Network management Toolkit

Story: Inventor celebrates 30 years of Ethernet

Posted by: philip overy (Thursday 17 November 2005, 1:35 PM)

arpanet didn't introduce packet switching, the uk National Physical Laboratory did it in 1963 or thereabouts. and it didn't stay in the lab either....

When I started work in computing in 1975 at BNF Metals Technology Centre in Wantage, UK they were already contributing to Lockheed Dialog or were preparing to over the Experimental Packet Switching Service* run by the Post Office (not even yet called British Telecom? not sure) and X25 already went right round the world, over GEC switches and BT lines (or GPO?). BBN stripped down X25 to make their arpanet work, but they didn't improve it which is why there is TCPIP Version 6.
*ie the Americans weren't using TCPIP either...

Not only that, Arpanet was still military, so pretending that universities (except American ones and UCL-CS here) could use it would be disingenuous. It was donated to the world because it was inherently insecure - nowadays touted "as a virtue". Quite a lot of the world's computer development happened courtesy of Coloured Book X25 protocols and Cambridge Ring LANs. and to be less jingoistic, quite a lot also went via BITNET which I later supported in the UK and IBM's RSCS network and internal VNET.

Xerox Parc was/is a wonderful institution. Just as they deserve the credit for windows, not Microsoft, NPL deserves the credit for packet switching. True. Teddington doesn't sound as glamorous.I 'd also claim that RINGS are extremely useful if replaceable for deterministic networks - and rings were used with fibre optics for a long time when it wasn't that easy to make drivers for anything else (eg FDDI). System X/ISO Switching System Number 7, the UK's/international definitely better 32-channel not 24-channel phone system was another unsung great leap forward - it had so much flak in the press it deserves a mention in redress. When Clifford Stoll traced the KGB Internet hackers out of Berkeley it was ISO digital networks and X25 monitoring that tracked them to Hamburg, possibly not a plus, thinking of 'The Net'...

I have only ommitted JAPAN from this diatribe out of sheer ignorance - I don't know what they're up to, but I bet it's just as interesting. Garage Projects make great hype but that's not how the world works.

My father studied physics and radar at Bristol University (the original, UK, Bristol). At that time Gray Walter was messing about elsewhere in the university on what he called "tortoises". Gray Walter was an American anglophile - I expect he would have preferred his name of "tortoises" to continue instead of "turtles" as the computing community chose to name the digital era's version. If you're interested in the "turtle" concept, switch to dry land for the pre-war story...

I have a serious point - if you want to read the right early papers, you can use heavyweight technology like Google and Science Citation Index techniques or you can simply know the right name and the right story. It saves so much effort to know the right place to look in the first place for background information. ALOHA and Ethernet were great ideas and the first users of randomisation, a very pervasive concept in digital communications nowadays, but neither of them arose in a vacuum. or an all-American comms discipline.

phil overy
boring old real time programmer