7 ms·
It's easy to underestimate the importance of actually deploying those first long-distance internet links, at substantial expense. Links within a campus weren't
by tlb 8y ago
It's easy to underestimate the importance of actually deploying those first long-distance internet links, at substantial expense. Links within a campus weren't compelling enough to invent things like email to make use of them, because you could just walk over to the other guy's office. But high-speed (for the time) links between campuses made people want to use them for real work.
- nostrademons 8y agoI wonder if they knew how powerful this infrastructure would be at the time they did it. This presentation [1] (by Larry Roberts, actually) indicates that basically nobody other than ARPA actually wanted this capability - the other big institutions at the time either didn't want to share their data (universities) or thought the whole technical approach was crazy and brain-dead (telcos). [1] https://www.youtube.com/watch?v=qkD4HVRnGJE#t=580 https://www.youtube.com/watch?v=qkD4HVRnGJE#t=580
- kev009 8y agoLarge institutions had these capabilities since the 1960s. There were parallel universes like Tymnet and companies would set up first party networks using vendor protocols like IBM's SNA (which had a few precursors like QTAM). Users were concentrated onto large machines but machine to machine communication was equally important. This history and significance seems appallingly preserved. There were even parallel universe packet switched networks by phone companies like Bell and MCI in the US. The novelty of the ARPANet was that it allowed for the construction of a resilient open access network (Tymnet and various phone companies packet switchers were centrally administered walled gardens). Once the code was coupled to UNIX in the form TCP in BSD, it rode that wave and the sequel of "open systems" a few years later into ubiquity across all operating systems.
- ams6110 8y agoTymnet existed into the 1990s at least. I used it for dial-up access to servers at work for remote nighttime support of batch jobs. I don't miss those 2am calls.
- csours 8y agoThe auto assembly plant I worked in had SNA (active, but unused) in 2006.
- C1sc0cat 8y agoYes as a former member of TEC (Tymnet Engineering centre) the whole OSI networking tech is poorly documented. I remember running map reduce code on the largest(non black) UK PR1ME network in the 80's which was cool. OSI's problem was it was designed in a world where each country would usually have a single PTT - in the UK it also suffered from not billing as an all you can eat model.