From owner-freebsd-hackers Mon Jun 22 16:58:48 1998 Return-Path: Received: (from majordom@localhost) by hub.freebsd.org (8.8.8/8.8.8) id QAA26854 for freebsd-hackers-outgoing; Mon, 22 Jun 1998 16:58:48 -0700 (PDT) (envelope-from owner-freebsd-hackers@FreeBSD.ORG) Received: from smtp04.primenet.com (daemon@smtp04.primenet.com [206.165.6.134]) by hub.freebsd.org (8.8.8/8.8.8) with ESMTP id QAA26772 for ; Mon, 22 Jun 1998 16:57:55 -0700 (PDT) (envelope-from tlambert@usr02.primenet.com) Received: (from daemon@localhost) by smtp04.primenet.com (8.8.8/8.8.8) id QAA26863; Mon, 22 Jun 1998 16:57:54 -0700 (MST) Received: from usr02.primenet.com(206.165.6.202) via SMTP by smtp04.primenet.com, id smtpd026820; Mon Jun 22 16:57:45 1998 Received: (from tlambert@localhost) by usr02.primenet.com (8.8.5/8.8.5) id QAA00953; Mon, 22 Jun 1998 16:57:43 -0700 (MST) From: Terry Lambert Message-Id: <199806222357.QAA00953@usr02.primenet.com> Subject: Re: LSMTP and Unix (fwd) To: shocking@prth.tensor.pgs.com (Stephen Hocking-Senior Programmer PGS Tensor Perth) Date: Mon, 22 Jun 1998 23:57:42 +0000 (GMT) Cc: freebsd-hackers@FreeBSD.ORG In-Reply-To: <199806220210.KAA14210@ariadne.tensor.pgs.com> from "Stephen Hocking-Senior Programmer PGS Tensor Perth" at Jun 22, 98 10:10:57 am X-Mailer: ELM [version 2.4 PL25] MIME-Version: 1.0 Content-Type: text/plain; charset=US-ASCII Content-Transfer-Encoding: 7bit Sender: owner-freebsd-hackers@FreeBSD.ORG Precedence: bulk X-Loop: FreeBSD.ORG > Curious - I think the giveaway is in the part of the quote where they state > they VMS was engineered to be able to handle lots of async I/Os. It's quite > possible that the code they're using relies on this. So if your thread > implementation isn't up to scratch (and this is one area where NT shines, see > the part of the quote about scheduling et cetera) you could lose out big time. > Various benchmarks have been released that show NT can schedule huge numbers > of threads very efficiently. On the other hand, its process scheduling is > pretty sad. My feeling is that the filesystem issues may not be a big deal, > unless you filesystem code isn't multi-threaded. More likely, given that the second number was alluded to relate to network performance, they are probably using the NT "SendFile" mechanism. This would be logical, considering the line terminators in the file system and the line terminators in the protocol are both CRLF. This implies that they are "short-cutting" the wire-off/wire-on translations -- specifically, that they are storing the "." session terminators verbatim in order to not have to byte-unstuff/byte-stuff on the way in and out. This is a very old trick. You can see it in the sample POP3 code that came with the NT SDK. Terry Lambert terry@lambert.org --- Any opinions in this posting are my own and not those of my present or previous employers. To Unsubscribe: send mail to majordomo@FreeBSD.org with "unsubscribe freebsd-hackers" in the body of the message