink in the armour came when the Astronomy community
> needed DECNET access to the SPANS network, and the HEP community
> needed DECNET to our VAXes at CERN.
> We had to bang our fists up and down on the table at meetings
> to get permission to use DECNET.

I remember organising these meeetings to provide the opportunity for
folk like John to bang his fist.  I think he may underestimate the
extent to which those of us running JANET worked in ordwer to allow a
proprietary network (DECNET) to operate across JANET, which had a
very large community of users other than the rather small group who
worked in High Energy Physics.  The important point here (Ahmad
please note) is that it is ESSENTIAL for the users to continue to
attend meetings, in order to make their views known.


> I was a very happy young person the first day I ran a remote
> X-windows session from Glasgow to my workstation in Geneva.
> That's something you all take for granted today.
>
> JANET originally then began to carry IP over X.25,
> and the rest is history, ie. the research community got
> what they wanted with open connectivity to the Internet.
> Remember - if you are an ac.uk user
> JANET didn't bring you the Internet. User pressure did.

I really must protest at that!!!  It implies that the JANET policy
was intended to retsrict access to Internet services.  It was not.  I
know.  I was there when the policy was formulated.  It may not have
moved as fast as some might have liked, but that was in order to
safeguard the overall interests of a large community of users, and to
avoid getting sucked into the exclusive use of proprietary protocols.

>
>
> Amazingly enough, there is still life in the 'old' JANET yet,
> with (I think) some JNT PADs and X.25 lines operating here.

I think that is still the case;  and what is more the folk using that
equipment are quite happily inter-working with the rest of JANET,
and the Internet community.
>
> I should have spoken up at MEDNET, but I still see this centralist
> philosophy operating in the NHSnet:
> "We'll choose the protocols and you'll run them.
> We'll negotiate deals for leased point-to-point lines and you'll use
> them. Cost and incompatibility with everyone else on the planet
> are unimportant as we know better".
If the use of a central deal can gain more favourable terms from the
suppliers, then why not.  Unfortunately, BT have always proved very
good at resisting pressure to provide favourable deals (they always
used to argue that they were legally prevented from doing so)  the
other PNOs are much more amenable.  The really important points are
to ensure that charges are not usage related; although they may of
course be bandwidth related.  I know of no more effective mechanism
for restricting growth of use of a network than a usage related
charge.  And as I said earlier, it was never the intention with JANET
to restrict the ability to interwork, as distinct from using
purchasing power to acquire better deals.  Unfortunately, the
products themselves were simply not as good as those provided to the
Internet protocols.

>
>

 Well, as Ahmad says, people are voting with their feet,
> just as we did ten years ago in JANET.

I note that Jonh's address, like mine, ends in ac.uk;   that is
JANET.  The moral;  don't take your bat home.  Ensure that your
umbrella organisation is able to do what JANET did, and say 'OK,
there are better products out there, let us use them'

Mike Wells

==========================================================
Professor Mike Wells
Department of Physics,    The University of Leeds
Leeds LS2 9JT,            United Kingdom
Phone: 0113-233-3879      E-Mail   [log in to unmask]
==========================================================


%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%