Holger,
I am curious as to the basis of your claim that Linux is broken
because it doesnt support T/TCP type setup (data+SYN/FIN)?
As Alan pointed out 1644(T/TCP) is "experimental" i.e not standard and
therefore doesnt have to implemented. Regardless of that fact, As to the
validity of T/TCP, do you understand the DoS implications in T/TCP?
Unless those security issues are addressed, a couple of roundtrip times
is not worth justifying deploying it.
Now, regarding the Linux stack, could you elaborate a little more what you
dont like about it? Sure BSD is documeted by stevens in a book, but that
is a rather bad excuse to asse5rt that BSDish is better. You have to
provide better reasoning than that. I am not looking for a flame war -- so
please respond after youve taken a few deep breaths.
cheers,
jamal
PS:- Yes, i am familiar with source code of both stacks -- and i think
Linux is better. Infact, i was happier with 2.0 because it was much
much more mysterious. Ah, the sadistic joy i got, when
some "TCP/IP ... ahem ... BSD" expert tried to trace the code. I recall
someone telling me he got a headache as a result ;-> and ofcourse decided
to go back to BSD "because it was better and had been around for 15
years". I think he was just saying "i know BSD, and this Linux code is so
different and most of the things i know dont apply -- since i have
invested my money and time on stevens vol2, i'll go back there".
BTW, rumor has it Stevens is writting a book on the Linux stack. We'll
then have to bury BSD in some deep grave (i predict that will be the
demise of BSD).
-
To unsubscribe from this list: send the line "unsubscribe linux-net" in
the body of a message to [EMAIL PROTECTED]
