Received: from localhost by CS.UTK.EDU with SMTP (cf v2.9s-UTK) id OAA19264; Fri, 5 Mar 1999 14:43:42 -0500 (EST) Received: by cs.cs.utk.edu (bulk_mailer v1.12); Fri, 5 Mar 1999 14:42:39 -0500 Received: by CS.UTK.EDU (cf v2.9s-UTK) id OAA19151; Fri, 5 Mar 1999 14:42:38 -0500 (EST) Received: from out0.mx.skynet.be (out0.mx.skynet.be [195.238.2.35]) by CS.UTK.EDU with ESMTP (cf v2.9s-UTK) id OAA19103; Fri, 5 Mar 1999 14:42:29 -0500 (EST) Received: from [195.238.19.121] ([194.78.31.228]) by out0.mx.skynet.be (8.9.3/odie-relay-v1.0) with SMTP id UAA08741; Fri, 5 Mar 1999 20:41:32 +0100 (MET) In-Reply-To: X-Mailer: CTM PowerMail 2.3.1 X-URL: http://www.skynet.be x-sender: blk@foxbert.skynet.be Date: Fri, 5 Mar 1999 18:20:21 +0100 To: Philip Hazel , Pete Resnick CC: John C Klensin , cos@leftbank.com, drums@cs.utk.edu From: Brad Knowles Subject: Re: FWD: Re: To MX or not to MX? Message-Id: <19990305182021.011465@relay.skynet.be> Mime-Version: 1.0 Content-Type: text/plain; charset=US-ASCII Content-Transfer-Encoding: 7bit List-Unsubscribe: On Fri, Mar 5, 1999, Philip Hazel wrote: > >The right way to fix all of this would be to start again, and start >registering services rather than hosts. You've got this today. They're called SRV records. Hardly anyone uses them yet (because they're so new), but they exist and are handled properly by current versions of BIND. -- These are my opinions -- not to be taken as official Skynet policy ____________________________________________________________________ |o| Brad Knowles, Belgacom Skynet NV/SA |o| |o| Systems Architect, News/mail/FTP Admin Rue Col. Bourg, 124 |o| |o| Phone/Fax: +32-2-706.11.11/12.49 B-1140 Brussels |o| |o| http://www.skynet.be Belgium |o| \/\/\/\/\/\/\/\/\/\/\/\/\/\/\/\/\/\/\/\/\/\/\/\/\/\/\/\/\/\/\/\/\/\/\/ The Sky is no longer the limit or layer four switches) to hide multiple hosts behind a single name. You can separate the POP3/IMAP servers from the outbound mail relays, but it is not strictly necessary (although I think it's a good idea). Unfortunately, MXes don't scale all that well, either. Once you get to hundreds of inbound mail machines for your domain, you can't realistically list all of them in your replies, so you end up having to use either DNS or NAT/switch tricks to get back down to a smaller subset that any one sender might know about at any one particular point in time. Based on what I've seen here so far, If I have to choose between SHOULD and MAY (or SHOULD NOT), I would vote for SHOULD NOT. I think we should explicitly state that how MUAs would deliver mail to their upstream servers is a private arrangement issue (since, as far as we're concerned, they're not truly on the Internet and we're not expecting them to do MX resolution), and then we should describe some typical private arrangements. As Pete points out, there are just too many different ways in which things might radically vary at different sites, and I don't think that it would be a good idea to try and encourage the adoption of any one of those alternatives to the detriment of any other. There's just too many "legacy" systems out there that probably wouldn't conform to any solution we might recommend (whatever that might be). How these guys go about doing their business has been unregulated and pretty much undocumented for too long, for us to come along at this late stage and try to impose our view of order upon them. Now, when it comes to the Message Submission protocol (whatever that might be), I think this would be an ideal time to make use of SRV records. But I think this is a matter for a different document, if not a different group. -- These are my opinions -- not to be taken as official Skynet policy ____________________________________________________________________ |o| Brad Knowles, Belgacom Skynet NV/SA |o| |o| Systems Architect, News/mail/FTP Admin Rue Col. Bourg, 124 |o| |o| Phone/Fax: +32-2-706.11.11/12.49 B-1140 Brussels |o| |o| http://www.skynet.be Belgium |o| \/\/\/\/\/\/\/\/\/\/\/\/\/\/\/\/\/\/\/\/\/\/\/\/\/\/\/\/\/\/\/\/\/\/\/ The Sky is no longer the limit