My Introduction to Networking and Unix

In retrospect I should have started blogging years ago, but I never seemed to have the time. It may have been a case of writer burnout from writing 3 books on the Internet in the early 90s and a number of technology columns from trade magazines at the same time. In any event here I am, attempting to blog and be relevant and up to date with Internet technologies.

I consider myself a 2nd generation Internaut. The first generation were the Vint Cerfs and Robert Kahns of the world who really did invent the protocols which drive the Internet today. The second generation folks, like myself, used and deployed and occasionally contributed to the technology. We were the early adopters. I was first introduced to the ARPANET (the precurosr of the Internet) in 1978 when I was still in graduate school studying Computer Science at the University of Texas at Austin. I knew nothing about networking. There was not even a graduate level course at UT at that time. I thought it was interesting, but my field of study was databases and I did not give a lot of thought to networking until a few years later.

I was also introduced to Unix at around the same time (1978). I found it interesting and elegant, but again I was not that taken by the technology. It was only later when I returned to UT to work in 1980 after a year working at Execucom Systems, did I get excited by both Unix and the Internet. I really became familiar with Unix when I worked as a research lab manager in the Computer Sciences Department. The department had a small PDP 11/60 running PWB Unix from Bell Labs. I was assigned to be the system administrator of the system. When I say this system was small, I do mean small. This machine had 64K of memory and 3 80 Mbyte disk drives. This was when I started my love affair with Unix. I also learned C at the same time from the "White Book" by Kernighan and Ritchie (http://en.wikipedia.org/wiki/The_C_Programming_Language_(book)).

I also got to play on the Computation Center's small PDP 11/70 system which ran Version 7 Unix. Since it was connected to the ARPANET, I also learned a bit more about the ARPANET protocols. This was right when the ARPANET was converting over to TCP/IP from NCP. Internetworking was emerging as a big technology. This was also when Ethernet was first being deployed at UT.

Some of my colleagues in the Computation Center were very interested in networking and how it could be deployed on the UT campus. UT was in the process of designing its own campus area network with its own set of protocols. Since they were standardizing the student computing environment around DEC VAX VMS systems, the core set of protocols envisioned to be used was based on DECnet. Unix and TCP/IP were considered too experimental and researchy by the powers that be at the time. One of my colleagues wrote a white paper to the administration explaining why TCP/IP should be used for campus computing. His argument fell on deaf ears and he moved on to other endeavors.

In 1983 the Computer Sciences Department was embarking on improving its own research computing infrastructure. They had acquired a large (for the time) VAX 11/780 plus several VAX 750 computers to support their research program. In addition a whole flock of Sun workstations was on order. I was still at the Computation Center, but I interviewed and was hired as one of the system and network administrators for this new environment. I was more interested in the Unix end of things, but I quickly fell into learning more about TCP/IP. I worked for John Quarterman who was hired as the system manager. He and I later formed an Internet consulting firm (Texas Internet Consulting). The two of us deployed the new VAX systems and also installed 4.2 BSD Unix on the systems. The Sun workstations came with an early version of SunOS which was also based on BSD Unix. We integrated the whole networked environment together. This was in the days before NIS or even NFS. We used rdist to copy around password files and software updates to all the workstations from the VAX.

We also embarked on some network evangelizing with the other pockets of Unix on campus. We later wrote up our experience as a paper called "The Yeast Theory of Networking" which was published in Sun Expert magazine afew years later. UT was still going down the DECNet path at the time. We decided to push TCP/IP in an organic way. We collaborated with the Unix group in the Computation Center and used our ARPANET connection as leverage to provide the campus with valuable services. We decided that two services which folks would find useful would be email and Usenet News. What we did was meet as a small group on a weekly basis and push these services around campus. DNS was also being deployed at this time and we used that as the naming service infrastructure. One of the first things I did was to register the UT domain name. I originally wanted to register bevo.edu (Bevo was the name of the longhorn mascot of the football team), but that idea was nixed and the domain utexas.edu was registered instead. Around this registration I developed the email system and the campus domain naming system.

At the same time, we connected the Astronomy Department (they were also running 4.2 BSD) to our nascent network via a SLIP connection using a dedicated internal phone line and two spare line drivers. With this connection we were able to push both email and Usenet News services to their computer system. We ended up replacing the SLIP link with Ethernet within about 6 months. Similar organic network extensions were pushed to other departments with an interest in connecting to the networked services we provided and managed. Within 2 years we handed over this organic network to the Computation Center and TCP/IP became the de facto protocol of choice for the UT campus.

It was an interesting and exciting time. I learned a lot about a variety of technologies. I essentially learned TCP/IP from the RFC documentation. I also wrote an RFC (http://www.ietf.org/rfc/rfc1027.txt). I also learned a lot about Ethernet. I did everything from install network cards to installing the OS and other software on the various systems we managed. It was fun and I learned a lot and got to play with a lot of cutting edge technology which helped me a lot later in my professional career.