Bora> There is plenty of analysis out there that states that with a physical link
Bora> with Bit Error Rate (BER) of 10^-7, TCP gets very limited throughput
Bora> (approximately 30 percent or less) due to the losses being interpreted as
Bora> congestion and premature clamping on the window. Hence I think it is WELL worth
Bora> the effort of making TCP/IP work over links with physical errors.
Bora> For 1500 byte packets 10^-7 translates to roughy 0.0012 packet error rate
Bora> which is much less than 1 percent but the throughput is still dismal.
Sorry to be so pedantic, but in one instance the laser beam is focused
on the IP header and one states that TCP header information will not be
used. When specifically asked how this is implemented, the focus shifts
to TCP. Again, my humble apologies for being so precise about this
point.
My understanding is that IP and TCP are not to be coupled. However, there
seems to be momentum to couple the TCP congestion control algorithm with
FEC to help TCP distinguish between noise (your example of 10-7 BER) and
forward congestion.
>From a purists perspective, this coupling of TCP and IP is not consistant
with the layered protocol design objectives of TCP/IP.
Can you provide the exact URL of the analysis?, and I'll do more homework
before beating the purists drum.
Thanks!
This archive was generated by hypermail 2b29 : Mon Feb 14 2000 - 16:14:30 EST