chris@grkermit.UUCP (Chris Hibbert) (07/09/84)
I have installed the propogation delay calculator provided by someone on the net, and have been watching its results for a while. It has always showed that a fairly solid 2% of the articles take longer than two weeks to get here. I wondered why, and whether it was a temporary situation due to downtime somewhere on the net. I waited and waited and finally decided to track things down. I added a few lines to the program to print out an article-id every time it hit a new maximum delay. My results (over a period of a month and a half) show that most of the slowest messages to get here start inside of HP somewhere. Here are the New Maximums the program reported getting this time, with the paths for the messages that started from HP. New Maximum: 1 ; <312@mhuxj.UUCP> New Maximum: 4 ; <642@eagle.UUCP> New Maximum: 6 ; <1346@cbosgd.UUCP> New Maximum: 7 ; <1120@sri-arpa.UUCP> New Maximum: 8 ; <1117@sri-arpa.UUCP> New Maximum: 9 ; <53100042@trsvax.UUCP> New Maximum: 13 ; <81100002@hpfclk.UUCP> Path: grkermit!genrad!decvax!ucbvax!ucbcad!tektronix!hplabs! hp-pcd!hp-dcd!hpfclk!marysue Posted: Wed Jun 13 17:25:00 1984 Date-Received: Tue, 26-Jun-84 07:59:14 EDT New Maximum: 19 ; <22700001@ucbcad.UUCP> New Maximum: 21 ; <6500024@hp-pcd.UUCP> Path: grkermit!genrad!decvax!ittvax!dcdwest!sdcsvax!sdcrdcf!hplabs! hp-pcd!john Posted: Fri Jun 8 09:24:00 1984 Date-Received: Fri, 29-Jun-84 06:33:12 EDT New Maximum: 22 ; <68500004@hpfclq.UUCP> Path: grkermit!masscomp!clyde!burl!ulysses!mhuxl!ihnp4!zehntel! hplabs!hp-pcd!hpfcla!hpfclq!robert Posted: Sat Jun 9 13:55:00 1984 Date-Received: Sat, 30-Jun-84 18:15:26 EDT Propagation Delay No. of Cumulative Days Articles Percent 0 5 2% 1 817 32% 2 943 66% 3 518 84% 4 186 91% 5 85 94% 6 58 96% 7 48 98% 8 10 98% 9 12 99% 10 4 99% 11 1 99% 12 1 99% 13 2 99% 15 2 99% 16 1 99% 17 2 99% 18 1 99% 19 3 99% 20 1 99% 21 4 99% 22 3 100% Distribution of news feeds Count Percent Site 1412 50% genrad 1363 49% masscomp 2 0% grkermit Decvax is only two hops away, so I have trouble believing that the problem is getting the articles here. It looks to me like they must be getting lots of places pretty slowly. Does anyone at HP want to look into this? Chris decvax!genrad!grkermit!chris allegra!linus!genrad!grkermit!chris