History of the Telstra Research Laboratories

 

Computing

The Evolution of Computing
The Gerber System for Phototool Production
TACONET
Administrative Computing
Switching & Signaling Computing Facilities
Computer Facilities and Networks

“I think there is a world market for maybe five computers.”
Thomas Watson, chairman of IBM, 1943

“Computers in the future may weigh no more than 1.5 tons.”
Popular Mechanics, forecasting the relentless march of science, 1949

“I have traveled the length and breadth of this country and talked with the best people, and I can assure you that data processing is a fad that won't last out the year.”
The editor in charge of business books for Prentice Hall, 1957

“But what ... is it good for?”
Engineer at the Advanced Computing Systems Division of IBM, 1968, commenting on the microchip.

“640K ought to be enough for anybody.”
Bill Gates, 1981

“There is no reason anyone would want a computer in their home.”
Ken Olson, president, chairman and founder of Digital Equipment Corp., 1977

up to the top of the pageThe Evolution of Computing

The first computing devices used in the Laboratories were mechanical computers such as the Monroe Calculator, the CURTA calculator (http://www.hpmuseum.org/ffcurta.htm) and the slide rule. A cylindrical slide rule used by the first Director of the Laboratories, S. H. Witt, was displayed in the Laboratories foyer. Slide rules were a cheap and popular device and used well into the 1970's.

Electronic computing facilities at the Laboratories were established in 1963 with the purchase of a Control Data Corporation (CDC) 160A computer (http://www.cs.uiowa.edu/~jones/cdc160/) complete with central processor and paper tape input and output via an ASR35 Teletype. The processor had a core memory of 4012 words each 12 bits (bytes had not been invented yet), assembler and Fortran compiler. Subsequent upgrades added a line printer, two vacuum column reel-to-reel tape drives, a plotter and doubling of the core memory to 8024 words. This equipment was located at 10 Lonsdale Street, Melbourne (Block T/Taubmans).

Click for larger picture




160A Computer at the Laboratories 1969.
The console
operator is Norma Light
and the
other is Gillian Roberts, a
programmer.


Fortran was the main language used, but the CDC supplied 2 pass compiler had so many limitations and bugs that it was re-written by Arthur Thiess within the Laboratories, and a copy was given to CDC for the benefit of other 160A users.

The computer was located in the centre of the floor space on the second floor because of floor loading criteria and was reinforced from the ground and first floors to carry the weight of the system. A special room was constructed, complete with air-conditioning and false ceiling to house the system. Air-conditioning was rare in those days so staff inside the computer room were privileged.

The system components were lifted in during a weekend by a crane in a rear lane through a temporarily removed large window frame. Power poles in Lonsdale St impeded lifting from the front of the building.

One of the early tasks for the 160A computer was the design of special filters to enhance the group delay transmission characteristics of data lines used in the NASA space program. The Laboratories later received NASA recognition for their efforts.

One interesting use for the 160A computer was in a study of the amplitude statistics of a speech signal. The idea was to use a commercial PCM system to produce a coded stream of voice samples (7-bit coding, 8000 samples per second) and build an interface that would pull one channel out of the 1,544 Mbps stream and insert it into the 160A and store it on mag tape in real time. The stored samples would then be analysed later at leisure. The timing constraints arising from the combination of the 125 microsecond intervals between samples, the processor speed, and the start/stop times of the mag tapes meant that some tricks were needed.

  1. There wasn't time to analyse each sample as it arrived, so they had to be stored on tape.

  2. The tape couldn't start and stop for each sample, so samples had to be written in blocks of 1000.

  3. The CPU couldn't input samples to memory and at the same time write them to tape. Fortunately, the 160A had, in addition to its normal internal busses, a "buffer" channel which could be used for transfers between memory and peripherals. This transfer only needed a few machine instructions to initiate it and then the 160A hardware would automatically complete the transfer whilst the CPU was free to do other things.

Thus, the idea was to have the CPU read 1000 samples into a block of memory. Then, in the 125 microseconds before the next sample was due, it would initiate a buffer channel transfer of this block to tape. The CPU would read in the next 1000 samples into a SECOND block of memory (by which time the first lot were written to tape), then initiate the transfer of this second block to tape, and then read a further1000 samples into the first block of memory, continuing thus in this alternating fashion for as long as needed.

