Thx Moak.

What I actually intend to do is to prove the scalability of the Gnutella networks.

One thing is to analyse the network structure to do so. What I try to find out is how network protection mechanisms of clients may influence my measurements that should prove the theoretical results. That is what this is all about.

I send out a Ping every 5 Minuts with a TTL of 7-20. Interestingly I get some "strange results": