Return to BSD News archive
Path: euryale.cc.adfa.oz.au!newshost.anu.edu.au!harbinger.cc.monash.edu.au!news.mel.connect.com.au!munnari.OZ.AU!news.ecn.uoknor.edu!paladin.american.edu!europa.chnt.gtegsc.com!news.msfc.nasa.gov!bcm.tmc.edu!newshost.convex.com!usenet From: "Alan E. Ross" <alan@popi.net> Newsgroups: comp.unix.bsd.bsdi.misc Subject: tcp_do_rfc1323 Date: Wed, 17 Jan 1996 11:21:31 -0800 Organization: CONVEX Computer Corporation, Richardson, TX USA Lines: 27 Message-ID: <30FD4C3B.276A@popi.net> NNTP-Posting-Host: mobil.convex.com Mime-Version: 1.0 Content-Type: text/plain; charset=us-ascii Content-Transfer-Encoding: 7bit X-Mailer: Mozilla 2.0b5 (Win16; I) Hello all, As a follow up to John Chu's Email regarding telnet sessions being slow on his machine, somebody suggested that the option to make the networking s/w conform to RFC 1323 might be causing a slow down. BACKGROUND: RFC-1323 has to do with TCP window scaling, and time stamping to make connections to "very fast" networks run faster. It stuffs more stuff into the pipe ( window scaling ) and stamps it so that it does not overrun it. In my case, I have a Pentium/100 w/ 32Mb memory connected to an ISDN 128kb line running BSDI 2.0.1. I would presume that I would NOT benefit from the RFC-1323 extensions to TCP. I looked in /usr/src/sys/netinet/in_proto.c and found that "tcp_do_rfc1323 = 1", based on this I am assuming that my kernel is configured to utilize the 1323 extension. Can anybody give me some numbers or experiences on what I might expect to see as far as throughput improvements/degradations if I recompile my kernel with "tcp_do_rfc1323 = 0"? Any feedback on the situation would be most greatfully accepted. Regards, Alan Ross