- Looking
back, in my professional life I have many times found myself in the middle
of a dichotomy, often fruitful, sometimes transitory and unstable.
- My first
experience of the UNIX operating system was in 1981, and it was then I
began to vacillate between the researcher who deals with 'deep' subjects
(concurrent programming, parallel systems) and the community geek who would
sell his soul in the laboratory to fine tune the tools and, by extension,
the environment, in which all other researchers can work profitably.
- My first
real experience with the world of networks - ("networks" is a
compliment when talking of connecting some minicomputers at the speed of
a serial terminal ..1-10 kbit/sec) - was with one of the first releases
of "Berkeley UNIX", perhaps the 4.1 BSD. Access to the operating
system code was exciting for two reasons: to understand how the arcane
mechanisms and perverse networking protocols functioned, but also to debunk
the cosy club of high-profile Californian programmers, who wrote code just
as mediocre as other programmers'. This gave me an excuse to participate
in one of the very first conferences of the European association of UNIX
users (EUUG, founded in 1981), in Dublin in 1983.
- It was
then I understood the distance there is between the "guild" of
UNIX programmers and the world of academia. The former had their own scientific
and cultural dignity, because their work was truly at the boundaries of
IT knowledge, but they did it with a collaborative spirit, which we still
find now in the Open Source community, in sharp contrast with the Hobbesian
world ("war of all against all") of academia (especially in Italy).
I began to sign my articles and reports for the Unix environment as "Joy
Marino", as they would be largely ignored in an academic career.
- After
a couple of years I collaborated in the founding of the Italian version
of the association of UNIX users, i2u, and here found another dichotomy,
sometimes salutary, but in the medium term, dangerous. Many of the founders
of i2u, unlike other EUUG members, had an industrial rather than university
background. At that time Olivetti and AT&T were partners - and UNIX
was one of the traded goods - but also Sperry/UNIVAC (soon to become UNISYS),
Siemens, IBM, and the newly founded SUN saw the importance of pre-competitive
collaboration through an association of users. An analogous situation existed
in the US, where there was an industrial type association (/user/group)
and one purely academic (USENIX), which just about managed to hold conferences,
not joint, but at least in the same town and on the same days. I treaded
water in the middle, sometimes feeling like a yo-yo, sometimes seeing the
undoubted advantages that siding with the industrial faction might bring.
- The next
time I encountered Yin and Yang was 2 years later in 1987. Participating
in the European conferences of EUUG I sneaked into a very informal meeting
of the group co-ordinating network connections that EUUG supported and
sponsored; the name "EUnet" began to be heard, but it was still
a vague, undefined entity.
- In this
way I discovered that another group of enthusiasts, from all over Europe,
had taken on the task, often purely on a voluntary basis, of maintaining
and operating a "network" of connections among UNIX computers
via modem, which was absolutely rudimentary and unreliable, and yet they
managed to channel large quantities of email both within the community
of users (Unix, in Europe and the US) but also with other types of networks
all over the world, through "gateways" at application level,
which translated the obscure and incompatible mail addresses that every
technological family had adopted.
- This was
another "magic circle" of initiates whose dedication to the cause
did not baulk at all-nighters checking that the modems and circuits were
working properly, lest their precious email be delayed by a few hours (much
of the system functioning was thanks to the efforts of a group we jokingly
called the "Dutch mafia": technicians mostly from CWI in Amsterdam.
Funny, because during the course of its evolution EUnet would have contact,
of another kind, with the Russian mafiathe real one...).
- Italy
had remained somewhat on the sidelines of this new adventure and, together
with an enterprising student (Alessandro Berni, who had contacted me via
email) on the one hand and the support of the industrial partners in i2u
on the other, I took on the job of "running" the network, once
again thinking of the tools and environment in which others could work
better.
- Alessandro's
email had travelled halfway round the world to get to me, passing from
the IBM BITNET mainframe network to UNIX, and I then discovered it had
been sent from the Computing Centre of my university…I have no idea how
he managed to get it through.
- The network
grew exponentially, simply due to the increase in use and number of users.
At the end of the 80s the principal junction lines had become dedicated
data lines and no longer phone circuits, the costs incurred - and shared
by all - were considerable, comparable to those paid by EUUG and by national
associations put together. While still a not-for-profit 'operation' the
service provided by the network had become indispensable to users and a
full time job to those running it. A technological upgrade was called for.
