From owner-freebsd-hackers Thu Dec 14 0:55: 4 2000 From owner-freebsd-hackers@FreeBSD.ORG Thu Dec 14 00:55:02 2000 Return-Path: Delivered-To: freebsd-hackers@freebsd.org Received: from menelao.polito.it (menelao.polito.it [130.192.3.30]) by hub.freebsd.org (Postfix) with SMTP id BCF5237B400 for ; Thu, 14 Dec 2000 00:55:01 -0800 (PST) Received: (qmail 848 invoked from network); 14 Dec 2000 08:54:53 -0000 Received: from truciolo.polito.it (HELO truciolo) (130.192.16.81) by menelao.polito.it with SMTP; 14 Dec 2000 08:54:53 -0000 From: "Fulvio Risso" To: , Cc: Subject: Kyxtech: freebsd outsniffed by wintendo!!?!? Date: Thu, 14 Dec 2000 09:53:29 +0100 Message-ID: MIME-Version: 1.0 Content-Type: text/plain; charset="us-ascii" Content-Transfer-Encoding: 7bit X-Priority: 3 (Normal) X-MSMail-Priority: Normal X-Mailer: Microsoft Outlook IMO, Build 9.0.2416 (9.0.2911.0) In-reply-to: <200012122113.eBCLDX305049@mass.osd.bsdi.com> X-MIMEOLE: Produced By Microsoft MimeOLE V5.00.2919.6700 Importance: Normal Sender: owner-freebsd-hackers@FreeBSD.ORG Precedence: bulk X-Loop: FreeBSD.ORG Hi. Sorry for bothering you again. This is the last public reply on this argument. - I'm very happy to hear (from many of you) that what "Jessem" said was not the FreeBSD-people thought. - I like FreeBSD myself; we have several FreeBSD servers on our network - I apologize for not having sent a summary of our tests on the FreeBSD mlist. - Fast answers to the several questions: Buffer sizes We did not make any test about creating 1MB buffers. However our architecture does not have the problem that "large buffer" = "large time used to transfer this buffer to user level" because we are able to transfer small pieces of data according to the network status. In other words, we have a transfer similar to the BIOCIMMEDIATE mode of BSD, so the transfer should always be optimized. Context switches Are kept the lowest, because several packets can be transferred at the same time. Safety We still keep 2 buffers (kernel and user). We're evaluating a 1-buffer architecture (taking into account the security problems) UFS filesystem We used FAT to strore packets because the UFS filesystem was on a second hard disk, so that the disk could have made some difference. We can reinstall BSD and repeat the test. We'll do that for sure. Improving BPF We made source code freely availble (http://netgroup-serv.polito.it/WinPcap/). We're very happy if someone will implements these ideas in BSD, so we can discuss (and compare) ideas between us. To understand more We submitted a paper "An Architecture for High Performance Network Analysis". You can find more details (and updated tests) at http://www.polito.it/~risso/research/WinPcap.pdf Thank you again, fulvio and the WinPcap team To Unsubscribe: send mail to majordomo@FreeBSD.org with "unsubscribe freebsd-hackers" in the body of the message