Some History of U/A Networking and the NetMgrs Group

 One thing to be noted as you his read history: you'll see time after time that it was department units leading the way with new technology, not a centralized, University bureaucracy.

The original "network", such as it was, consisted of Control Data Corporation (CDC) remote job entry (RJE) stations in various places like Optical Sciences, Engineering (the building by the Student Union), and probably other places.

These allowed people to submit card decks to be run on the CDC-6400 main frame, and get their printouts back at the RJE site. Those locations were later upgraded to Digital Equipment Corporation (DEC) PDP-11 based RJE stations which also supported asynchronous terminals. These stations were connected to the DEC System-10, running DECNet (or Advanced Networking Functions (ANF)-10). The home built inter-computer link allowed one to use those terminals and RJE facilities to submit batch jobs to the CDC mainframe as well as run interactive sessions on the DEC-10.

In the meantime, the UCC acquired a Gandalf Private Automated Computer eXchange (PACX). It connected asynch terminals to asynch computer ports. This allowed many more terminals than ports, at a lower cost per connection. Since most terminals weren't in use all the time, this was a big win, as you had a hardwired terminal (at up to 120 characters/second, phenomenal speed at the time), and could generally get connected to a computer by pushing the "Break" key and entering a few keystrokes.

The PACX maxed out at 512 terminals and 256 computer ports, couldn't make certain combinations of connections, rendering some port/terminal pairings undoable, even if both ends were free, and turned out to be a flakey beast also. It was phased out in favor of the M/A-COM Linkabit IDX-3000. Unlike the PACX, which used Gandalf line drivers over a 2 pair dedicated circuit to each terminal, the IDX ran a T1 signal over the same 2 pair, and used a multiplexer to expand that out to 24 serial ports at the far end. The T1 signal could be repeated, allowing terminals to be placed in out of the way buildings like Kaibab-Huachuca dorm.

The IDX-3000 supported 128 T1 circuits, which worked out to 3,072 ports which is where its name came from. It also provided direct T1 connections via DMZ-11 boards to PDP-11 systems, such as the DEC-10 communication front end systems and DEC VAX systems, reducing the footprint and power requirements for computer room communication equipment. Even the IDX ran out of ports, so we had a dual IDX-3000 system eventually. Some juggling was done to avoid cross connections since that was just wasted capacity.

In the early 1980s Steward Observatory & Lunar and Planetary Laboratory (LPL) bought a 12 strand fibre bundle and about a dozen grad students & computer staff from Steward & LPL installed it in the tunnels. The steam side was chosen because it looked less threatening than the high voltage side and neither FM or UCC had any advice to offer. The IDX lines were strung along the tunnel ceiling.

Because commercial hangers and radius' were judged be to flimsy for the price (given that people with very large wrenches would be working near the fibre), craftsman machinists from the LPL machine shop custom made them so the fibre cable (multimode, 62.5/125) would not "kink" in the tunnel system. Volunteers held, dragged and supported the cable while it was snaked through the tunnel system and affixed (something akin to a Chinese New Year's celebration). The fibre cable was dragged and installed with the fibre pairs already terminated (termination quotes were over budget).

Service loops were left mounted in the tunnel sumps on garden hose hangers rather than re-terminate. The flimsy hangers were fine in those protected locations, and remain undamaged.

P/TCP wasn't run on the fibre but uucp was from at 19200 baud ... so it was a "fibre network" of sorts.

The various VAXes were also connected by DECNet from various campus locations. There was a VitaLink satellite link to the "John von Neumann" supercomputer center in Princeton, a.k.a. JvNC. The 56kb/s link from CCIT to JvNC was installed in 1987 or so. It was actually a link from a dedicated VAX-11/750, named JVAX, which became an honorary part of the CCIT VMS VAX cluster. JVAX was even on Bitnet, with Bitnet name ARIZJVAX; it was connected to Bitnet via the much larger RVAX.

The original plan was for each faculty user to be given an account on JVAX. That was before the campus Ethernet amounted to much.

Five DEC LanBridge-100's (2 port 10mb switches but about 4U high), connected to a common roll of cable (1 meter between connections) under the floor on one side, and the other connected to a copper-to-fibre adapter to make the run to the 5 original Ethernet connected buildings.

Charlie Glickman and the Engineering folks led the broadband contingent on campus. (There was a time when the winner of these two technologies, data over broadband vs. old style Ethernet was uncertain.) They brought broadband to LPL about that time. Charlie gave LPL a couple of serial ports running over the broadband as well as video. Charlie's coax network extended to the AG buildings as well. Charlie and Robert MacArthur spent a whole summer in the tunnels with a little TMK TV testing the signal in various places as it got pulled.

UCC lashed their fibre to the Steward & LPL fibre as the TIPs wiring replaced departmental PBXs. The fibre bundle ran from UCC (current CCIT building) to the Steward machine room. One pair was used to connect a shared MV10000 computer. Other pairs were jumped to LPL, Optical Sciences, and NOAO, making the Steward Observatory machine room an early interconnect-hub, in a passive sense.

In 1986 LPL installed a thicknet (10Base-5) Ethernet backbone that included Flandrau. The Planetarium link was replaced with fibre in the spring of 2003.

When UCC was being reorganized into CCIT (the reorganization occurred in 1986, give or take a few months), the University Telephone dept was removed from FM and dropped into CCIT. From UCC, and other data processing divisions, came the data communication folks. From this mixture, Telecom was born. Dan Roman was the first person placed in charge (and by fate, was the last person in charge).

Dan who formed the first Telecommunications Advisory Group (TAG). He asked every person he knew who had a beef with the old UCC establishment to attend and offer advice.

