Relay-Version: version B 2.10 5/3/83; site utzoo.UUCP Path: utzoo!mnetor!uunet!seismo!gatech!hao!ames!lll-lcc!pyramid!ncc!lyndon From: lyndon@ncc.UUCP (Lyndon Nerenberg) Newsgroups: comp.os.minix Subject: Re: (none) Message-ID: <1498@ncc.UUCP> Date: Tue, 21-Jul-87 01:08:36 EDT Article-I.D.: ncc.1498 Posted: Tue Jul 21 01:08:36 1987 Date-Received: Wed, 22-Jul-87 06:03:40 EDT References: <366@louie.udel.EDU> Organization: Nexus Computing Corp., Edmonton, AB Lines: 32 There is a lot of merit in looking at NFS/RFS type enhancements to Minix, however I'm still at a loss as to what would be the best underlying link level protocol. Given that the "average" teaching environment won't have a lab full of PC's tied to a server via Ethernet (nor does the support code currently exist) we should probably examine the options we have to "network" via serial (RS232) links. The first thing that springs to mind is SLIP, however there have been some recent RFC's released addressing inter-host access using an "asynchronous line discipline" designed for PC to "host" access via dialup. Given that Minix doesn't currently support UUCP, perhaps we could look at building a slow speed network over top of this for applications similar to SMTP, FTP, TELNET, etc. In this environment, a rewrite of (ISO) layers 2 and 3 should allow us to implement these existing protocols on top of the new async spec without having to mess with the already well defined server/client and application layers. (bad analogy - I know) I realise that there is currently alot of work being done with Phil's code (myself being one of the participants), however this might be an excellent chance to work with some of the alternatives currently available to us without negatively impacting existing implementations. Lyndon Nerenberg VE6BBM alberta!ncc!lyndon pyramid!ncc!lyndon -- Ollie for president: the tradition continues.