Xref: utzoo news.software.nntp:1287 news.software.b:7571 Newsgroups: news.software.nntp,news.software.b Path: utzoo!henry From: henry@zoo.toronto.edu (Henry Spencer) Subject: Re: NNTP/News propagation question Message-ID: <1991Apr26.153021.27016@zoo.toronto.edu> Date: Fri, 26 Apr 1991 15:30:21 GMT References: <2657@brchh104.bnr.ca> Organization: U of Toronto Zoology Followup-To: news.software.b (This really has nothing to do with NNTP; I've pointed followups at news.software.b) In article <2657@brchh104.bnr.ca> rgreene@bgrgs1.bnr.ca (Bob Greene) writes: >Now, if I'm sitting at my terminal in Dallas, and I post >something setting distribution to "dfw", what happens? >All of our local Dallas, TX machines see it, then we >pass the article on up to our Ottawa machine, now it's >not in dfw, so does it dump it on the floor even >though all the rest of dfw is only accessible through >it? Much depends on what news system your Ottawa machine is running and how it is configured. The general answer is that it doesn't dump it on the floor, but it may or may not pass it onward. >In more generic terms, since no news machine knows the >geographical locations of it's neighbors, doesn't that mean >to guarantee correct distribution, each news machine >must propagate each news article to all it's neighbors... No, it means that each news machine must be **configured** to know which sites to pass which articles to. It doesn't have to know who's where, but it does have to know who gets what traffic. >even though that article may not be in it's own >distribution? ... There is no notion of a machine having "its own distribution". >I have a feeling that the articles just get dumped on the >floor, since the whole Usenet news system was really >intended to be run with your neighbors being geographically >proximate... Whatever gives you that idea? It's never been true. -- And the bean-counter replied, | Henry Spencer @ U of Toronto Zoology "beans are more important". | henry@zoo.toronto.edu utzoo!henry