JNET

From ZaInternetHistory

Well before the Foundation for Research Development (FRD) started the Uninet project in 1987, it was apparent to the group at Rhodes that it was essential to find a way to link the Rhodes Cyber to the network of IBM computers that linked the Universities of Potchefstroom, Witwatersrand and Pretoria together, and that also linked into the CSIR. That small network was connected at 9,600 bps on leased analog lines, and by all accounts there was no great use made of it. However, the network existed and was significant in that it connected different and autonomous institutions together. The technical work of maintaining this network was done by the networkers at Potch, so Potch was the obvious connection point for linking to the Rhodes computers. The cost of a leased line between Rhodes and Potch was simply unaffordable, so other means had to be found. In practical terms, the only option was to use a dialup connection, the only modems available worked at 2,400 bps. The communication protocol on this network was the very protocol that was used by Bitnet, viz IBM's RSCS, which used a synchronous communication protocol.

It was quite out of the question to attempt such a dialup link without first gaining some experience with the RSCS protocol. This link would have linked computers of different vendors and under separate autonomous control. The potential for chaos was simply far too high.

So, we started work at Rhodes to link a local computer to one of the Cybers by means of the RSCS protocol . This local computer was on our doorstep in the Physics department at Rhodes, in the form of a small VAX run by Justin Jonas.

Gateway Cyber/VAX

In the Physics department at Rhodes, there was a small Vax 11/730 computer. Some time in 1987 Mike Lawrie came across a package called JNET, which enabled the VAX to run the RSCS protocol. It interfaced well into the VAX operating system, and in particular into the VAX's email system. Thanks to the vision of the Control Data SA account manager for Rhodes, viz George Campbell, Rhodes had obtained the Network Job Entry Facility (NJEF) for the Cyber, which enabled the Cyber to run the RSCS protocols. Funds (about USD900) were found by the Physics department to purchase the JNET package from Joiner Associates in the USA. The political credentials of Rhodes were provided in order to deal with the sanctions issues which were then rife in the USA. The datestamp on the JNET tape, which was still at Rhodes in 1997, was 11 October 1987. By the 18th January 1988, good progress had been made with the connection between the Cyber and the VAX and we could make use of all of the facilities that it offered.

The linkup between the Cyber known as RURES and the VAX known as RUPHYS was established, and it worked very well. Modifications had to be made to the Rhodes mailing system that ran on the Cyber in order for mail to use the RSCS protocols as a transport mechanism. This was not too difficult for the Computing Centre staff to do. The exercise taught us a great deal about how to set about building a network of computers that ran different operating systems, and were controlled by different departments. The link operated at 9,600 bps on Microcom modems that had been obtained cheaply from a commercial company (name is forgotten) that had recently changed from analog to digital circuits and thus no longer needed analog modems.

In due course, when the Cyber/Fidonet gateway was working, it was a simple matter to include the RUPHYS host in the tables that routed email to the correct hosts. By an extension, carried out by Justin Jonas of the Rhodes Physics department, a microvax, called RUCHEM, in the Chemistry department could also be included in this network. A means was found to get email to a host, called RUBIS, in the Business Information Systems department - this being motivated by Malcolm Sainsbury. Simple? Well, not quite at the time, but simple in retrospect. One has to bear in mind that this was the only "Bitnet" style network that ever existed in South Africa, and we had to find our own way ourselves in order to make it work.