I love Rome

Date October 24, 2012

I found this today:

# hosts
# Generate standard hosts tables.
# Newnet doesn't really use it, it's just for easy reference.
# Neither does oldnet now, since it's doing DNS right.

It lives on line 327 of a Makefile used to build and distribute various configurations around our infrastructure. And it's old. "How old", you might ask?

Those lines have been in that file for 19 years. I know that, because I have primary historical evidence, as they say in archaeology.

When I asked about this "oldnet" thing, my boss told me to look up a document, Tenwen: The Re-engineering Of A Computing Environment. It's a peer reviewed paper from LISA'94 that chronicles the creation of the fabric underlying the network that I administer today.

In the paper, under Fixed Name Service (September 1992):

Once name service was working, we modified the Newnet hosts to do hostname lookups correctly.

This fixed the naming problems for Newnet, but Oldnet was still broken. While we wanted to limit the amount of work we put into Oldnet, we also felt that having name service working correctly could only make it easier to administrate Newnet. We therefore modified Oldnet to use the nameserver on Newnet, and quit using host files on Oldnet as well.

So, those lines date to September, 1992. I was 13 years old.

One time, a couple of weeks after I started working here, one of the other admins asked me what I thought of all of it. The infrastructure, the network, and so on. It took a couple of seconds for me to put my thoughts together, and I said, "Have you ever been to Rome?"

He had, so I continued, "This network is Rome".

He was confused, understandably, so I explained.

Some networks are like Phoenix, AZ. They're relatively new and shiny. They're brought up for the purpose they serve, and for the time in which they exist. There's not a building in Phoenix over 125 years old. It was built according to relatively modern methods, and the streets are largely laid out in a grid, because there wasn't a whole lot to get in their way. The city was started from essentially a clean slate.

Some networks, particularly in non-technological companies, are much more like Ephesus in Turkey. Ephesus is amazing...it's kind of like a metropolis the size of Orlando just got up and left. Seriously, I was there, and it was mind-boggling. It's almost exactly like it was 1500 years ago. There has been no modern development there, so when you walk through it, you're in the midst of miles of ruins.

Rome, on the other hand, is a different sort of beast. Romulus founded the City of Seven Hills 2,500 years old, if you believe the legends, and it has been occupied for pretty much the entire time since. When you climb out of an underground Metro station full of artifacts excavated while digging the subway tunnels you just left, you may find yourself facing an Roman gladiatorial arena turned into a Catholic church. The modern Rome grew up out of the old, and there's a continual architectural history of the past 2500 years waiting on you to find it.

It's the exact same way here on this network. The new Windows desktop infrastructure runs smoothly alongside the host database mentioned in the paper above, written 20 years ago. It relies on the host database, actually. Like Rome's streets, sewer, viaducts, and aquaducts, these antique systems were vital to the time, and if they aren't still in use, they paved the way for the modern systems that replaced them.

Being someone who likes archaeology, antiquities, and old libraries and junk stores, working here is a pleasure, because there's always something new to discover, and like I did today, a rich history behind it.

  • http://lonesysadmin.net/ Bob Plankers

    Great post, sums up many academic environments nicely. A few of us at the UW were just talking about this the other day, wondering specifically how many hosts on campus still have BITNET configured. We routinely posit that we have one of every network device ever made running on our network, to this day. :)

  • SPremeau

    Last semester a student project documenting the account creation process of an small computer science department discovered that the core of the process was created in 1986. It's been extended incrementally since then, but the core process that manage accounts and class groups are basically unchanged.