Xref: utzoo comp.mail.misc:5437 comp.mail.uucp:6507 Newsgroups: comp.mail.misc,comp.mail.uucp Path: utzoo!utgpu!news-server.csri.toronto.edu!rpi!uupsi!schoff From: schoff@uu.psi.com (Martin Schoffstall) Subject: Re: BITFTP grief! Message-ID: <1991May16.145758.6817@uu.psi.com> Organization: Performance Systems International, Inc. References: <81678@bu.edu> <1705@aupair.cs.athabascau.ca> Distribution: na Date: Thu, 16 May 91 14:57:58 GMT Peter, Most service providers are interesting in guaranteeing a level of service reliability every hour of every day of the year, across all of their customers. No one every succeeds, but that is the goal, if you get 95%+ you survive, if you get 99.9%+ you thrive. Mailing files is one of those things that doesn't work very reliably, due to many factors from error correction, to memory utilization, to installed base of mailer constraits, to resource constraints, etc. I've seen this debate for a decade, and I'm sure it continue forever. To transfer files we use a tried and true mechanism: FTP. Why does BITNET do this? Because it fits in their model, part of their model (from my biased perspective) is an acceptance of unreliability. Where else do you have dozens of huge organizations cut off for days, whole countries similiarly, nuked disk spools continually, and have to use non-standard MTU's to dump it (really tunnel it) through the Internet. [But it has those great applications! Or so they believe.....] Marty ---------------- In article peter@ficc.ferranti.com (Peter da Silva) writes: > >The real solution would be for UUNET and other sites to provide their own >remote FTP type facilities to their cusomers. I've brought this up before, >and been poohpoohed, but so far as I can tell it would solve most of these >sorts of problems. And sites could do their own choking. Why is it that >Princeton in the strange world of BITNET is the only place that thinks to >do this? >