Xref: utzoo news.admin:6117 alt.config:1094 Path: utzoo!utgpu!jarvis.csri.toronto.edu!rutgers!usc!csun!fedeva!premise!mirror!billc From: billc@mirror.UUCP (Bill Callahan) Newsgroups: news.admin,alt.config Subject: Re: UUCP in Germany Message-ID: <28148@mirror.UUCP> Date: 27 Jun 89 14:12:33 GMT References: <3202@wheat-chex.ai.mit.edu> Reply-To: billc@prism.TMC.COM (Bill Callahan) Organization: Mirror Systems, Cambridge Mass. Lines: 33 In article <3202@wheat-chex.ai.mit.edu> tmb@wheaties.ai.mit.edu (Thomas M. Breuel) writes: -o The idea of a central backbone (unido) charging for all of a country's - USENET traffic seems wrong. The costs that unido has for pulling articles - over the Atlantic should be redistributed equally among its immediate - neighbors in Germany. Each neighbor could then decide to let other - hosts connect to them and redistribute their costs further. - Central network administration (just like central economic planning) - always seems like a good idea at first but from all the examples I - have seen always fails miserably. I agree with this point (as well as many of the other ones) quite wholeheartedly. I understand that unido has costs from getting a U.S. newsfeed which need to be redistributed, but it seems that they have chosen one redistribution scheme which they choose to maintain and enforce, when the possibility exists that another would be superior. The one that Mr. Breuel suggests (having unido only charge its immediate neighbors) seems to have a lot of advantages. One of them is that it takes the burden of enforcement off the shoulders of unido. All they have to do is make sure that the sites they supply directly pay for the service. If a site fails to pay up, they can cut the feed directly. None of this "blacklisting." I am aware that the immediate neighbors would have to pay a steep connect fee, but they can defray the cost by charging connect fees to *their* downstream neighbors. A well connected secondary site could wind up gettting news for *free*, which is fair, since it's providing a service. A little leaf site way downstream may get articles several days late, be probably would have to pay very little for the connection, plus local calling costs, of course. Ultimately, I think a system like this could lead to fairer cost distribution, higher connectivity, and a much simpler (i.e., market oriented) administration.