Thanks for the explanation. It seems since version 4.4 that kernel net.inet.tcp.rfc1323 is set to 1 by default, thus causing all the TCP connections to use the RFC1323 extension.
The effects are: 1. bigger TCP header. 2. more processing time at sending and receiving hosts. 3. VJ TCP/IP header compression algorithm does not compress most of the time. I am not sure turning on the RFC1323 support on by default is such a good idea. > The TCP timestamp option is used to obtain better round-trip time > estimates than can be obtained without, and these estimates turn out > to be important in networks with large bandwidth*delay products. > > Timestamps in the timestamp option also cycle much more slowly than > sequence numbers on an active high-speed connection and can thus be used > to detect and discard old duplicate packets with apparently valid sequence > numbers. > > RFC 1323 explains the details. > -- > G. Paul Ziemba [EMAIL PROTECTED] > FreeBSD unix: > 11:06AM up 16 days, 14 mins, 7 users, load averages: 0.03, 0.03, 0.00 > > To Unsubscribe: send mail to [EMAIL PROTECTED] > with "unsubscribe freebsd-net" in the body of the message > To Unsubscribe: send mail to [EMAIL PROTECTED] with "unsubscribe freebsd-net" in the body of the message
