Take a look at something you d/l'ed, right on the news-client, and you'll see right at the top of it the 'magic' behind usenet, i.e., the 'peering'. Lets say you d/l'ed from giganews, so that is the last place the file was 'recieved', and then the 'path' is listed down to where it originated. Here's a simple one:
Path: border1.nntp.dca.giganews.com!nntp.giganews.com!nx02.iad01.newshosting.com!newshosting.com!post02.iad01!not-for-mail
So, the file originated at newshosting, went right straight to Giganews. One 'hop' in the electronic 'bucket-brigade'.
How about this one:
Path: border1.nntp.dca.giganews.com!nntp.giganews.com!news-in-01.newsfeed.easynews.com!easynews.com!easynews!news-out.octanews.net!indigo.octanews.net!auth.brown.octanews.com
Got an extra 'hop' in there, originated from octanews, then went through easynews, then on to Giganews.
There can be a dozen or more 'hops' in there, eventually every news-server gets the file. Exactly how it keeps track of it is that every piece of every file has a unique ID number, the news-servers may see that a single piece of a file (say 1 part out of 130+, say) is missing, and it'll query each of the news-severs that peer with it for that missing piece. Generally, the 'closer' (single hop or better) ones posting server is to one of the majors like newshosting, easynews, or giganews is, the better, but over the last few years things have gotten much better, so much so that posting in multiple groups is simply NOT nessesary, and any more par files than 5% is a real waste of time (unless the posting server is a real dog).
FYI. that unique ID number is generated by the news-client posting the file, NOT the server; you can see that unique message ID, again, in your newsreader, i.e.:
[email protected]
That's 25digits long, each digit is 40+ alpha-numeric-symboltype, so figure out the unique combinations. 10^xxx power. 40x40x40x..... 25 times.