- In Europe
at the time the EU, or EC as it was then, sponsored initiatives to establish
networks of computers, provided that they used "standard" protocols,
the magic word being the palindrome ISO-OSI. In fact this was a showdown,
another dichotomy, between the network model adopted in the US (ARPANET,
NSFnet, TCP/IP, in one word: "Internet") and the European Community,
both the political side and, more especially, the academic one, affected
by the NIH syndrome (the first time an official EC document mentioned the
word "Internet" was in the Bangelman Report…1994).
- EUnet
was caught between this Yin and Yang and almost at the same time: a) decided
to participate in a research project sponsored by the EC with a project
for "migration to OSI" and b) submitted a request to EUUG for
financing of migration of the European network to TCP/IP. Neither was successful,
but the panorama was changing and, even without any real financing, all
the principal EUnet nodes adopted Internet protocols over dedicated lines.
- The Italian
node, which we hosted at Genoa, first got a 19.2 kbps line to INRIA in
the Sophia Antipolis tech-park (Antibes), then a direct line to Amsterdam
at 64kbps, and a 9.6 kbps link with CNUCE in Pisa (the first example of
a "gateway" in Italy between different protocols).
- It was
at this point that I found myself in another Yin and Yang situation, one
that would profoundly affect my life. It was no longer possible to run
the "network" on a purely voluntary basis.
- The "service"
I ran had a budget and paid technicians, but was based on mere division
of cost between the "affiliated" users and was not able to finance
itself even to the extent of buying a UPS system. In other parts of Europe,
managers of the UNIX portions of national networks heard the siren song
of venture capitalists and began to set themselves up as businesses.
- EUnet
itself was no longer a group of volunteers, it had a general manager, a
board that took decisions on economic commitments worth some millions of
dollars and, especially, had a strategic agenda completely different to
that of the association (also because with the recession UNIX, even if
stretched to "Open Systems" in the broadest sense, languished).
Then there was the advent of "competition": business initiatives
based on the same protocols (UNIX+Internet) and targeted at the same potential
users - for example, in the UK, PIPEX and Demon.
- In the
US Rick Adams had worked a miracle finding finance from the USENIX association
to convert the principal node of the worldwide UNIX network, "uunet"
into a commercial company, UUNet Inc. - UUnet Inc. was the principal Internet
operator worldwide until it was acquired by Worldcom in 1996, and it grew
further with the merger with MCI in 1997.
- After
the bankruptcy of MCI-Worldcom in 2003, operations resumed under the name
MCI and it was finally acquired by Verizon in 2006. In Europe the Board
of EUnet (on which I sat) set up the company EUnet Ltd, escaping from the
now troublesome control of the association.
- But by
now it was almost too late to emulate the success of the US Unix operation:
some national networks affiliated to EUnet had been acquired by new private
owners, others could not choose between a voluntary type set up and a commercial
one (Yin and Yang, again).
- In Italy
I tried to replicate the operation that had succeeded in Europe in 1993,
but with no luck. So then I was practically forced to choose one - and
only one - side of the many dichotomies I had found. This was how ITnet
was founded.
- To add
the finishing touch to the symbolism, I was also forced, making a virtue
of necessity, to abandon the "good guys", my friends across EUnet
with their idealism, and side with the "baddies", those unscrupulous
entrepreneurs who bet their shirts on the commercial success of the Internet.
I thought this was the end, but it was only a new beginning.
- It was
1994. The Internet boom was about to hit Italy.
- Looking
back now with some detachment, what seemed at the time to be tensions between
irreconcilable points of view now seem to be healthy dialectical opposition.
- It was
only by means of the vice-like grip of the Yin and Yang situations that
I have seen myself, and the other dilemmas faced by those who made the
history of the Internet in Italy, that it was possible to lay the solid
foundation for that nebulous thing we call the Internet.
- I suspect
it could be no other way: no "Intelligent Design", no "Masterplan"
blueprint could have produced the Net. It took the tension of opposites,
the rubbing of shoulders and compromises between academia and business,
science and technology, volunteers and entrepreneurs, to bring all this
about.
- A lesson
that should not be forgotten, when, for example we hear talk of "innovation".
|
