Relay-Version: version B 2.10 5/3/83; site utzoo.UUCP Posting-Version: version B 2.10.2 9/18/84; site wdl1.UUCP Path: utzoo!linus!philabs!cmcl2!seismo!umcp-cs!gymble!lll-crg!dual!amd!fortune!wdl1!jbn From: jbn@wdl1.UUCP Newsgroups: net.unix-wizards Subject: Re: Re: Time to live values for IP netwo Message-ID: <316@wdl1.UUCP> Date: Wed, 6-Mar-85 00:58:23 EST Article-I.D.: wdl1.316 Posted: Wed Mar 6 00:58:23 1985 Date-Received: Sat, 9-Mar-85 20:34:34 EST Sender: notes@wdl1.UUCP Organization: Ford Aerospace, Western Development Laboratories Lines: 15 Nf-ID: #R:brl-tgr:-881000:wdl1:17100067:000:745 Nf-From: wdl1!jbn Mar 5 20:00:00 1985 The time that a fragment is held by IP reassembly before reassembly times out is supposed to be controlled by the Time to Live field (ref MIL-STD-1777, para 9.2.4). So excessive Time to Live values can tie down reassembly buffers when fragments are being lost. I don't know if 4.2BSD uses the field in this way, but some implementations might. 1 minute (TTL=60) is probably a reasonable value. If transit times in your network are approaching one minute, you probably have serious congestion problems. If you have 4.2BSD running on slow links, you almost certainly have congestion problems; 4.2BSD is very badly behaved. See my RFC896 for some technical discussion. John Nagle Ford Aerospace and Communications Corp.