benn@sphinx.UChicago.UUCP (Eldarion ap Aragorn) (11/14/85)
[] Dear Net-readers: Here are further comments on the cost of the USENET. They are followups to the first summary I posted on costs. It is not too late to write in. From John Mackin In article <1293@sphinx.UChicago.UUCP> Erik Fair (ucbvax!fair) writes: > In the last two years, it has been AT&T at various locations, > Hewlett-Packard, and the Center for Seismology Studies in Arlington, VA > that have contributed greatly to the maintainance and expansion of the > network (seismo feeds both Europe and Australia. hplabs feeds Korea. > Ponder the cost of that for a moment). I just felt that I had better point out that the cost of traffic over the seismo!munnari link is borne by us here in Australia, NOT by seismo. Erik seems to imply the contrary. I'd appreciate it if you'd include this in one of your upcoming summaries. Thanks. John Mackin, Basser Department of Computer Science, University of Sydney, Sydney, Australia From: ihnp4!erc3ba!jfka Heart-warming, delightful, risque! Don't miss it! Rated PG From: Phil Ngai Ok, here's one of my pet peeves. People who argue the cost of USENET can be reduced by "improving backbone connectivity" are full of sh**. By definition, a backbone is a site with long-distance feeds. Were connections set up exclusively to minimize long-distance phone calls, hop counts and propagation delays would go up dramatically. Moderated newsgroups will look quick by comparison. What's worse, phone costs would not go down much if at all. The difference in cost between calling Los Angeles vs calling New York from San Francisco is really quite small. (maybe 20%) And there are many cases where a path using only local phone calls simply can not be found no matter how high a hop count you are willing to accept. Those of us who live in densely populated areas like Silicon Valley may make the mistake of thinking the whole country is like your home. Well, consider Kansas, or Idaho, or Nevada, etc. Long distance calls are going to have to be made. And there's a lot more to a feed than efficiency considerations. Intra-company communications, vendor/customer relationships, and personal/professional relationships are probably the dominant reasons the net has its current topology. And these can not easily be changed for something with as low a signal to noise ratio as USENET. In summary, I claim people who call for "backbone connectivity improvements" are as credible as Ronald Reagan pledging to "reduce waste in government". Both are attacking a mythical problem, mythical in the sense that actual cases are the exception rather than the rule. It is such an obvious target that undetected cases are very unlikely and we will spend our energy better by working on other things. Phil Ngai {ucbvax,decwrl,ihnp4,allegra}!amdcad!phil This was brought to you as a USENET public service by an end-site reader who's tired of sponging. Thomas Cox ...ihnp4!gargoyle!sphinx!benn
rick@seismo.CSS.GOV (Rick Adams) (11/15/85)
While munnari and mcvax do indeed pay the per packet charge and connect time, I should point out that the $1000/month that seismo pays for the x.25 link is hardly insignificant. This $1000 is before anything is sent and is for the "priviledge" of being conencted to Tymnet. As an aside, the international rate structures are so screwed up that if seismo initiates the call, it costs at least twice as much as if mcvax or munnari initiates the call. ---rick