The software for acquiring the PCM samples onto tape was written in assembler. The software for the later analysis was written in Fortran.

The 160A computing facilities grew and in 1969 the facilities were moved to 31 Flinders Lane, Melbourne which became known as the Laboratories Computing Centre.

In the late 1960's, the Laboratories began investigating Fluidics, the technology of using fluids to propel items or to operate digital enumerating systems. Displayed at the 1969 Laboratories open day, was a Fluidic Binary-Decimal Decoder, demonstrating the potential for using Fluidics as a computing device.

In the late 1960's and early 1970's, programmable calculators such as the Hewlett Packard 9800 family (http://www.hpmuseum.org/) were becoming popular and many Laboratories Sections purchased these units. These calculators used credit card sized, flexible memory/program cards and some Laboratories staff were known to utilise the calculators to plot the price movements of the Stock Market in their "share club" which included the infamous Poseidon shares (on the rise).

In 1972, the HP 35 slide rule pocket calculator was released. It was the first calculator with transcendental functions and Reverse Polish Notation. These and subsequent models were very popular in the Laboratories.

The Computing Centre facilities grew, and by 1972, the Laboratories were also renting mainframe time from Control Data Australia, IBM Call 360 and Honeywell systems accessed via Remote Batch Terminals and teletypes that were scattered around the Laboratories. A Remote Batch Terminal is a card reader/tape-drive/printer/etc that is distant (remote) from the host computer.

In 1973, the Computing Centre purchased their first VDU’s and a Data General Nova minicomputer.

In 1975, the Computing Centre facilities in Flinders Lane, Melbourne were relocated to 22 Winterton Rd., Clayton (Block CN).

In 1978, the HP 9845 computer (controller) was released with a BASIC language and HPIB (Hewlett Packard Interface Bus – later became an IEEE standard). The BASIC language was optimised for measurement and data analysis. These units became very popular in the Laboratories for controlling instrumentation and peripherals. In the early 1980’s, the Hewlett Packard 85 was released and this also became a popular controller. By the early 1990’s, personal computers complete with HPIB interface cards were the favoured controllers and the use of HP controllers declined.

By the late 1970’s, many Sections of the Laboratories had purchased their own minicomputers such as the Data General Nova and Digital Equipment Corporation PDP. The Nova boasted a “Real Time Operating System” which was useful for laboratory experiments where data was collected in real time.

By the late 1980’s, many Laboratories Sections had purchased UNIX based desktop computers such as SUN’s and also created private LANs. Similar systems such as Apollo’s and SUN clones (Osbourne) were also purchased. A problem that arose at the time with these systems was host naming. Duplicate host names appeared and some names were considered offensive or inappropriate. Details of the subsequent battles between officialdom and the users about hostnames cannot be given here but it wasted many hours.

In 1998, Telstra entered into a policy deal with IBM (called “PC Lease”) to purchase all of Telstra’s Personal Computers and lease them back. The leasing costs included maintenance and automatic replacement every 3 years. Prior to this, the Laboratories purchased and maintained their own PC’s via an “Instrumentation” group that was disbanded in 1996. In a research environment, PC hardware and software configurations are continually changing. Unfortunately, the PC Lease conditions did not allow for continual hardware changes causing many problems for Laboratories staff.

up to the top of the pageThe Gerber System for Phototool Production

In March 1975, The Laboratories purchased a “Gerber Interactive Design System IDS 2” to produce phototools for printed wiring boards and later, thick and thin film devices and the occasional semiconductor. Prior to this , all phototools were manually produced using stick-on tapes. Originally located on the ground floor of Watkins Building (the shell of Watkins is part of Telstra's headquarters at 242 Exhibition St), it was an important tool for hardware development in the Laboratories. The system was based around three HP 2100 series processors, and was revolutionary for its time. The main processor had 48K of ferrite core ram and controlled two satellite workstations each having 4K of ram. The first controlled a green screen VDU editing station or a photoplotter table., the second controlled a digitising table where large scale PWB layouts on mylar film were transposed to the digital domain. Other peripheral devices were an ASR33 teletypewriter for "bootstrap" loading via paper tape, a paper tape reader and punch for drilling tapes, and a HP7970 reel to reel mag tape for archiving. The section also had an Excellon CNC drilling machine, to enable very fine drilling tolerances and therefore denser PWB's.

