On Jul 26, 19:31, Bill Pechter wrote:
> Several transceivers connected to the same cable
allow several machines
> to communicate. In a sense, the transceivers together with the coax
cable
> form the hub.
>
> The T-pieces are, indeed passive (all 3 connectors connected in
parallel
in the obvious
way), but they're not really the hub.
-tony
It's really that the hub doesn't exist on a bus network like 10Base2
or 10Base5... all the Tee connectors are doing is replacing
the vampire tap on thicknet...
The thinnet "Transceiver" is really built on
to the card on
most PC's and is the same (basically) as the old thicknet
transceiver attached to my Sun my Unix boxes -- they can go from
Thinnet to Thicknet by swapping the N-Connector top to the
BNC connector top. And thinnet can go to thicknet with just an N to BNC
adapter -- but the max length and specs drop to thinnet specs.
Hubs really distort the logical ethernet bus topology.
Not really. Logically, the innards of the hub are the bus. The ports are
the taps. Then, since the stations connected are typically some distance
away, you use an appropriate technology to get the cable to them, and
that's why there are additional transceivers (one on the host card, one on
the hub port) to drive the signal over a resonable distance. It certainly
changes the physical layout of course (short bus long drops instead of a
long bus with short drops).
Now 10/100 switches really screw with it.
Those switches are just bridges, which have been around for decades (albeit
with fewer ports).
--
Pete Peter Turnbull
Network Manager
University of York