Computing Science Computing History
Computing Science Computing History(Mainly by Pete Fenelon)
Initially obtained PDP-11s for research, later acquired VAXes. Student users moved off Computing Service machine for academic year 1986-7.State of the art in 86-7:
Undergraduate terminals were 20 Visual 200s (VT52 clones) in V/044, plus an assortment of elderly devices in the labs. MSc students used BBC Micros with a temrinal emulator eprom in X/A block.
Exarch became funstar in 1987-88, bishop became the lab machine, minster became the main u/g Vax whilst remaining the dep't mail and news gateway.
For AY1987-88 the U/G terminals moved to V/058 and increased in number to 28.
Student usernames were of the form PeterF, IanR, etc. Potential clashes were solved by subsequent arrivals having more of their surnames left in. Electronics students used to initially be taught Pascal by Comp Sci, but used "reversed" usernames: PFrench IPayton, etc.)
There was an amusing confusion early in my first term when I saw Pete French trying to log in as PeterF instead of PFrench.... :)
Staff usernames were the usual mixture of initials, forenames, surnames, nicknames and whatever else people wanted.
The staff meanwhile moved onto a network of Sun-3 fileservers and workstations. Servers were:
To this were later added the imaginatively-named
For AY1988-9 the Vax maintenance contract became too expensive to justify keeping the students on them, so two shiny new HLH Orions were purchased, imaginatively named Orion1 (undergrad) and Orion2 (MSc).These ran a 4.2-alike Unix and were based around the Fairchild Clipper. The Visual 200s were replaced by DEC VT220s and migrated to the labs.
Student usernames changed to initials suffixed by a hyphen and letters starting at a for the first student to arrive with a particular set... e.g. pf-a, pf-b, etc.
HLH rapidly lost interest in building and supporting computers and became an Apple dealer. The Orions were incredibly buggy, though fast for their time.
Before the next major round of purchases, there was the Great Renumbering when all the machines were given addresses in York's Class B -- previously, they'd used arbitrary IP addresses.
The next major round of purchases in the early 90s saw the arrival of the "BM" concept. This was aimed at replacing all character terminals in the Department with bitmapped screens; however, Sun or similar equipment was perceived as being too expensive. Hence, inspired by the ideas of Plan9, it was decided to create a simple environment for providing a diskless workstation with a bitmapped, windowed display and local editing etc. talking to a more powerful "compute server" for little more than the cost of a terminal.
A large number of 386/16s were purchased, with (for the time) a good spec -- 4Mb of RAM, Tseng ET4000 video cards driving Interquad non-interlaced monitors, NE2000 ethernet cards.
As compute/file servers, four IBM RS6000s were purchased after a competition that also included DEC and Sun. These were:
It never really obtained a foothold with the staff.
A number of Sparcs had also appeared in the HCI and High Integrity groups.
With the adoption of JIPS, the Torch TripleX was no longer required and a reshuffle of Sun servers took place, the upshot of which was that Terminal was no more,. minster became a sun3 and johann became a more powerful one. I forget which box went where.
During the early 90s, the Sun fileservers started to become less reliable and some users began to demand local disk and a machine which could at least boot itself -- combined with the arrival of more powerful PCs a number of users began to buy high-spec x86 boxes; most of the kit bought as "BMs" continued to be used either as fast terminals or had windows and a number of TCP/IP installed on them.
The next major change in computer science was the replacement of ageing and by now unreliable Sun-3 kit with SGI hardware in 1994. A Challenge S fileserver and a large number of Indy workstations replaced the Sun3 fileservers and workstations for most of the department's staff. In the meantime, the High Integrity and Realtime groups had largely gon down the route of Sparcs, Macs and high-spec PCs. MSc students switched to a Novell Netware network for most of their work.
Undergraduate usernames started to be brought into line with Computing Service ones.
With the arrival of the SGIs the shape of the network started to change. Previously, there had been a "backbone" linking the fileservers together, with a separate "serving ether" for each machine's diskless clients. With one main fileserver replacing many , the serving ethers became more geographically based. FDDI to the Computing Service was installed, as was a very smart 3Com hub.
The HISE group by this point had its RS6000 plus two Sparc-10/51s, one of which continues to host most of the department's WWW pages :)
Charles Forsyth's interest in Plan9 saw a plan9 network arrive, with several important functions (dns, ftp, "official" www) moved over to this, as well as general file/compute service for a number of staff. The undergraduate PCs switched to plan9 from Orbit, though could also connect to a Novell/Windows service. The U/G fileserver Student1 was replaced by a Challenge/S (Atlas).
Related pages:
Computer history on campus
Folklore Index Page
Disclaimer
S020 Home Page
Send us Mail