The Laboratories also had an in-house Printed Wiring Board (PWB) production facility, where a range of PWB's were made including multi-layer boards (This was situated in 1/M1). The two facilities worked closely together. The PWB production became computer controlled in the mid 80's using a Hewlett Packard system with variable board insertion and dwell times governed by the size and other characteristics of the panel. It controlled the actual plating currents in some baths and had large relay boxes to do all the switching.

The in-house printed wiring board production facilities were closed in 1987 and transferred to the Workshops in Dodd St South Melbourne (now the West Gate Freeway) and then to Virginia Park with the Excellon Drill. This closure, together with the increasing availability of PC based photo tool production and a reduction in hardware based development, led to a reducing need for the Gerber system and it was decommissioned in 1989.

up to the top of the pageTACONET

In 1976, the Telecom Australia COmputer NETwork (TACONET – the name from a staff competition) was created. At the Laboratories Computing Centre, a DATA100 Remote Batch Terminal was connected to TACONET and the original CDC 160A was retired (to the RAAF training school in Laverton Vic). The Data100 was a mini-computer system that emulated a General Electric GE115 remote computer to act as a remote batch terminal for a GCOS mainframe. Alternative emulations were available, for example, to connect to CDC mainframes like Control Data Australia's Cybernet. The same mini, with different peripherals could be used as a data entry (Key to disk/tape) system.

TACONET was a collection of closely networked Honeywell mainframes located in Dandenong Rd, Clayton. Melbourne and later in Sydney. They were accessed by Remote Batch Terminals over dedicated 1200 baud lines using synchronous modems for batch jobs. The Honeywell mainframes were actually not interconnected until 1981. Keyboard terminal access was provided by “VIP terminals”, a Honeywell acronym for Visual Information Projector. The terminals utilised a Honeywell proprietary synchronous protocol called VIP7700, connected on dedicated 4800 baud lines. Dial-up access was initially provided using STC roll paper teletypes using 300 baud asynchronous modems and later by various “glass” terminals (visual display units) using 1200 half duplex, 1200/75 and 1200/1200 full duplex asynchronous switched network modems.

In 1976, a number of terminal rooms were established around the Laboratories, each containing a number of VIP terminals connected to TACONET. The TACONET facilities were heavily used by Laboratories staff.

TACONET was intended to provide all the computing resources that would ever be required in the organisation. To purchase any device that was classed as a computer, a justification describing what this device could do that TACONET could not do had to be supplied to the bureaucrats. The Laboratories developed considerable expertise in wording such justifications.

In the mid 1970’s, the first microprocessors appeared. The Laboratories developed considerable expertise in developing systems using microprocessors and were seen as a leader in the field within Telecom Australia. A Laboratories staff member created Cross Assemblers for a number of microprocessor families that ran on TACONET. These Cross Assemblers were given the acronym ORGASM (OnlineRemoteGlobalASeMbler) which was not well received by management of the time. A Cross Assembler converts Assembly Language code to microprocessor binary code.

By the late 1980’s, the use of TACONET facilities by Laboratories staff was declining and internal support was reduced. TACONET was officially decommissioned by Telstra in July 1999.

up to the top of the pageAdministrative Computing

The late 1970’s saw the introduction of personal computers running the CPM operating system including software for word processing. Some Sections purchased these systems for use in document production.

One of the first CPM systems purchased was used by the Information Section of the Laboratories as a word processing system for publication production. It was called a “Comfile” and used twin 8 inch disks. It was replaced by a multi user system which had a huge 20Mb disk and 64K memory per terminal.

In 1978, one VDU per office desk was adopted as a principle in the Laboratories (as required) in preparation for staff performing some of their own administrative functions.

