AEIOUnet

Trinity College Dublin was, of course, the breeding ground for many figures of Internet-consequence to emerge.Dennis Jennings, for example, who although he came to be associated most with UCD, actually did his Ph.D. in TCD. One of these was Mike Nowlan, who is probably the person with the best claim on bringing up Ireland’s first Internet connection. (We’ll talk about that more shortly.)

Another of these figures of consequence was Alan Judge, who was brought in by Mike and another founder of the company, Cormac CallananLater director of the anti-child-pornography organisations, hotline.ie and inhope.org. close to the very start of IEunet. Alan was a part-time postgrad and also part-time Technical Director of IEunet for many years, back when it was still possible to actually do that.He later went on to be director of R&D in Indigo, in turn bought by Eircom, and we’ll hear more about that shortly.

In the good old days, as was illustrated in the Ireland On-line section, mostly everything was email, and you didn’t have an interactive Internet as such: you dialled up periodically and “sucked down” your email via a process known as UUCP, famous for commands that sounded quite like ducks quacking.For example, uux and uucico. Or ancient Greek renditions of frog noises, such as Brekekekèx-koàx-koáx. European networking in general could be said to have grown out of that aforementioned Unix computer operating system, from a series of organisations called EUUs – European Unix User Groups. The seminal pan-European organisation, EUnet, was actually incorporated in Ireland, and Mike Nowlan was on the original board (and, indeed, was a 25% shareholder). Of course, this was quite separate to IEunet, the Ireland-specific Internet connection company, in an arrangement which still seems perhaps a little complicated.For example, IEunet was set up under the aegis of the Irish Unix User’s Group, which was very closely associated with DECUS (Digital Equipment Corporation User Society). In fact, the name dec4ie, IEunet’s main server for some years, originated thus: it was a gift from DEC to the IUUG and was lent out to IEunet on a temporary basis. Indeed, IUUG had a contractual entitlement to a board position in IEunet, but this was eventually forgotten about and the IUUG fell into disrepair. – Nick Hilliard

Although the larger historical progress of European networking is a fascinating one, and the details of some of the political wars around the protocols, national or otherwise incumbent telecomms companies, and lots and lots of research money, are touched on elsewhere, I will leave a broader examination to works such as A History of International Research Networking. Instead I am going to focus specifically on the Irish story.

Consequently we can begin the IEunet story with Michael Nowlan and Cormac Callanan providing UUCP and related services on a commercial basis in approximately 1991. Attentive and thoughtful growth from the early stages demonstrated that there was a sustainable commercial basis for Internet services in Ireland, and furthermore that it was possible to actually make money by offering them. (Neither of those things could be taken for granted at that time in a small country like Ireland.)

Michael himself had always been interested in the mathematical side of things, and ended up doing science - in this case, Chemistry - in UCD’s science campus of the time, Earlsfort Terrace (later across the road from Ireland On-line):

I started in ‘67 and I was the first year that they were fully there. The first year was difficult, but second year I had such a ball. I have never had as good a year as in second year. It’s after that you realised that you’re able to make it, first year you were sort of expecting you’d be thrown out, and I suppose I got comfortable comfortable with girls at that stage. It was bloody marvellous!

Even back then it was obvious that chemistry was becoming more mathematical and therefore more computer-based. Indeed, in his final year, Michael believes he had a FORTRAN course, and after that experience was determined to have a career in computing.Note: a number of other people have had FORTRAN experiences and gone very much the other way, just to be clear. After a Masters in Computing Applications, which was effectively a conversion course, he went to TCD to fill a job doing administrative-related programming. Apart from a year abroad in Sweden, connecting computers together in 1976,It wasn’t actually called networking, but it was as good enough a match as you’d get at the time – connecting together multiple mainframes with channel-to-channel adaptors and front-end computers which were 16-bit mini-computers and cheap teletypes that went BZZZZZZZ printing our nuclear launch codes or something. he never left TCD. But it provided plenty of scope for his talents.

