Xref: utzoo news.misc:1491 news.config:640
Path: utzoo!linus!husc6!cmcl2!phri!roy
From: roy@phri.UUCP (Roy Smith)
Newsgroups: news.misc,news.config
Subject: Re: The USENET Backbone (Updated: 17 May 1988)
Summary: Some backbone links appear to be bogus
Message-ID: <3283@phri.UUCP>
Date: 25 May 88 03:33:48 GMT
References: <4118@medusa.cs.purdue.edu>
Reply-To: roy@phri.UUCP (Roy Smith)
Organization: Public Health Research Inst. (NY, NY)
Lines: 52

spaf@cs.purdue.EDU (Gene Spafford) writes:

>            husc6
>              |              [If condensing somebody's words is paraphrasing,
>  philabs---cmcl2            does this constitute paragraphing?]
>              \
>             rutgers

	Our main feed is cmcl2.  I did some stats on the paths of the
articles we had laying around in our /usr/spool/news.  Out of 14007 files
(not articles; no attempt was made to only count cross-posted articles
once) I got 13834 paths coming directly from cmcl2.  Of those paths, here's
the breakdown by third site:

8006 nrl-cmf	
4148 husc6	
 559 brl-adm	
 553 rutgers	
 274 beta	
 262 yale	
  11 arizona	
   5 vx2	
   5 acf8	
   4 acf5	
   2 esquire	
   2 acf3	
   1 polyof	
   1 kaplan	
   1 acf2	

	While I'm willing to believe that the majority of the traffic over
the philabs link might be in cmcl2->philabs direction for some reason, I
find it amazing that not a single article came the other way.  Also, it
seems that the vast majority of the articles reaching cmcl2 come from
nrl-cmf, a link which doesn't show up at all.

	OK, I can understand that maybe nrl-cmf doesn't want to be listed
in the backbone map.  But, when two of the three official backbone links
only supply a total of 4% of the incoming news flow, and all three only add
up to about 30%, the only assumption I can come to is that something is
wrong.  If my observations for my local bit of the net hold true all over
then I must conclude that the backbone map, while pretty to look at, really
isn't worth much.

	It would be interesting if some other folks in different parts of
the net ran the same kind of stats (doesn't take much more than some egreps,
seds, sorts, and uniq's, and a bunch of CPU time) and posted the results.
-- 
Roy Smith, System Administrator
Public Health Research Institute
455 First Avenue, New York, NY 10016
{allegra,philabs,cmcl2,rutgers}!phri!roy -or- phri!roy@uunet.uu.net