From owner-freebsd-hardware Thu Oct 8 02:50:32 1998 Return-Path: Received: (from majordom@localhost) by hub.freebsd.org (8.8.8/8.8.8) id CAA02891 for freebsd-hardware-outgoing; Thu, 8 Oct 1998 02:50:32 -0700 (PDT) (envelope-from owner-freebsd-hardware@FreeBSD.ORG) Received: from Tasha.STARDreams.org (maccess-01-070.magna.com.au [203.111.85.70]) by hub.freebsd.org (8.8.8/8.8.8) with ESMTP id CAA02732 for ; Thu, 8 Oct 1998 02:50:06 -0700 (PDT) (envelope-from kevla@studentmail.dis.unimelb.edu.au) Received: from EventHorizon (EventHorizon.STARDreams.org [10.144.144.1]) by Tasha.STARDreams.org (8.9.1/8.8.8) with SMTP id FAA05379 for ; Thu, 8 Oct 1998 05:50:00 -0400 (EDT) (envelope-from kevla@studentmail.dis.unimelb.edu.au) Message-Id: <3.0.3.32.19981008194944.0097da90@studentmail.dis.unimelb.edu.au> X-Sender: kevla@studentmail.dis.unimelb.edu.au X-Mailer: QUALCOMM Windows Eudora Pro Version 3.0.3 (32) Date: Thu, 08 Oct 1998 19:49:44 +1000 To: freebsd-hardware@FreeBSD.ORG From: Kevin Lam Subject: Re: Full duplex ethernet In-Reply-To: <199810072350.NAA20011@pegasus.com> Mime-Version: 1.0 Content-Type: text/plain; charset="us-ascii" Sender: owner-freebsd-hardware@FreeBSD.ORG Precedence: bulk X-Loop: FreeBSD.org At 13:50 10/7/98 -1000, you wrote: >Which ether-card drivers currently support full duplex? > >Recently, someone here said that full duplex for 10Base-T was a waste >of time. But after some research it's beginning to look like that's >not true at all. > >10BT switches are available, and cost way less than 100BT switches. > >I've gotten reports from several places saying that in practice, switched >full-duplex 10BT is faster than half-duplex hub-based 100BT. You might be confusing a few things here. Firstly, full-duplex does not equate to switching.. however it is usually implemented alongside switching. It is true that switched 10BT would be faster than hub-based 100BT, but it would hold true even if it was switched half-duplex 10BT. The vast improvement in performance comes from the switch, which effectively creates a dedicated point-to-point channel for communications on the network, being much more powerful than shared 100BaseTX. In a switched architecture, collisions, which greatly impact performance, are almost eliminated, however, the two stations exchanging data could also collide with each other, even in a switched architecture. Full duplex merely eliminates this from happening at all (as both may transmit simultaneously). Switching plays the greatest part in reducing collisions and segregating communications, then full-duplex forms the icing on the cake. Anyway, for what it's worth, the xl driver (3Com Fast EtherLink) has full-duplex functions at both 10 and 100Mbps listed as available media. -- K "Deliver yesterday, code today, think tomorrow." To Unsubscribe: send mail to majordomo@FreeBSD.org with "unsubscribe freebsd-hardware" in the body of the message