When he got back to TCD, they had gotten (in words that will warm the heart of many an early-generation nerd) a DEC-20 with a PDP front-end, and Michael PurserIn another nod to history, the founder of Baltimore Technologies in 1976, before it’s sale to Dermot Desmond in 1996 and subsequent wild stock-market ride. was looking to try to connect UCD and TCD, in an early version of HEAnet called the Irish Universities Network. That run from the early 80s to the early 90s was seminal for TCD and for Ireland, Internet-wise. According to Michael, it gave birth to the first Internet connection into Ireland:The more complicated background is that apparently a Maths lecturer called Tim Murphy in TCD discovered UUCP, as referenced above. To quote Michael, “He was a UNIX guy and much of the Maths Department were heavily UNIX and despised us with our VAX VMS and we thought they were rank amateurs and we were all wrong! So Tim developed with someone - and I think it was a guy called Brendan Lynch who was a Computer Science postgrad student - they developed 7- and 6-bit versions of UUCP that would work over an X.25 pad and they could suddenly get into the University of Kent, which mean Trinity was actually connected enough for email, Usenet, and so on.”

By a month, it was Trinity connecting to the University of Canterbury in Kent, in 1991. Literally a month or so ahead of UCD.

By a stroke of good fortune, we have Michael’s original email describing this connection to have happened on the 17th of June, 1991. (Perhaps the truest thing said in that mail was: “No gaurantees [sic] of reliable service are offered at present, it is quite likely that the line with [sic] go down at no notice.”).It was 19.2kb/s, approximately one metric bajillionth of what you have in your pocket or your hand right now.

Showing perhaps how these things failed to be perceived as historic at the time, the connection was only announced to staff and postgraduates in Trinity. Continuing the tradition of slightly odd arrangements, however, was the question of how exactly Trinity and UCD were connected, which was through a European-flavoured organisation called IXI:

IXI was a European, an EU-sponsored X.25-network for academia. And so, in 1991, UCD connected there. And it took them a couple of months to run IP over it. Serendipitously or not, Trinity got in ahead. Trinity and IEunet because the router was owned by Trinity but the [connecting logic] board was owned by IEunet. It was a rather… strange relationship. I think that, sorry, I KNOW that was the first in academia; I know there were connections to BITNET before that, but they weren’t Internet connections. I suspect that someone like Alcatel in Galway may have had a private commercial connection, but I don’t know for sure. I always thought that there were some of these multinationals [that] may well have had their own private, but it wasn’t really public, well, it wasn’t available to anyone except company employees. So I think that’s when it was.

This experience, coupled with a deal with the Irish Unix Users Group (IUUG), made Michael determined to form a company to take advantage of this intriguing development, in co-operation with Cormac, who had been a postgrad and was working as a lecturer:

In 1991 we signed a deal with the IUUG to get members for them. For a while, the IUUG and IEunet were sort of inseparable. I mean, we made them inseparable when it was convenient! After a while it was clear that we needed to get proper buildings and things like that, so we formed a Trinity campus company, with some percentage owned by Trinity, roughly half and half I think. Cormac and myself did it in our spare time, which was anything but spare time, I mean, it was grossly subsidised by our employers. But it also brought benefits because it meant that the local organisations, DIT or Trinity actually got connected first. The bulk of [what we initially did] was email, and Usenet news. We also ran a gateway between UUCP mail and the X.25 network connection in the universities. So we got all of the universities connected to [the broader world of email] at that point. At this stage, the larger EUnet was very much an amateur organisation. I’ve got a list of meetings here and we always met at weekends. I think it’s a sign of amateurism…

At this early stage of IEunet’s existence, the Internet was still a mostly non-commercial space, and Michael has a vivid memory of:

