From owner-cvs-all Thu Jun 15 22:42: 4 2000 Delivered-To: cvs-all@freebsd.org Received: from redfish.go2net.com (redfish.go2net.com [207.178.55.5]) by hub.freebsd.org (Postfix) with ESMTP id 2030F37BC45; Thu, 15 Jun 2000 22:41:58 -0700 (PDT) (envelope-from marcs@go2net.com) Received: from marcs (helo=localhost) by redfish.go2net.com with local-esmtp (Exim 2.11 #1) id 132osV-0006tG-00; Thu, 15 Jun 2000 22:41:15 -0700 Date: Thu, 15 Jun 2000 22:41:15 -0700 (PDT) From: Marc Slemko X-Sender: marcs@redfish To: Mark Newton Cc: cvs-all@FreeBSD.ORG, cvs-committers@FreeBSD.ORG Subject: Re: cvs commit: src/sys/kern uipc_socket.c uipc_socket2.c src/sy In-Reply-To: <20000616145956.B91282@atdot.dotat.org> Message-ID: MIME-Version: 1.0 Content-Type: TEXT/PLAIN; charset=US-ASCII Sender: owner-cvs-all@FreeBSD.ORG Precedence: bulk X-Loop: FreeBSD.ORG On Fri, 16 Jun 2000, Mark Newton wrote: > On Thu, Jun 15, 2000 at 10:18:11PM -0700, Marc Slemko wrote: > > > I think the real point here is that, if this sort of specialized > > performance is required, then great, people can do it as a random LKM > > or whatever that drops in using reasonably generic and appropriate > > hooks. > > Smirk. > > freebsd# kldload /modules/apache.ko No reason to smirk about it. That is a reality on some OSes right now. NT. AIX. Linux (well, there is a kernel httpd there, don't know for sure if there is any public code to do what I'm talking about, which is slightly different) The kernel code is actually only a very simple cache with very simple and minimal parsing routines. Then it punts anything not in the cache or that is too hard down to a normal Apache httpd or whatever. It has been done. It is a nice magic show, since it lets you use the same server for static and dynamic content, it is just that the static stuff only gets punted down to userland for cache population while the dynamic stuff is always punted. Heck, that way you can even keep the disk IO out of the kernel. Do such tactics result in kick-ass benchmarks? You bet, you can take it to a point where you can't get any faster on given hardware (well, assuming your standard x86 server hardware) no matter what you do, since you run into bus limitations. Will IIS have a similar feature? Dunno, I know MS was working on it, just not sure if they will release it. Is it of any actual real world use? Well... I have my doubts. To Unsubscribe: send mail to majordomo@FreeBSD.org with "unsubscribe cvs-all" in the body of the message