Header image  
Home Michael Cowie  
 Home  Home      

The Central Systems from 1986 to 1996

From 1979 to 1987 a mainframe ICL computer called an ICL 2960 was supposed to provide the central computing service to the University. By the late eighties it was looking quite incredibly ancient even to those working there at the time.

There was a point around 1980 where electronic things stopped being electro-mechanical, and started using far more solid-state electronics and the reliability shot up. It's very difficult for people today to understand why so much effort had to go into everything then, but everything had so many parts, and so many connections to fail.

Unfortunately this machine was bought in 1979 - just on the wrong side of the reliability boundary, and everything around it looked about a million times better. I've written some quite damning things about its reliability below. The things I write didn't happen constantly - but if they happened twice a week in a place where you were trying to get something done it was bad, and sometimes they'd happen several times in a day.

The computer had a punchcard reader and minor upgrades were still supplied on a box of punchcards as late as the mid-eighties! The reader frequently misread these crisp, new cards, occasionally firing some of the cards onto the machine room floor, but usually crumpling some of them like a concertina when it wanted to play up. Meanwhile the tape drives, visible at the end of the room, had a vacuum mechanism to hold the tape in place and if the loading of it went wrong the large tape viewing window filled from the bottom upwards with a big pile of writhing brown computer tape that looked like like demented spaghetti, and you couldn't open the door and try and retrieve what was left of your tape until it had finished sucking away at it. On one occasion one of the tape drives did this to a 12" tape that been sent here all the way from Australia, with one copy of valuable data on it. I loaded the tape myself to make sure that it was done right and carefully, and the drive still went beserk with it! Fortunately, after 15 minutes winding the tape back carefully onto the spool by hand, the second time I loaded the tape on a different drive it loaded properly and we got the data off it. None of the five drives were entirely reliable, although everyone had their lucky drives that seemed to be functioning well at that time.

In addition to 12 inch reel-to-reel tapes, the main storage was on 3 giant 177MB hard drives with removable platters for backup. The units were each the size of a washing machine, and the tops of two of them are just visible in the bottom left of the second picture. The platters, once removed, were wheeled on big trollies to University House once a week to safeguard the data, and replaced with ones that had just been copied to.

The whole computer took up most of a large machine room, and was painted in a white and a garish orange colour that things were painted in during the 1970s.

Two engineers from ICL kept the whole thing running by visiting very often. One engineer found everything very funny, and wasn't in the least bit upset when he couldn't fix the computer. The other guy was really good at fixing the machine, but struggled to stay calm when he discovered the bodges that the other guy had done to make it work again in his lighthearted manner.

By 1987, the ICL 2960 could only support 12 users, badly, and the service had been taken over by the emergency purchase of a couple of Digital VAX computers running VMS - the OS written by the people who later wrote Microsoft's Windows NT. The service was moved down to Bath University, using the new JANET network to connect. Bath still had an ICL service that wasn't a joke, and the one here was shut down, and there was a party on its last day.

The VAX computers were replaced in 1989 by Unix-based equipment - initially an innovative multi-processor design by Sequent, an American company called the Symmetry. This was really good for its day, and worked so well that there is very little I can tell you about it. It ran a variant of the Unix operating system, and it rarely went wrong. I wrote the backups for it using what I thought was quite an innovative scheme that worked like a clock using multiple interlocking sets of 4 tapes, with the weekly main cycle incrementing a 4 week cycle which incremented a 16 week cycle giving, automatically, backups that became less frequent over increasing time into the past for a full 64 weeks.

The Symmetry provided a consistently good platform for the University's central computing needs until the mid-nineties, when it was replaced by a number of servers that look like the modern ones.

The crudely made pano to the right is of the machine room as it was in 1996 with the Symmetry- the room was a real mess then because it had no one managing it, and the networking was also quite untidy. The scene shows the last Symmetry cabinets with a red 'S' symbol in a square on them, and there are Sun and HP servers running the main service at this point elsewhere in the room against the far wall.



Machine Room in 1986

The Computer Centre machine room as it was in 1986. I do recall that the photo records fairly accurately the operators' workload for a lot of the time. One or two of them were interested in how everything worked, and read the manuals and/or studied for OU courses, but the rest were always sure to bring a copy of the newspaper to work each day. How times have changed.

The machine room as it was in 1986

Bob Bird, one of the operator, dealing with printout from the 2960's main printer unit.Visible at the back of the room, the white VAX mini computers supported almost all of the University's computer users at this time.


Computer room at the end of the 1980s with operators Bob Salkeld and Jenny Bland. Main service now provided by the Sequent Symmetry behind them; backups still to 12" reel to reel tapes in foreground. The computer service was run pretty well then. The Systems Group had yet to be taken over by the Comms and Micros Group, later called the Technical Services Group.


Below: Computer room by 1996. The standard of structure and order in the room shows how the room had become a dumping ground, despite the fact that it was still filled with the systems that ran much of the University. The people running them just had to cope with the shambles.. The huge racks of 12 inch tapes hadn't been used for years, because backups were now done to small 'Exabyte' backup cassettes. The old tapes weren't archives, it just was no one's job to tidy up the room. The sectioned off room with the tinted windows was full of junk, and stayed that way for years too. A lot wasn't documented or labelled then; all the stuff behind the scenes was a mess, and only things that people saw were made to appear nice. The machine room was a secure room, and so no one was going to see what it was like.

Unfortunately your browser can't display this Java applet - maybe the browser has Java turned off?

Note: if you can't see this, it may be that the Java console may have been turned off in your browser. If so, you'll need to turn it on. If this pano doesn't work with on-campus managed computers, it may not let you install Java if you haven't got it.