Roger Clarke's Web-Site

© Xamax Consultancy Pty Ltd,  1995-2024
Photo of Roger Clarke

Jenny Sinclair's Anniversary

The Network Anniversary

Jenny Sinclair

The Age, Melbourne

© The Age, 1999

This article was published in The Age on 22 June 1999

This document is at

See also Roger Clarke's Brief History of the Internet in Australia

The Internet first came to Australia 10 years ago with a fragile link from Hawaii.

It started with a ping ... Ten years ago tomorrow, or the day after, depending on what time zone you're in, Australia was connected to the Internet by two men at either end of a fragile, convoluted link between the universities of Melbourne and Hawaii.

The man in Melbourne was a technical staffer, Robert Elz, who was already the honorary custodian of Australia's international computer-based communications.

In Hawaii, Torben Nielsen, an academic with a dream of a Pacific network, had been given cash by NASA to set up Internet links with countries around the Asia Pacific.

The subject line of what was probably the first message into Australia was very much to the point: ``Link up ...'' it said. As well as being the start of something big, that message was the culmination of years of planning and diversions into technological dead-ends by the computer scientists and technicians in Australia's leading universities and the CSIRO. Other Australian networks at the time included ACSNET, a University of Sydney-designed network that relied on intermediate computers to forward mail; Spearnet, a proprietary Queensland/New Zealand system; and CSIRO's own network as well as a handful of public commercial networks like Telex.

Internationally, Europe's universities had at least five disparate networks with names like RARE, ROSE, SURF and JANET.

The universities and CSIRO had struggled with the communications problem for years. In a 1988 report to the Australian Vice Chancellors' Committee (AVCC) that reads like a document from the distant past, Dr Brian Carss wrote: ``Mail as the traditional mode of communication is proving too slow and unreliable in areas where developments are occurring almost on a daily basis. The telephone too is proving unsuitable ... because of its cost and lack of permanent records ... recently FAX (sic) has been widely introduced and its acceptance in the business sector is evidence of the need for rapid communication ...''

The reason the first Australian link came to Melbourne University was a combination of history and the presence of Robert Elz, according to Elz's then boss, Emeritus Professor Peter Poole.

Poole was the university's first professor of computer science. When he got the job, he was allowed to buy ``a computer''.

He chose an Interdata 832, with one megabyte of memory, worth $200,000.

At about the same time, Poole hired Elz, a ``brilliant'' student of information technology and law.

Poole says Elz was one of a group of technically-minded Australians who introduced the Unix operating system here, even contributing to its development. The Interdata 832 was running on Unix by 1977.

It didn't take Poole and Elz long to discover and use the Unix-to-Unix Copy Protocol (UUCP) command, which sends a file to another Unix machine.

The only other Unix machine they could find was at the University of Woollongong, so a link was set up and file transfers started over a telephone line.

Poole had been working in the United States and England, where network technology was more advanced.

``The first time I sent e-mail was in 1972,'' he says, and he was excited about the potential of this rather clunky technology.

Poole sent Elz to a US meeting where he visited the Digital Equipment Corporation (DEC), which was running a network called DECNET. A link was established between Melbourne's and DEC's machines over a phone line paid for by DEC, allowing other Australian universities to send files to the US via Melbourne.

That machine was linked to another in the Netherlands, which in turn had a link to the United Kingdom; a message sent to the UK would have to specify each machine along its path and could take days to arrive. But it was a vast improvement over the postal system, and bulletins about conferences and technical papers began to circulate freely.

``There was even software ... there was a rudimentary e-commerce,'' Poole says. Users would take copies of software and, if they liked it, would pay for it, much like shareware today.

There was also pornography and online gaming, but in a form so text-based and makeshift - composite images printed out on line printers and word-only games - that they seem quaint beside the X-rated streamed video and 3D color gorefests you can find online today.

For a while, DEC paid for the link, but ``eventually the load got so much that they said `hey, you guys have got to stop making phone calls','' Poole says. ``So we had to do it some other way.''

Elz stepped in to run the mail and news transfers, billing calls back to each university. Mail and news was sent over an X.25 packet network, and bills for the calls reached up to $100,000 a year.

But that was a stop-gap measure, and the establishment of a universities network called AARNET was soon being considered by the Australian Vice-Chancellors Committee and CSIRO.

That network, which became today's Australian Internet, was kick-started by a mere $800,000 from the Australian Research Council, plus funds gleaned from the universities' budgets.

Elz says that if there hadn't been money available from NASA for the Hawaiian satellite link, a far cheaper, slower Internet Protocol link into the UUNET (now the Internet service of MCI Worldcom in the US), would have been used. But that would have been so slow as to be not really any different from the existing X.25 link.

