I first came into real working contact with IT in college in 1980/81. Studying applied mathematics at NIHE in Limerick, a required module was computer programming. We learnt four languages – Basic, Cobol, Pascal and Fortran – and worked on two different systems – an ICL 2904 and a VAX 11/780 from Digital Equipment Corporation (DEC). We programmed the 2904 via punchcards. Believe me, this helped improve one’s coding syntax discipline. I found I could code easily and quite liked it. I also enjoyed learning about the computing environment on which the software ran.
Second year students were required to do work experience and I ended up being hired by the IT department in the college. The original library system in the college had 5-digit book codes. I guess they never thought the library would exceed 100,000 books. They were rapidly running out of book numbers and would not be able to add new books to the library system. I was asked to rewrite the Cobol software to support 7-digit codes. The problem was the library system ran on Cobol on the ICL 2904 and I was not about to undertake a task this big on punchcards. For efficiency I needed to get both the code, and later the data files, to the VAX for editing and then back to the ICL for compilation. I really wanted a network, but I didn’t have one. I ended up using mag tapes, but from that point I understood clearly the potential value of networks and, in particular, of internetworks for systems from different manufacturers.
In 1987 I ended up in the Netherlands building DECnet networks between VAXes and other network-enabled devices. I learned to configure DECnet and began to grasp the concepts of networking.
In 1988 I was back in Ireland again working as a systems and networks manager for S3, a silicon design subsidiary of Philips Nat Lab (Natuurkundig Laboratorium). It was founded and part owned by Professor Maurice Whelan of Trinity College. The company used a VAX for design analyses and a token ring network for chip design, running workstations from Apollo Computer. Shortly before I arrived in S3 some engineers from the Netherlands had added a second Ethernet card to the VAX and an Ethernet card to one of the Apollos, bridging the two networks. They also bought a TCP/IP stack that Novell had produced for the VAX. By default the Apollos ran TCP/IP as their networking protocol.
It was my first time to see TCP/IP and the help pages on the systems were the only reading material I had. The connection was already in place before I arrived, but I had to figure out how to get the two networks to talk to each other. In doing so we configured what may have been the first TCP/IP routers in Ireland, one on a VAX the other on an Apollo – a small start but I had mastered gated and host tables. The value of joining different networks together was something I understood from my college days and it was a powerful idea to move freely across systems from different vendors.
I began researching the TCP/IP protocol. From an industry magazine I learnt about a company called Wollongong who did a co-resident TCP/IP stack for a VAX (meaning it could use the DECnet Ethernet card along with TCP/IP) which would be a big saving.
I read in a chip design magazine about a company, Mips Computer Systems, that had just released a new RISC (reduced instruction set computer) based processor. It could run at speeds up to 100 times faster than a VAX. More interestingly they also brought out VAX-compatible compilers for both Cobol and Basic. The Mips systems ran Unix – something I now understood reasonably well from my work with the Apollos – and TCP/IP.
I was aware that the VAX installed base in Ireland was large. I could see that if you could connect VAX systems and networks to Mips systems via TCP/IP, and compile the VMS source code to run on the Mips, you could run your software much faster. This had to be a very interesting business proposition, in particular to companies with time sensitive computing requirements.
In S3 I dealt with DEC resellers and other suppliers of VAX-related products. Tom Moynagh was a founding partner in a company called Parity Modules. Parity had just been sold and he had come in for some money. I told him about my thoughts on TCP/IP for VAXes and how we could connect to the VAX-compatible Mips. We agreed we would set up a company to market this technology in Ireland. I thought that X terminals (network-only Windows machines) would also be big. So we called the company SysXNet – systems, X terminals and networks. (We later changed this to just Sysnet.) I persuaded two college mates to join me on the engineering side. Declan Mulcahy was also with S3 and had got me the introduction for a job interview there. Mike Ryan was probably the best programmer I knew at the time. Tom Moynagh asked Paul Shields to join us as a salesman and thus in 1989 was launched the first TCP/IP integration firm in Ireland.
Tom Moynagh had lots of contacts and we hit the ground running. We had many early wins but sometimes the leap forward in technology was too much for conservative customers to trust. Nevertheless, on the whole, business was going very well.
Technically during this period we did not discover too much that was new. I think I got a commercial TCP/IP stack going with DOS, but most of the problems had been cracked. It was mostly about implementation. Business-wise we discovered that the X terminals technology was a lame duck. Ironic when I think of the success of Chromebooks now.
The three engineers were not happy with with the direction of the business and decided to leave. Declan went his own way and Mike and myself founded BakerRyan in late 1991. I worked mainly on VAX/VMS and Mike on Unix.
