Shareaza horizon V gnutella horizon

[humuhumu]

Brief test to get an idea of Shareaza / G2 network against the 'normal' gnutella network.

1. Told Shareaza to use G2, G1 v0.4 and G1 v0.6 protocols. Connected to 4xG2 hubs / peers / supernodes (whatever you like to call them), 4xG1 hubs.

Search for 'britney' on all types of file, allowed to run for ten minutes.

Found 39158 nodes.


2. Same test, ONLY using the G2 protocol. Connected to 6 hubs (all G2 - all it could find after fifteen minutes of looking for eight G2 hubs). Same search, same length of time.

Found 18872 nodes (48.2% of total found before).


So on the basis of that search, it's looking around 50-50. Half of normal G2 / Shareaza results are from clients connected to the G1 protocol.

Not exactly scientific, and the search was done 5am EST, so should expect much higher numbers later on in the day, but gives an idea of the size of gnutella V G2 (or Mikes Protocol, or Shareaza or [insert name of choice]).

Will run the test again later if I get the time...

[Sephiroth]

Its impossible since Mike Protocol("gnutella2") "hubs" also connect to Gnutella so you can try it that way but it will always be a mix of MP and Gnutella hosts..

Horizon is a bad way of trying to estimating the number of people because so many factors influence it and when your done many of your results would be from people who have not been online for a few hours or when you started it. Things like uptime, host connections and how many searches you do can inflate the value..

This method is not efficient and does not work and does not provide any kind of accurate picture at all.