At 03:09 PM 3/30/99 -0800, Jon Mansey wrote:
>Can anyone point me to some research or tables showing how BER affects
>TCP/IP thoughput?
>"Unix IS user friendly... It's just selective about who its friends are."
One could say the same thing about TCP being BER friendly. Most RFC
2001-compliant implementations assume that congestion has occurred if a
packet is lost, and halve their rate. Most TCPs also send segments of
either 512 or 1460 bytes in size - which with the TCP, IP, and link layer
overheads becomes on the order of 5000 to 12000 bits. A BER on the order of
10E-4 therefore stops them cold, and a BER on the order of 10E-N will limit
a single TCP stream to averaging 10E+(N/2) bits or less per round trip time.
That should break out something like this:
bits rate in kbps
per for 600 ms
exponent BER rtt delay
4 1E-04 1000 1.666666667 (actually, close to zero,
5 1E-05 1E+4 16.66666667 due to frame size)
6 1E-06 1E+5 166.6666667
7 1E-07 1E+6 1666.666667
8 1E-08 1E+7 16666.66667
9 1E-09 1E+8 166666.6667
10 1E-10 1E+9 1666666.667
11 1E-11 1E+10 16666666.67
12 1E-12 1E+11 166666666.7
This archive was generated by hypermail 2b29 : Mon Feb 14 2000 - 16:14:53 EST