Remember that almost all "computer calculations" for a Kilobyte is
1024. In networking, 1 Kilobyte is 1,000 bytes. So you have to be
careful about using the term "byte" FWIW, almost all network people
use bits/sec as that's what we're used to and there can be no confusion.
Also, you have to account for the various header if you are trying to
nail down the EXACT throughput (besides the fact that file size reported
by OS assumes 1024 is 1KB.) You have to account for PPP header, HDLC
header, or Ethernet header including preamble and CRC. *IF* you want
the most accurate number I mean.
Also, at any given moment, the network is 100% utilized or 0% utilized.
Either you have a bit on the wire or you don't. So if you're going to
average it, you need the appropriate sample size and live with the fact
that it's just an average.
That is all correct. I'm just wondering why is the graph inconsistent
with the results of simple calculations, given that both emerge from the
same data.
--
Best regards,
Michal Kepien