CSIRONET Part1 скачать в хорошем качестве

CSIRONET Part1

CSIROnet has a 3-decade history, during which it has pioneered, innovated & achieved a great deal. In this multipart episode, we look at the foundation of Australia’s 1st Wide Area Network & one of the world's 1st. In February 1948, Trevor Pearcey had written an article in The Australian Journal of Science, entitled “Modern Trends in Machine Computation”, where he seemingly predicted the internet. He wrote, “It is not inconceivable that an automatic encyclopaedia service operated through the national teleprinter or telephone system will one day exist.” At the time, Trevor was in the process of developing Australia’s 1st computer (4th in the world) called CSIR MARK1 (later known as CSIRAC), for the Council for Scientific & Industrial Research (CSIR), soon to be renamed the Commonwealth Scientific & Industrial Research Organisation (CSIRO) in 1949. In 1958, Trevor Pearcey met with the then head of the Division of Mathematical & Statistics, Edmund “Alf” Cornish, in London, where they discussed the CSIRO’s computing needs. Alf suggested to Trevor that the CSIRO should have a central computing laboratory & would he be interested in setting something up for the organisation. Trevor agreed & returned to Australia to join forces with CSIRO Mathematician/Computer Programmer Geoff Hill, where they began a survey of CSIRO’s computing needs. By 1958, the CSIRO had “Divisions” scattered all over the country & they all had computing needs. Those physically close enough to access a computer were often near a University, but for everyone else, it meant travelling long distances. Pearcey proposed a central computing network based in Canberra, where a large, capable machine (a CDC 3600) would be based & several regional hubs (Sydney, Melbourne & Adelaide) with smaller machines (CDC 3200s). These four machines were made “compatible” so that work could be sent from the regions to the machine in Canberra, initially via a courier-based network. Scientists would send their punchcards physically via a courier service to Canberra, where they would be processed & sent back via the courier service. However, Trevor had always envisaged that eventually all machines would be “wired” to the network. The proposal was accepted & in 1963, the Computing Research Section (later the Division of Computing Research - DCR) was established. In Jan 1963, DCR appointed key staff members Terry Holden & Brian Austin, with many others following soon after. A tender was put out for the acquisition of the 4 computers & a local agent for Control Data Corporation put-in an unbeatable quote. The quote was accepted. The trouble was that the proposal for the smaller machines was a work of fiction & CDC had no knowledge of it, sending the local agent into a spin with CDC in the USA. Somehow, he managed to convince CDC to build 3 x 3200 machines for the regional hubs. After a late arrival, the computers were delivered, albeit lacking software to make them function properly.  Brian Austin was able to modify the CDC-supplied “Scope” operating system’s source code. Together with others, he wrote the DAD (Drums & Displays) operating system. The team were architecting a multiuser computer system, where a high-speed, persistent drum storage unit would be used in conjunction with Display Terminals connected to the 3600. Brian engineered the interactive use of the 3600 along with CDC engineers.  A “Job Stack” program was written to handle the submission of jobs for the 3600. Henry Hudson wrote code for a very lightweight text editor initially called FRED, then TED & eventually ED, which was crucial for the local network to run seamlessly. Henry Hudson also wrote code called “BreakIn”, which allowed priority usage of the 3600. Later, as Discs were added, Brian Savvas wrote code that ran on the 3200's, allowing “Random Access” read, write, copy, & paste to information stored on the discs. All these innovations were completely developed in-house. In 1967, Brian Austin & Peter Ewens started experimenting with an ASR33 Teletype machine & connected it to the CDC3600 via a PDP8S over a short line. He was successful in reading punchcards, sending the information to the 3600 to execute. The experiments continued with the team setting up in another building a kilometre away. They strung a line along the National Botanical Gardens fence, which borders the CSIRO's property, & connected the Teletype via the PDP-8 to the line & then to the 3600. Again, they executed code from the ASR33 & this was the first use of a remotely accessed computer on CSIROnet. In 1969, CSIRO realised they could connect all machines in its compatible, aligned “network” into an actual wired network across Australia, using telecommunication lines. And so, in 1969, CSIRO leased lines from PMG (Postmaster General’s Department, later known as Telecom & then Telstra), and used PDP-11s, driving "Node Code" to serve & receive packets, enabling Australia’s first, wired, Wide Area Network. Continued in Part2.