The first Wang systems arrived in the Laboratories in late 1984 and the Wang “VS” systems in late 1986. These systems were used by Laboratories staff for administrative tasks.

The WANG server (a set of Wang DOS based computers) was established in Block M2 (next to the Ground Floor PABX room) and provided “star” connections to desktop workstations and printers using a variation of co-axial cable called “twin-ax” for transmit and receive traffic. Later, an extension to server equipment in M8 was installed using “hard line” steel encased co-axial cable (certified to 750 MHz). The system used a network disc meaning that each terminal did not have its own disc drive but accessed a central drive over the network (later SUN systems had a similar system they called “diskless clients”). The WANG system was technically advanced for its time.

One WANG 300Mb fixed drive cost around $30,000 in 1988. The Laboratories assisted Wang to get TCP/IP working on the VS systems.

The VS system was a mini-computer with 900Mb of disk. The main VS was located in the computer room at M2 with a smaller system at 22 Winterton Road, Clayton (Block CS). The two systems were first connected via a dedicated 9600 baud synchronous line and later via the Ethernet link between the two sites. The operating system was Wang proprietary as was the “Wpplus” word processing system. The VS systems were installed throughout Telecom by Information Systems and interconnected the Laboratories VS system with TCP/IP allowing email transmission between the VS network and Unix and VMS systems in the Laboratories. The systems communicated via 9600 baud dial up synchronous links. All Laboratories staff had access to the system so they could interchange documents with other areas of Telecom.

The Accounting and Human Resources Laboratories staff accessed the Honeywell and IBM mainframes via PCs with special software that emulated VIP (Honeywell) and SNA (IBM) terminals. Local support was provided by CFN (see later) for these systems.

The WANG systems were decommissioned in mid 1994.

up to the top of the pageSwitching & Signaling Computing Facilities

In the early 1970's, the Switching & Signaling Branch had begun investigating and developing all electronic, stored program controlled exchanges. This lead to the development of the No. 6 Signaling field trial and the IST model exchange. The computer involved with the No. 6 Signaling trial was a BTMX (British Telephone Manufacturing eXperimental) processor. It had a 16 bit plus parity word size, 16K magnetic core memory and a 2 microsecond instruction cycle. The three IST processors were nominally ITT 1600’s, the 16 bit version of the ITT 3200s used in the 10C exchanges. They may well have been re-badged Honeywell machines, and were certainly very like the Honeywell processor (516??) used by the then British Post Office Research Laboratories in their experimental digital exchange “Pathfinder”. IST ran on duplex processors in 32K of core (real core memory!). When the system was enhanced by addition of the Remote Switching Unit (1974-78), the processors were modified in-house to support 2-word instructions and the memory was increased to 64K and digital cassette input was added. Originally the input was paper tape only.

Click for larger picture



No. 6 Signaling field trial components at Windsor Exchange 1972.
John Drake at TTY, Mike Hunter looking on.
Twin rack – BTMX processor & peripherals.
Half hidden on left – Paper tape reader & punch.
On right – Line printer.

The Switching & Signaling Branch computing facilities grew rapidly. The first PDP 11's were a pair of PDP 11/10s arriving in late 1974. Legend has it that they were justified in the budget as "intelligent disk controllers" for the ITT 1600's, i.e. a way of adding hard disks to the 1600's. In fact they were used for software development for the RSU, with a cross-assembler (developed by external contractors) and a Fortran-like higher-level language developed in-house. Similar dubious justifications were used by many Sections in later years to obtain specific computer equipment.

When the IST model exchange was closed and decommissioned in 1991, an ITT 1600 and a selection of the IST cabinets and RSU racks were given to the Telstra Museum.

In addition to the DEC machines, Switching & Signaling Branch acquired in late 1970’s or early 1980’s an ICL PERQ, which was donated. ICL had managed to persuade the UK universities to make PERQ’s their standard computer and had similar visions for Australia. However, they never made much headway here. It was the first Unix machine in Switching & Signaling Branch and was an advanced machine with a large high-resolution screen of portrait orientation.

In 1975, the Switching & Signaling Branch moved from 140 Exhibition, Melbourne (Block C/Cheneys) to 22 Winterton Rd., Clayton (Block CS).

