UO USENET NEWS ARTICLE PROPAGATION PERFORMANCE

This statistical summary is intended to give you a sense of the role we've played in the propagation of USENET News article to other sites around the world. All data is taken from the Freenix Top 1000 USENET Site rankings.

System 9/96 8/96 7/96 6/96 5/96 4/96 3/96 2/96 1/96

Pith #29/5.38% #21/8.18% #23/7.03% #13/9.41% #13/8.32% #37/3.87% #27/4.53% #15/8.12% #20/6.14%

Arclight #17/8.37% #86/2.13% #40/4.29% #31/5.06% #20/6.38% #77/2.08%

Hammer

Blacklab [NR] #665/0.14% #816/0.10% #486/0.23% #362/0.36% #928/0.08%

Platform [NR] [NR] [NR] [NR] #245/0.6% #601/0.16%

Neonlights [NR] [NR] [NR] [NR] #371/0.34% #557/0.18% #736/0.11%

Smalltown [NR] [NR] [NR] [NR] #535/0.20% #977/0.07% #735/0.11%

CS #833/0.10% #718/0.12% #795/0.10% #832/0.09% #418/0.29% #471/0.23% #448/0.23% [NR] #415/0.23%

Sum % 13.85% 10.57% 11.52% 14.79% 16.49% 6.67% 4.98% 8.12% 6.37%

Max Single % 8.37% 8.18% 7.03% 9.41% 8.32% 3.87% 4.53% 8.12% 6.14%

Difference 5.48% 2.39% 4.49% 5.38% 8.17% 2.80% 0.45% 0.0% 0.23%

System	9/96	8/96	7/96	6/96	5/96	4/96	3/96	2/96	1/96
Pith	#29/5.38%	#21/8.18%	#23/7.03%	#13/9.41%	#13/8.32%	#37/3.87%	#27/4.53%	#15/8.12%	#20/6.14%
Arclight	#17/8.37%	#86/2.13%	#40/4.29%	#31/5.06%	#20/6.38%	#77/2.08%
Hammer
Blacklab	[NR]	#665/0.14%	#816/0.10%	#486/0.23%	#362/0.36%	#928/0.08%
Platform	[NR]	[NR]	[NR]	[NR]	#245/0.6%	#601/0.16%
Neonlights	[NR]	[NR]	[NR]	[NR]	#371/0.34%	#557/0.18%	#736/0.11%
Smalltown	[NR]	[NR]	[NR]	[NR]	#535/0.20%	#977/0.07%	#735/0.11%
CS	#833/0.10%	#718/0.12%	#795/0.10%	#832/0.09%	#418/0.29%	#471/0.23%	#448/0.23%	[NR]	#415/0.23%
Sum %	13.85%	10.57%	11.52%	14.79%	16.49%	6.67%	4.98%	8.12%	6.37%
Max Single %	8.37%	8.18%	7.03%	9.41%	8.32%	3.87%	4.53%	8.12%	6.14%
Difference	5.48%	2.39%	4.49%	5.38%	8.17%	2.80%	0.45%	0.0%	0.23%

Month-to-month variation reflects:

changes to server hardware and software,
network connectivity,
network load management policies,
peer additions and deletions,
migration of peers from one server to another,
participation (or non-participation) of peers in the Freenix survey data collection process, and
fluctuations in the flow and composition of news articles themselves.

Note that the CS news system is administered by the Computer Science Department, not by the Computing Center; it is shown here for completeness.

Also note that it is difficult to aggregate total propagation influence across multiple servers because in some cases two or more of the listed servers may appear in a single article's Path header; we do know, however, that the true estimate for the aggregate percentage falls somewhere between the sum of the percentages, and the maximum single percentage shown for any given month.

Joe St Sauver (joe@oregon.uoregon.edu)
October 8th, 1996