From owner-freebsd-questions Tue Feb 20 3:25:56 2001 Delivered-To: freebsd-questions@freebsd.org Received: from grumpy.dyndns.org (user-24-214-56-129.knology.net [24.214.56.129]) by hub.freebsd.org (Postfix) with ESMTP id 5837237B491 for ; Tue, 20 Feb 2001 03:25:52 -0800 (PST) (envelope-from dkelly@grumpy.dyndns.org) Received: from localhost (localhost [127.0.0.1]) by grumpy.dyndns.org (8.11.2/8.11.2) with ESMTP id f1KBPJm90855; Tue, 20 Feb 2001 05:25:19 -0600 (CST) (envelope-from dkelly@grumpy.dyndns.org) Message-Id: <200102201125.f1KBPJm90855@grumpy.dyndns.org> X-Mailer: exmh version 2.3.1 01/18/2001 with nmh-1.0.4 To: efinley@efinley.com Cc: freebsd-questions@FreeBSD.ORG From: David Kelly Subject: Re: how many switches? In-reply-to: Message from Elliot Finley of "Mon, 19 Feb 2001 19:45:34 MST." Mime-Version: 1.0 Content-Type: text/plain; charset=us-ascii Date: Tue, 20 Feb 2001 05:25:19 -0600 Sender: owner-freebsd-questions@FreeBSD.ORG Precedence: bulk X-Loop: FreeBSD.ORG Elliot Finley writes: > This isn't exactly a FreeBSD question, but since it's going to be an > all FBSD network... :-) > > I know that I can only go through 3-4 hubs before I have timing > issues. My question is: how many switches can I go through? it seems > like it would be unlimited as long as the packet passed through them > and the response came back before the application timed out. This is > the case isn't it? Since a switch uses store-and-forward. You mix the terms "hub" and "switch" as if they are interchangeable. A hub is very much like a CATV distribution amplifier. Contributes measureable delay in ethernet time but no buffering. A switch is smarter, but just how smart varies. Its not the application timeout which is important. You didn't say what error messages you were seeing. Anyway, too many hubs or too long of a network wire, the result is the speed of light isn't fast enough. In half duplex your NIC listens for anyone else on the wire during the first 64 octets. If it doesn't copy its data perfectly then it backs off and tries again later. A "collision". A long network can postpone the collision detection past octet 64 and results in a "late collision" which is reported with urgency as something is broken. -- David Kelly N4HHE, dkelly@hiwaay.net ===================================================================== The human mind ordinarily operates at only ten percent of its capacity -- the rest is overhead for the operating system. To Unsubscribe: send mail to majordomo@FreeBSD.org with "unsubscribe freebsd-questions" in the body of the message