02-23-2013, 05:02 PM
As is often the case, it is the "nearest to best" or the "best of the correct" answer.
None of the answer take into account Eric the disgruntled ex-employee who may take his garden snippers and chop through the cables. Neither do we know if he has brought in his collection 10 x 48 port switches from his garage and taken the network over the 500 barrier, thus causing several of those mentioned problems to occur.
The 100m is a theoretical number. Sometimes it can be less, sometimes more, but it is not an exact science.
The idea is that if you operate with no hubs, no bridges and just purely routers and switches in with less than 500 hosts and no cables stretching over 100m, there should effectively be no collisions.
P68-69 of Wendell's book explains this is in more detail - but basically a switch in full duplex mode, when you add in the buffering and that each port has its own bandwidth, theoretically collisions will not occur.
None of the answer take into account Eric the disgruntled ex-employee who may take his garden snippers and chop through the cables. Neither do we know if he has brought in his collection 10 x 48 port switches from his garage and taken the network over the 500 barrier, thus causing several of those mentioned problems to occur.
The 100m is a theoretical number. Sometimes it can be less, sometimes more, but it is not an exact science.
The idea is that if you operate with no hubs, no bridges and just purely routers and switches in with less than 500 hosts and no cables stretching over 100m, there should effectively be no collisions.
P68-69 of Wendell's book explains this is in more detail - but basically a switch in full duplex mode, when you add in the buffering and that each port has its own bandwidth, theoretically collisions will not occur.