A small correction. It is common fallacy that ethernet maxes out at 40% of nominal capacity. This non-fact was started by a very old paper by Matcalf etc. that used a pretty crude model that came up with that ~40% figure. Actual tests show that ethernet is effecient up into the 90% range on collision domains with hundreds of nodes. See the O'Reilly ethernet book for details. Note that this is not talking about switched/full-duplex ethernet which is better still.
Slashdot Mirror
If you read to the end TFA really seems like nothing more than an elaborate set up for a pun.
Maybe this should be written as 1 ????? 10 base pi should be *approximately pi^2*
A small correction. It is common fallacy that ethernet maxes out at 40% of nominal capacity. This non-fact was started by a very old paper by Matcalf etc. that used a pretty crude model that came up with that ~40% figure. Actual tests show that ethernet is effecient up into the 90% range on collision domains with hundreds of nodes. See the O'Reilly ethernet book for details. Note that this is not talking about switched/full-duplex ethernet which is better still.