In 1982, the Switching & Signaling Branch network connected (email only) to the Internet via Melbourne University (and hence the world) using a 1200 baud dial-up system.

In 1982, the Telematic and Message Services Section (TMS) established a multi-user Unix network, based on a Digital Electronics (DE) Unity minicomputer and Ungermann Bass (Ethernet 1) terminal LAN in Block M6. In 1984, TMS established a dedicated connection to ACSnet (Australian Computer Science Network) via Monash University for email and file transfer connection to the Internet at 9600 baud. It was around this time that the Laboratories domain “trl.oz” was created.

By the mid 1980’s, the Switching & Signaling Branch computing resources were the largest in the Laboratories and were used by many other Laboratories sections.

up to the top of the pageComputing Facilities and Networks

In 1985, the Laboratories Computing Facilities and Networks (CFN) Section was established to look after the overall computing needs of the Laboratories. The computing resources of the Switching & Signaling Branch were taken over by CFN and used as the basis for a centralised Laboratories computing environment. These facilities were located at 22 Winterton Rd., Clayton (Block CS).

One of the first projects was to provide a Laboratories wide switched 9600 baud “LAN” called RESNET, based on low cost hardware developed by Monash University known as MONET. The RESNET system had severe limitations and by 1989 a full Ethernet LAN had been established on the Monash site and RESNET was decommissioned.

CFN established a system for remote terminal access to the Laboratories computing environment in 1986 with the purchase of a “Secure ID” system from Security Dynamics. This system used a PIN together with a card to gain access. The card had a display with a number that changed every few minutes (PIN challenge system). To login, you had to enter the PIN and the number on the card. In the mid 1990’s, it was determined that the system did not fit the corporate computing standard and it was decommissioned.

In 1989, ACSnet became AARNet (Australian Academic and Research Network) and the Laboratories Internet connection was upgraded from 9600 baud to 64,000 b/s ISDN (switched, long hold call). Later, this was upgraded to 128,000 b/s (dual B channel multilink). In 1997, this was replaced by a Frame relay connection, with 128Kb/s guarantee (CIR) and burst rates to 2Mb/s.

In 1989, the CFN computing facilities were moved from 22 Winterton Rd., Clayton (Block CS) to Block M2 on the Monash site.

In 1991, CFN installed an Asynchronous Transfer Mode (ATM) system as the backbone network between each building on the Monash site. The network operated at 155Mbit/s and was one of the first ATM systems installed in Australia. Later, it was replaced with high speed Ethernet.

In 1993, the computing facilities were again moved to specialised facilities in Block M1. About this time, Telstra began moving towards a corporate computing standard using desktop personal computers, Microsoft Office products and Novell Netware servers. To comply with the Telstra policy, CFN purchased and installed a number of servers including a Microsoft mail server. Up to this point, all email to/from the Laboratories was SMTP based. The Laboratories Microsoft mail server included an SMTP gateway that, for a time, was the only one in Telstra (and managed by the Laboratories for many years).

The high end computing facilities supported by CFN included both DEC systems running the VMS operating system and SUN systems running Unix. As more development moved to PC’s and Unix became the favoured high end operating system, the DEC systems running VMS moved out of favour.

In 1996, it was recognised that Laboratories had some specific computing requirements in addition to supporting the corporate standard. CFN staff maintained the specific needs of the Laboratories as follows:

In 1999, CFN staff were offered tenure in a Telstra subsidiary company called Advantra (changed to Telstra Enterprise Services in 2002 and later merged back into Telstra) that was setup to service Telstra’s Information Technology needs. All CFN staff either accepted redundancy or stayed in the Laboratories.

By 2003, all of the Laboratories Information Technology needs were served by Telstra Enterprise Services who maintained a presence at the Laboratories to serve the specific needs of the Laboratories.

In 2005 TRL was closed.

See also:
> Closure of TRL. About 80% of the staff accepted redundancies and the rest obtained positions in other areas of Telstra.
> The Grepper A poem about a well known system administrator.

< Back to History of TRL page

 

Last Updated: Aug 2011