Having launched the first Irish TCP/IP integration company with Sysnet, we were constantly looking out for competitive products. The Unix market was also starting to explode at that time. We decided to take on TGV Multinet for VMS (a competitive product to Wollongong) and also to become a System V reseller with a product called Interactive Unix. We also sourced a free TCP/IP stack for DOS and later deployed the Winsock programming interface. Many commercial packages for Windows systems came out at that time, including FTP Software’s PC/TCP, NetManage’s Chameleon and Hummingbird. We stayed with the free code because it worked well. The market was exploding and new TCP/IP integration businesses were forming almost daily.
Around that time the internet proper was starting to take hold. The first supplier was IEunet in Trinity and I think BakerRyan was their first or second commercial customer. I turned a VAX into a router that could do dial up on a serial line and run a protocol called SLIP. We had a connection in no time and from the start it worked smoothly. I remember having to sign a long ‘Terms of Use’ agreement. It said, among other things, that we would not use the internet for commercial purposes. Before the internet connection could happen we also had to get our IP address allocation and this involved an application to RIPE in the Netherlands. It also required a domain name and this involved getting ourselves registered with the IEDR in UCD.
Once we got on the internet we had to start to master protocols that were not required in LAN-to-LAN integration. SLIP was one, of course, but there were also things like DNS, Mail Daemon and others. Business was good but not great. A lot of the more technical customers from Sysnet came with us, including the Valuation Office and the Department of Enterprise and Employment. These were the very first government bodies to connect to the internet (as far as I know). Gilbert Storrs at the Valuation Office kept the first email exchange and sent me a copy years later.
It’s good to have the time stamp. This was from a Dell system running Interactive Unix. Shortly afterwards Eamon Lairde in the Department of Enterprise got me to connect to the internet from a VAX. Both organisations accessed the internet through IEunet, because BakerRyan had become an IEunet reseller. It was also during this period that we started playing with Mosaic, an early web browser.
The Central Information Technology Service (CITS) at the Department of Finance advised other government departments on computing strategy. It was very resistive to the deployment of the internet and equally resistive to the use of TCP/IP on LANs.
I recall on one occasion being asked by Eamonn Lairde to give a presentation on TCP/IP to the heads of IT for various departments and government agencies. CITS representatives were there and ridiculed TCP/IP, saying it would never catch on. I argued that the internet was an unstoppable force and that LAN protocols would never scale. With Eamonn Lairde I encouraged them to look for a Class B internet address, which was then still available to agencies of the size of the Irish government. They never took up the suggestion – a decision, I believe, that cost the state over the years. Fortunately the government departments had sufficient independence for those with vision to make decisions and get things done. When it became obvious the internet was here to stay, most departments were well positioned.
At the end of 1994 Mike Ryan and myself decided to go our own separate ways. Mike went to do Unix consulting (with Sun I think) and I set up Baker Consultants Ltd (later Baker Security and Networks) in January 1995. I kept the VAX Multinet product and all the networking business. For the most part I dropped the Unix end of the business. At this stage the internet was starting to explode in use. Baker became an internet provisioning partner for PostGem. We were also a DNS provider and have remained a boutique DNS provider to large organisations with more complex DNS needs.
In 1995 we got our first leased line to the internet. It was an M1020, which was little more than a permanent dialup at 19.2K. We also set up our first web page. I believe this was one of the first commercial web pages in Ireland. I did not keep records of any of this, so the exact date and time are lost to me.
Ireland On-Line was starting to gain traction and there was a lot of internet business opportunity. We stayed away from anything to do with modems (or consumers). However we sold a lot of leased lines to enterprise customers and would manage the service (firewall, routers, etc.) for them. When Esat Telecom bought PostGem in 1999 I believe we represented about 30% of their commercial business.
Internet provision was now becoming a commodity and we were never going to be a commodity player. We wound up our internet leased line business and focused on firewalls. I was joined by Claudio Corrocher in Baker and we took on the Lucent firewall business. Until then we had been using VAXes as firewalls (or hardened routers). The internet was becoming a more threatening place and more specialist protection was needed.
We rolled out a managed firewall service based around the Lucent Brick. This was definitely the first centrally managed internet security service in Ireland. With Claudio’s guidance we won many large accounts, including Hilton Hotels International, Kingspan and Enterprise Ireland. Soon we were managing firewalls in over 100 countries.
I may be wrong, but I believe we are still the only firm in this area and from this era that is still trading as the same entity. Baker Security and Networks has always remained focused on TCP/IP networks and the management and protection of same.
Last edit: January 2016
© Robert Baker 2016