motz
dimecres, 21. de novembre 2007

price of the net

1969: imp, interface message processor, sort of early router. price per imp $150.000, designed by bolt beranek &newman, bbn. a company that is still around today. still contractor of darpa? i have no idea.

1984: new kind of router for the nsfnet was needed. nsfnet inhaled arpanet and got itself inhaled later on by aol. the only companie that accomplished the spec for the switching gateways, as it was called back then, was - again - bbn; but their price was high: $250.000 for one router; with a discount from darpa you could get it for $180.000 ..

fuzzball, programmed by dave mills, made the game. One programmer in charge for the whole new network of networks, as nsfnet gets described. in fact one guy in charge of the internet backbone. he didn't get a ot of sleep during those days, dennis jennings says, at that time head of nsf.

i didn't ask for the price of snowball, but it must have been less than $10.000, because that was the price jennings felt is all right for campuses and regional networks and the backbone of nsfnet.

2007 cost of a cisco router: between $1,123.00 and $264.00 (i guess one would take a better one for something like a backbone ...)

... Comment

as ever with your knowledge and instincts, this is a very good and interesting story. it's well worth further elaboration. if you want to do it for radio once, get me on the racket, please.
inflation cleaning should be done. the story of the cisco founding should go in it. it should also say what the big ix ciscos today do and cost what the difference between an olden time interface processor and a router is.

AR: now you can buy a router for $50, that would be able to handle mid80ies traffic with a load of 0,066. btw, i think ORF was one of the first companies in A to bye ca. 12 of those grey cisco backbone pots. amazing machines. we had one that was configured at startup, rebooted twice and never again until its retirement in 1998.
AR: in terms of org nsfnet inhaled arpanet but in nearly all other aspects arpanet ate up nsfnet.
AR: BBN did imp not by designing but by pimping an Honeywell GP minicomputer. Computer pimping still meant soldering AND programming then.
AR: AFAIK Fuzzball never was a product. It was a tape and kind of tarball that people adapted to PDP11s and ICL minicomputers. I think it came for the price of the tape as many things then. Maybe there was a license. The hardware was either there (quite obsolete already) or had to be purchased at a price clearly higher than $10.000. Ethernet was standard on those boxes but the telephone interface was an addon, i think. satellite interface was a box of its own talking to the cpu by all kinds of different electric and logical interfaces. So, a fully equipped hw platform for fuzzball would cost quite a bit, but was already present and mostly underused in technical ac institutions in the mid80ies.


Early Cisco AS MPR at CERN in 87

... Link

sure bbn pimped a honeywell, yet ornstein was in charge of hardware and kahn i think worked on the software ... under the leadership of frank heart.

i didn't care too much about inflation calc, and the story somehow ends in the 80s. but there is a nice snipped to cisco, karrenberg told me: cisco legende

(audio/mpeg, 542 KB)

... link

sorry to disturb you again, aber was karrenberg sagt, ist einfach nur bullshit. In diesem O-Ton ist einfach alles total falsch.
pdp11s waren keine mainframes, sondern relativ billige minicomputer.
sie brauchen mehr strom als ein pc, aber kein kraftwerk.
die essenz der cisco-gründung ist, dass die bosack-lerners einfach mit relativ fertigen forschungsergebnissen aus stanford richtung markt abgehauen sind und die anderen Leute aus dem RZ da um eben diese Ergebnisse abgezockt haben.
die idee bei diesen routern war, dass die cpu ziemlich schwach sein kann, das os ziemlich abgespeckt usw., weil die kiste ja kein zurechtgebogener general purpose rechner sein muss, sondern spezialisiert sein kann.
Der Markt hatte 1984 mehrere 100 wenn nicht mehrere 1000 Abnehmer.
Mich machen "Auskenner", die nicht mal recherchieren und einfach falsche Legenden nachplappern, ohne auch nur den Kontext annähernd im Auge zu haben, total krank.

... link

nanana, ich leit's weiter ...
pdp 11 kommt im ot ja nicht vor?
wenn jemand das wort "legende" verwendet, wäre es wohl eher an mir, das alles nachzurecherchieren ... in anderen worten: ich kann mich da nicht aus der pflicht nehmen.

cisco wurde 84 gegründet, jennings war 84/85 für's nsfnet verantworlich und suchte nach einer lösung für seine specs. wenn er sagt, er halte sich zugute cisco in diese richtung weitergetrieben zu haben, dann weiss ich noch immer nicht wie die situation vor 84 war ...

... link

kurzschlussgedanke, aber die honeywells waren auch minicomputer und ich meinte ja den o-ton von karrenberg. jennings ist ok. 84 gab's immer noch imp nachfolger von BBN, ich weiß nicht, was sonst noch und viele viele "Bastellösungen", von denen fuzzball wohl die interessanteste war.

... link


... Comment

Online for 8550 days
Last update: 3/11/23 17:00
status
Youre not logged in ... Login
menu
... Home
... Tags


search
calendar
novembre 2024
dg.dl.dt.dc.dj.dv.ds.
12
3456789
10111213141516
17181920212223
24252627282930
novembre
recent updates
human "The mind is what
the brain does." (margaret boden) Mind As Machine. A History...
by motzes (3/11/23 17:00)
when industry looks old i
have no idea how i came here, but i still...
by motzes (13/12/22 21:10)
holography explained it has been
20 years since i met nils abramson and heard about...
by motzes (20/2/22 10:22)
digital dilemma as seen in
the year 2000 . Intellectual Property in the Information Age...
by motzes (28/1/22 8:56)
anti colonial connectivity "... it
was after all, the early days of Intelsat, when having...
by motzes (16/8/21 11:20)
old stories revisited ... ...
makes one search again, along the lines given. brought me...
by motzes (6/7/21 14:27)
history writing gerade
im ohr: ein interview mit verkühlter stimme. aufnahmedatum: 2016.
by motzes (30/3/20 15:42)
Nice Thanks for uploading this.
It's an amazing window on the early history of interactive...
by Kayla (1/3/20 15:51)
gibberjabber interesting, die eingefangenen bots
werden in ihrer wortwahl aggressiv.
by motzes (26/10/19 20:41)
rätsel Daniel Schwenter, Philosophischen und
Mathematischen Erquickstunden, Dritter Theil, 1653 | https://archive.org/details/bub_gb_bGM_AAAAcAAJ
by motzes (22/10/19 19:06)

RSS Feed

Made with Antville
Powered by
helma object publisher