Shortly afterwards, the new head of CCIT, John Monnier, went out for a permanent Director of Telecommunications (now that the organization was defined and staffed). Bob Leach got the job.

Bob attended his first TAG meeting, turned white, and never returned. A short while later he formed his own advisory group (buy replacing the current TAG members mostly with crony's) and things were never the same again.

Bob's group were at a higher administrative level and more sedate, but they for the most part were not in the trenches and were unable to provide the detailed feedback and advice that Dan wanted - and needed. (Dan was now 2nd in charge of Telecomm). Dan formed his own advisory group again, being careful not to call it the Telecomm Advisory Group, least Bob take offense.

A whole bunch of "Official" Advisory Committees got formed. One of which was the Telecom Advisory Committee (TAC) and one of the big issues we got to discuss was "How do we pay for the data network?" Sound familiar? The NetMgrs were quite involved in that discussion. We gave the same answer back then as we should be now - pay for it like electricity on this campus. After all your computer needs both!

The first NetMgrs meeting occurred around the second week on November, 1988 (It was not called NetMgrs before that). The occasion was marked by three events. The Robert Morris worm of the previous week on Nov 2, comfortable chairs and booths at the old Big A and beer.

There was a group of approximately 6 people who met to discuss what we were going to do about protecting the campus network, because of this worm thing. We got together to solve technical problems, like how do we subnet campus (something that, in those days was unheard of and difficult to do).

Thus, in the mold of the "VAXmgrs" informal group of VAX managers, we started doing "NetMgrs" in the same way (most NetMgrs were VAXmgrs at the time).

The main reason that NetMgrs was not an official U of A organization or decision-making body or policy builder is that there was a huge amount of distrust between the three main communities involved. There was UCC, CS, and there were all the other departments who had networks (MIS, LPL, Engr, Math, AG, and perhaps a few others). Being unofficial avoided answering the question no one wanted to ask: "What is the authority of this committee?" but also solved the problems everyone needed to solve: "How do we coordinate & manage a campus-wide data network?"

We had just come out of an awful era where UCC had dominated the central computing structure on campus. As part of it, they went where they wanted to go, did things the way they wanted to do, and then charged mightily for them based on their own budget and empire. The summary was that all the departments which needed to push computing out to faculty & students, especially undergrads, could only do it on terms acceptable to UCC.

Computer Science (CS) was the exception: they had always taken their own path, for obvious reasons, and were not beholden to UCC. But all the other departments who were trying to kick off their own computing services disliked UCC because they didn't listen, couldn't make services available at reasonable cost, were slow moving, autocratic and dictatorial, etc. In the early 1980s, departments started going out for their own computing via grants, decision packages, etc., not necessarily because they wanted to duplicate effort but because they didn't trust UCC to offer the services - a distrust well founded in prior experience.

Then the network came around, and the Internet, and suddenly there was a shift in needs and requirements. The network would have been most easily managed in a centralized way, but no one wanted to cede UCC control of it. CS had the first Internet connection, which everyone wanted a piece of. UCC had the Bitnet and JVNCnet connections, also very desirable. And, of course, the carrot for cooperation: a huge inter-building physical plant which they were putting in under the TIPS name. It was originally designed to push IDX (RS-232 terminals) out to every office, along with phones, of course, but by a massive stroke of luck turned out to also be 10BASE-T compatible (thank goodness!).

Now, while no one really trusted UCC, we didn't regard them as a monolithic entity. So while the folks at the top were considered politicos and budget grabbers, the folks at the bottom actually doing the work were considered to be trustworthy and trying to "do the right thing." Because of the management structure and there was always the risk that UCC would go back to its "old ways" of doing business and yank everything around at any moment. So there was a necessity to get together and coordinate and agree on things, not just IP space but LAT groups and AppleTalk names and DECnet areas too. The problem was that no one was willing to allow UCC to have an official committee to do this, because that would be tantamount to ceding control of the network to them from the start. Over time, people have begun to trust UA Telecommunications and enjoy more benefits of centralized coordination and purchasing and distribution of data services.

Tunnels

Tunnels connect most all campus buildings together for utilities. One wall usually has water (steam, chilled, drinking, and sewer), the other, high voltage lines. It's a big grid-work with on/off switches for lights at either end of main drags. Tunnels vary from new, wide, and dry, to old, narrow, wet, and dank. Brightest colored thing in there is the colorful fibre cable, which was run through orange interduct.

Basically the tunnels are poured concrete, square, "tubes", but some newer ones are now just galvanized pipe placed into the ground. They are big enough to walk through, but you don't drive a golf cart like at the Pentagon.

Most all buildings have basements with "doors" onto the tunnels. Some doors are something akin to jail cell doors, i.e., spaced bars with a big lock. Often doors to building basements were unlocked. Some large basement areas of buildings were still loaded with rations from the old civil defense days.

Some time ago a lot of the high voltage lines were replaced, before that you were careful to keep your hands at your side. Of course, once in a while there was a leaky steam pipe on the other side which rudely reminded you to keep your distance here as well. Can't see a small steam leak, but you knew it quick.

In certain places, the tunnel has cracks and water leaks in. Some of these cracks were caused by contractors shooting support bolts into the ceiling. FM is rightfully protective of this facility.

While walking in the tunnels to building 43 where the old phone switch was kept. In those pre-electronic days, you had mechanical dialing devices, making mechanical connections for users. You could begin hearing the clackity-clack-clack of the switch, 100 yards away. As you approached, it became more and more intense. You could peek into the basement and view the operation. Definitely spooky.