The first link ran into a borrowed Proteon P4100 router (borrowed from Hawaii) and a Sun 4/260 server in the machine room at the Richard Berry building, at the University of Melbourne's Parkville campus.

Apart from a mis-assembled cable, the main problem was in technical issues of timing and getting the 56kbps stream (carved out of a 64kbps link) to match the US 56kbps stream.

When the link did get up, overnight on 23/24 June, Torben Nielsen was able to log directly into the Sun machine (named ``munnari'') and write his ``Link up ...'' message to Robert Elz as a munnari user.

Geoff Huston, the first manager of AARNET, says the network aimed for speed first, and that turned out to mean using Internet protocols only.

``We made some harsh decisions. decision number one: we're not going to do voice ... a lot of us wanted to, but voice was out.

``Fax was out: don't do fax, don't do SNA (an IBM networking protocol), don't do OSI. It really was speed was the essence. We were reducing it down .. and trying not to get diverted into agendas that are just open-ended.

``We're doing data (TCP/IP, Decnet and X.25) and then pretty soon it was `If anyone wants us to do anything other than TCP/IP, put your hand up right now - oh good, not a problem'.''

ACSNET still exists in a sense, retaining its domain name of Nominally about 360 nodes on the network still remain and messages are still sent on it, mostly over Internet connections.

Professor Poole says the new system had to use whatever was available.

``There wasn't a cable in those days from Melbourne to Hawaii,'' he says. ``So it went over a cable from Melbourne to Sydney and then via satellite to the west coast (of the US) and then back to Hawaii (over a cable).''

For the first six months or so, the link remained in the care of Melbourne University. Then AARNET took over, but the physical link to the Internet terminated at Melbourne for at least six more years.

The development of Telnet (remote log-in software) and Gopher (software that classifies and helps organise Internet resources) allowed researchers to start making remote queries of other machines.

Huston says commercial researchers were the first outside the academic community to cotton on to the possibilities of the Internet and librarians were the first non-scientists.

``The Bureau of Meteorology, BHP research labs, Telstra Research labs were very early adopters.''

``They came in very quickly because their research folk worked very closely with the universities. They said `we want to talk with you, we're sick of sending you faxes'.''

And people being what they are, this powerful communication tool wasn't restricted to hard science even from the start. Elz says there were always social applications on the Internet - in the early '90s, a radio call of the Melbourne Cup was ``broadcast'' and, very early on, there were text descriptions of Test cricket matches in Australia.

After early ISPs like Connect, Western Australia's Dialix and the now-defunct RUNX in Sydney set up, and browsers and the Web kicked in to keep bandwith use growing. Telstra eventually bought out the backbone in 1995, leaving AARNET with its original universities and CSIRO.

Huston, who is now Telstra's manger of data networks, says the pioneers were somewhat surprised in the end that they ``won'' the battle against proprietary networks.

``We always thought, even when it was gathering steam in the early '90s, that we were showing the established communications sector what they should be doing. We never thought that those folk would then go and follow where we were.

``We thought that they were there to run the communications systems, we were there to play.

``There was an emerging realisation around '94 or so that this wasn't a game any more and that no one was going to come along and say: `game's over, we're now going in this (other) direction'.

``But when I look around at my peers, we're all working for commercial ISPs, none of us are left in the research sector.'' (Apart from Elz, who is still employed at Melbourne University's computer science department.)

``We're doing the same job; it's now business, it's not just a game. And it's now quite critical business for a whole swag of folk, but the most critical thing is that, for a while, there we were simply grabbing a little bit of spare capacity out of this gigantic voice network.

``We were taking a few fractions of a per cent, nibbling away at the edge - taking a little bit of their capacity and saying `we can run data'.

``We've used all that up, all of it. We've run through the lot.

``So now it really is time to get serious because from bottom to top we now have to build a data network.''

xamaxsmall.gif missing
The content and infrastructure for these community service pages are provided by Roger Clarke through his consultancy company, Xamax.

From the site's beginnings in August 1994 until February 2009, the infrastructure was provided by the Australian National University. During that time, the site accumulated close to 30 million hits. It passed 65 million in early 2021.

Sponsored by the Gallery, Bunhybee Grasslands, the extended Clarke Family, Knights of the Spatchcock and their drummer
Xamax Consultancy Pty Ltd
ACN: 002 360 456
78 Sidaway St, Chapman ACT 2611 AUSTRALIA
Tel: +61 2 6288 6916

Created: 12 July 1999 - Last Amended: 12 July 1999 by Roger Clarke - Site Last Verified: 15 February 2009
This document is at
Mail to Webmaster   -    © Xamax Consultancy Pty Ltd, 1995-2022   -    Privacy Policy