From owner-freebsd-hackers Mon May 21 10:24:18 2001 Delivered-To: freebsd-hackers@freebsd.org Received: from master.ndi.net (master.ndi.net [209.45.245.62]) by hub.freebsd.org (Postfix) with ESMTP id BC33E37B422 for ; Mon, 21 May 2001 10:24:14 -0700 (PDT) (envelope-from ccf@master.ndi.net) Received: from localhost (ccf@localhost) by master.ndi.net (8.9.3/8.9.3) with ESMTP id RAA92273 for ; Mon, 21 May 2001 17:10:54 -0400 (EDT) (envelope-from ccf@master.ndi.net) Date: Mon, 21 May 2001 17:10:54 -0400 (EDT) From: "Charles C. Figueiredo" To: freebsd-hackers@freebsd.org Subject: technical comparison Message-ID: MIME-Version: 1.0 Content-Type: TEXT/PLAIN; charset=US-ASCII Sender: owner-freebsd-hackers@FreeBSD.ORG Precedence: bulk List-ID: List-Archive: (Web Archive) List-Help: (List Instructions) List-Subscribe: List-Unsubscribe: X-Loop: FreeBSD.ORG Hi, I appoligize if this is the improper channel for this sort of discussion, but it is in the best interests of the FreeBSD following, atleast, within my orginization. I work in an environment consisting of 300+ systems, all FreeBSD and Solaris, along with lots of EMC and F5 stuff. Our engineering division has been working on a dynamic content server and search engine for the past 2.5 years. They have consistently not met up to performance and throughput requirements and have always blamed our use of FreeBSD for it. We have humored them time and time again; i.e. they once claimed the lack of some sort of RAID was keeping them from meeting their requirements, when he had already thrown brute amounts of hardware at their application. When we setup a load-testing environment with multiple types of RAIDs, all the systems, including the one without any sort of RAID performed identically. And poorly, at that. We have had a recent change in departmental structure, which unfortunately, weakened the more technical side of the top of the food chain. They have taken this as another opportunity to push for Linux-use within our environment. We do not want, nor feel the need for introducing another OS into the environment. The following are the points that the head of engineering claimed were their requirements and our shortcoming, which Linux would handle well: --- a) A machine that has fast character operations b) A *supported* Oracle client c) A filesystem that will be fast in light of tens of thousands of files in a single directory (maybe even hundreds of thousands) Requirement a) means that it won't run well on a Sparc processor as they are notoriously bad at character addressing, and since search makes extensive use of character operations (as does *any* web application server for that matter), using a Sparc processor will be a waste since the x86 architecture (AMD's and Crusoe's especially) do it much better. Requirement b) means it won't be FreeBSD. Yes, you can run Linux apps under emulation, but I'd bet dollars for doughnuts that this will be a support nightmare if we can even get it to work. Requirement c) means it won't be Solaris or FreeBSD since neither of them have a filesystem which handles this effectively. Linux on Intel fits the bill because it meets these three requirements *very* effectively. --- I find them to be mostly silly points -- (a) touching on integer math -- pretty moot point given the real meat of this. (b) is wrong, since there is a native port of the oracle client and (c) is just silly -- I'm sure softupdates on a modern BSD ufs is loads faster than ext2fs. Folks, please give me some real technical ammo -- reference internals, give a real technical comparison if possible. I don't believe this is some lame FreeBSD/Linux comparison -- I'm simply trying to tactfully and effectively deal with a zealot. :-) Any help would be greatly appreciated. -charles. To Unsubscribe: send mail to majordomo@FreeBSD.org with "unsubscribe freebsd-hackers" in the body of the message