Path: utzoo!utgpu!news-server.csri.toronto.edu!cs.utexas.edu!usc!zaphod.mps.ohio-state.edu!mstar!mstar.morningstar.com!bob From: bob@MorningStar.Com (Bob Sutterfield) Newsgroups: comp.mail.misc Subject: MX agreements (was Re: OMNET <--> Internet : How?) Message-ID: Date: 14 Feb 91 16:21:50 GMT References: <7105.27bab5df@cc.curtin.edu.au> <4328@dftsrv.gsfc.nasa.gov> <4330@dftsrv.gsfc.nasa.gov> Sender: usenet@MorningStar.COM (USENET Administrator) Reply-To: bob@MorningStar.Com (Bob Sutterfield) Organization: Morning Star Technologies Lines: 20 In-Reply-To: dweissman@amarna.gsfc.nasa.gov's message of 14 Feb 91 15:17:05 GMT In article <4330@dftsrv.gsfc.nasa.gov> dweissman@amarna.gsfc.nasa.gov (WiseGuy) writes: ...The MX for [some]NET uses [someone]'s POSTMAN gateway at [somewhere]. I know the chart you refer to but I'm not sure how [someone] feels about mass use of the [somewhere] gateway. If [someone] agrees to MX for someone else, they're agreeing to all possible implications of that action. Once an MX record is broadcast, it's the way that *everyone* is supposed to get to that domain. If [some]NET didn't agree up front to limit the load offered to their MX (how?), then the MX can either (a) ask politely that they reduce the traffic or (b) tell [some]NET to go find another MX. Besides the [somewhere] gateway can sometimes take hours or days to deliver (if at all). If [some]NET's customers complain loudly enough, then [some]NET's management will go find another MX. There are folks (e.g. UUNET, PSI, SURFnet) who do it for money, and in exchange accept responsibility to provide a pre-arranged level of service, which can be contractually specified to include traffic volume and passthrough delay time.