On Fri, Jan 17, 2014 at 3:46 AM, Eggert, Lars <[email protected]> wrote: > Nothing new under the sun. Proceedings from the very first IETF meeting, held > 28 years ago today: > > "Nagle presented his 'fair queuing' scheme, in which gateways maintain > separate queues for each sending host. In this way, hosts with pathological > implementations can not usurp more than their fair share of the gateway’s > resources. This invoked spirited and interested discussion. Zhang pointed out > that this was a subtle change in architecture away from a pure datagram > network. Callon reminded everyone that he had written a paper advocating a > connection oriented Internet Protocol several years ago." > > https://www.ietf.org/proceedings/01.pdf, page 5
It actually starts at page 88. Reading the rest was great. Wow. What a wonderful piece of history! Thank you for digging that up! Seeing the problems they were trying to solve at the time was very revealing. "1000 nodes running at 9600 baud, with one outage every 5 minutes will only use 20% of the bandwidth for routing information" And nagles preso retains many cogent points, although I guess "source quench" was all the rage at the time. > Lars > > _______________________________________________ > aqm mailing list > [email protected] > https://www.ietf.org/mailman/listinfo/aqm > -- Dave Täht Fixing bufferbloat with cerowrt: http://www.teklibre.com/cerowrt/subscribe.html _______________________________________________ aqm mailing list [email protected] https://www.ietf.org/mailman/listinfo/aqm
