25Meg = 25 x 1024 x 1024 x 8 bits = 209715200 bits= minimum transfer time of about
20.97 seconds.... Now that doesn't include all the packet headers, but we are
showing a healthy 66% or therabouts with a 30 second transmission.
But was that two machines linked with nothing else on? What happens when it's
multiple? The 40% figure is from IBM, but I agree that it approaches that limit as
the number of machines (and collisions) increase and also as the packets vary in
number of octets (smaller packets require the same number of header octets as
larger ones.) It is ethernet plus TCP/IP that yields the 40% figure. I frankly
don't know enough about other protocols to guess at a calculation. Ethernet
itself, well, I don't think we can talk about a communication technology for
computer to computer without involving a message protocol.
Anyway, thank you for the data.
Civileme
[EMAIL PROTECTED] wrote:
> R u serious on the 40% efficiency on Ethernet topology? any reason why? coz a
> rough benchmarking i did when using samba transferring a 25 meg file took about
> 30s on a 10base-T network.
<BIG SNIP>