[…] standing up at a RIPE meeting formally objecting that the Swiss networkNote: “the” Swiss network. blocked our commercial traffic. Because they felt that the internet was not a place for commercials. So they literally blocked our traffic. Happened in one or two places and has disappeared, obviously has disappeared. But it was monopoly days, you see, I mean, just thinking now, I felt we had a monopoly on [e.g.] Usenet news, which we brought in, but of course we didn’t, although we were the only people bringing it in at the time.Speaking of monopolies, there was a fascinating interaction between incumbent telecoms companies and this kind of entrepeneurial activity. Broadly speaking in Ireland, the incumbent telco, Telecom Eireann, did not go after startups during this period, although that was to change later. So Michael and Cormac found themselves mostly unmolested, probably because they were too small for a long while, and also had quasi-governmental support. It could easily have gone the other way, however.

Enter Alan Judge, TCD computer science student:

At this stage I was doing a Masters or a PhD and part-time managing mail systems on what was originally a DEC workstation of some form, which had four modems on top, and that was the entirety of the Internet in Ireland, back in the day. We had a small number of customers; companies like Lotus, I remember. The Lotus connection actually was interesting because the contact there was Jordan Hubbard,As of the time of writing, he had just left his job as “Director of Unix Engineering” at Apple to become CTO for iXSystems. who went on to become more notable in the FreeBSD world as a co-founder of the whole project, and then later [head Unix person] in Apple. [At that point], he was the tech guy in Lotus who cared about email.

The scene was set for a major milestone, however:

[…] probably the most significant event in the early history was in the summer of ‘91, Ireland got its first hardwired connection to the Internet, so TCP packets could actually [go back and forth]. I wasn’t involved at all, first hand, in that. I was doing a four or five month postgrad placement in the Open Software Foundation Research Institute in Grenoble but I know it came up that summer because we went from the point where I could mail people at home to the point where you could actually Telnet!

For context: back in the day, you used a Unix command called telnet to connect to remote computers where you had an account. It passed characters back and forth and was just like logging into a machine “normally”,That is to say, directly on the console or across a serial line, back in the day when normal involved things like that. except across a network.

At some point in the crazy growth of this Internet thing it passed from being a fairly humdrum and fundamental tool of Internet operations, to being something that only wizards did, to being almost comically obsolete and appearing on “Those crazy 80’s” TV programmes. But at this moment, it’s just Alan connecting to the TCD computers to read his mail. Unfortunately a small amount of doubts about the specifics of this first connection remain:

I’m not sure now, because I read the document that I sent you and it said it was a 9.6K line but I would have sworn it was a 19.2 line;I currently believe it was a 19.2kb/s line – Niall. [I guess] it might have been upgraded early on. It was from TCD to the University of Kent and Canterbury. It was operated by TCD, hence Ken Gordon’s involvement. They got a Cisco AGS, I’m pretty sure, with a big axial fan at the end, which was in a cupboard in Westland Row or somewhere. That was the Internet connection, and that’s basically the same era as when IEunet was set up. For a long time the two were interlinked. It was a campus company so it was all quite confused! I don’t recall anymore at what point the two became separate. There was a point where IEunet moved down to Westland Square and got its own premises and its own leased-lines but that was … a couple of years later. I orchestrated that move and by that point we had about a hundred-plus phone lines.

The relationship between Trinity and its on-campus companies was, as they say, not without friction. Bear in mind this was a substantial number of years ago. The predatory - or subservient, depending on your point of view - relationship between business and academia had yet to be established; academia definitely had the upper hand, and they were at least dogmatic and often actively unpleasant about defending their territory and their interests. At that time academic institutions were one of the few places in Ireland which had Internet connections at all, never mind good quality ones, so there was a compelling reason for IEunet to choose to be a campus company. Nick Hilliard, technical director of IEunet after Alan Judge left, remembers that being such a company required:

[…] TCD to be a significant shareholder. I don’t know what shareholding they ultimately had, but it was large, and they were troublesome.

