Date: Fri, 22 Oct 2004 17:55:45 +0200 From: Andre Oppermann <andre@freebsd.org> To: Brian Fundakowski Feldman <green@freebsd.org> Cc: freebsd-arch@freebsd.org Subject: Re: Removing T/TCP and replacing it with something simpler Message-ID: <41792D81.C030A26F@freebsd.org> References: <4177C8AD.6060706@freebsd.org> <71C3A1EA-238F-11D9-9171-000A95C705DC@chittenden.org> <41780672.6AAC705F@freebsd.org> <E0C34A72-2396-11D9-9171-000A95C705DC@chittenden.org> <417923BF.661D2A6D@freebsd.org> <20041022154517.GN1072@green.homeunix.org>
next in thread | previous in thread | raw e-mail | index | archive | help
Brian Fundakowski Feldman wrote: > > On Fri, Oct 22, 2004 at 05:14:07PM +0200, Andre Oppermann wrote: > > None of it. Neither NOPUSH nor CORK have any security implications. > > Those are only with the specification of T/TCP. Blocking the data > > is independend of 3WSH. Normally you have Nagle enabled (default) > > and when you don't fill an entire packet worth of data it will wait > > up to 200ms to send the packet in anticipation of more data from the > > socket. This screws the responsiveness of your connection. The first > > solution is to turn off Nagle (with TCP_NODELAY) but now you get a > > packet for every single write() you do. Fine for telnet and ssh but > > not the right thing for a database server. There you don't want the > > delay but at the same time you want several successive write()s that > > will go in one packet on the wire. Here NOPUSH and CORK come into > > play. > > Why is just tuning the delay a bad solution? If you tune it too low it ain't useful anymore (doesn't gather distant writes together) and too many timers too often. -- Andre
Want to link to this message? Use this URL: <https://mail-archive.FreeBSD.org/cgi/mid.cgi?41792D81.C030A26F>