From owner-freebsd-net Thu Feb 7 5:23:11 2002 Delivered-To: freebsd-net@freebsd.org Received: from web.cs.ndsu.nodak.edu (web.cs.ndsu.NoDak.edu [134.129.125.7]) by hub.freebsd.org (Postfix) with ESMTP id 2F55537B422 for ; Thu, 7 Feb 2002 05:23:08 -0800 (PST) Received: (from tinguely@localhost) by web.cs.ndsu.nodak.edu (8.11.4/8.11.4) id g17DN7H40434; Thu, 7 Feb 2002 07:23:07 -0600 (CST) (envelope-from tinguely) Date: Thu, 7 Feb 2002 07:23:07 -0600 (CST) From: mark tinguely Message-Id: <200202071323.g17DN7H40434@web.cs.ndsu.nodak.edu> To: freebsd-net@FreeBSD.ORG, knmurthy30@hotmail.com Subject: Re: Duplicate Acks and Fast Retransmit In-Reply-To: Sender: owner-freebsd-net@FreeBSD.ORG Precedence: bulk List-ID: List-Archive: (Web Archive) List-Help: (List Instructions) List-Subscribe: List-Unsubscribe: X-Loop: FreeBSD.org > In order to understand the effect of reordering better, > > a)the value of "Fast Retransmit Threshold" in tcp_input.c is modified to 1 ouch, hopefully this is an isolated network. > b)packets of the connection are distributed over 2 links. > > 1. Is it possible to have so many duplicate acks but not > still have a Fast Retransmit ? (TCP source clearly shows that > there is a possibility of this but is it not surprising > that probability the probability is so high?) yes, on a private network where you do not have congestion, lost, or in your setting of the fast retransmit to equal 1, even reordering of the packets. Having these number when traversing the Internet backbone would be very surprising. > 2. What does the field "Old duplicate packets" (in this case, ACKs) mean ? > (This was the first TCP session after reboot and the > earlier TCP sessions apparently closed without any problem). that is a wrap in the PAWS sequence number. To Unsubscribe: send mail to majordomo@FreeBSD.org with "unsubscribe freebsd-net" in the body of the message