TCD’s holding was managed by the director of innovation services, a post which was held at the time by [a man called] Eoin O’ Neill. To understand TCD’s position on its campus companies at time, I need to mention Iona Technologies. In common with all campus companies at the time, there was a running battle with Eoin about, well, basically everything. Eoin’s position was that because TCD was a large shareholder and these companies continued to exist only because of TCD’s good graces, that when it came to sell out time, TCD would milk it to the greatest extent possible. This had all sorts of consequences - inventors were paid a token amount for each patent, no money was available for incubation, everything was as slow as treacle, decisions [needing to be made were not made], and vetos were almost arbitrarily implemented on business plans. All in all, it was a nightmare.}

Nick relates the tale of how the Iona people became more and more upset with this until eventually they forced TCD’s (and Eoin’s) hand by threatening to close down the company unless TCD accepted their offer for buying out that shareholding. Of course, 6 months later, Iona received an investment from Sun which valued the company at a large multiple of what it had been, and TCD became determined that this would never happen again. Consequently Eoin redoubled his efforts at extracting maximum value and participating in every possible decision. With a partner that was adding substantial day-to-day value, this would have been tricky enough to manage, but apparently this was not the case, and tempers frayed both externally and internally. This apparently killed more than one potentially very large deal.

Nick again:

At one stage they had more-or-less worked out a difficult deal with PSI International,I’ve been unable to find another reference for this, sadly. and Cormac asked Eoin to drop over (he worked in the office across the corridor). When he walked in, his first question to the new American investors was: “So why should Trinity allow PSI to invest in the company?”. Perhaps a statement like that would have been ok if it had been one university selling to another, but this was the commercial world. […] After the meeting [PSI] told Cormac that the deal was off.

In common with many other companies at the time, it was recognised by the principals in the mid 90s that the insane growth rates were not going to be effectively met by the cash generated from organic growth, and presumably (then as now) the banks were not going to get involved. (Positively, anyway.)

Nick recalls that:

There was a lot of messing around at the time with the European EUnet people and what to do with EUnet as a whole; the UK operation had been subsumed into PSI already, EUnet Germany had just sold out to UUNet, and there were a lot of problems relating to the Polish operation, the Swiss operation, and several others.

This was given extra piquancy by EUnet’s new Managing Director, somewhat improbably (to Irish ears) called Wim Vink, who decided that the best way forward for the company as a whole was to perform a group merging under the EUnet International banner. Nick Hilliard suggested that Cormac Callanan may have been unhappy about it, although I have been unable to confirm this.

IOUnet

From Michael’s point of view, and as per a number of other stories elsewhere in this work, he is straightforward at acknowledging that they were good at what they were doing technically, but they weren’t good at business. Well,

[…] I certainly wasn’t. Even though I thought we were making money our accountant says, you’re not making money, or you’re trading insolvently, whatever it was. So this was coming up to ‘95 and we had someone come in and did due diligence. Unfortunately there was some crisis in the stock market and we were to seal a deal in the morning, seven o’clock, down in Fitzwilliam Square, some solicitor’s place - but some crisis in Wall Street hit the night before, and the whole thing collapsed. So then, subsequently [ISI] came in and took it. At that stage I started backing out because I had become director of Trinity and eventually Horizon Open Systems or Esat or whatever it was took it over. We were taken over, but quite satisfactorily taken over. But I certainly did not have business acumen.

Ultimately they sold out to Samir Naji of Internet Services Ireland, some while later and for a lot less money, which was implemented as a reverse merger. But although ISI had the money, IEunet had the customers and the cashflow and had pioneered the business to the point that it was obviously commercially viable. In an echo of the Club Internet story, EUnet Ireland owed EUnet Ltd lots of money due to crazy bad Internet costs for getting from Dublin to Amsterdam, and this formed part of the motivation for the sellout.

Unlike the other companies, IEunet had always concentrated on the business and commercial end, not the consumer end, so their cashflow problems, while considerable, were not the same order of magnitude of the other access businesses.

At this point, though, we are well into the early-to-mid 90s, and it’s probably appropriate to bring myself more thoroughly into the picture.