IP#s versus Users
One of the new features of the 2.6 sheep client is that each installation uniquely identifies itself to the server. That means I can finally count the number of users, rather than estimate it by counting unique IP addresses. After a couple weeks of beta activity I have some preliminary results: 2197 IP addresses to 1113 users. That means counting addresses is off by about a factor of 2 in the long run. When measured for one day (April 18th) it came to 505/472 = 1.07. I assume the difference results from users getting new IP#s as the days go by.
So when the 2.5 server was clocking 10000 addresses per day, it was really only getting 9300 users, if these numbers hold.
Posted by spot at April 20, 2005 06:25 PM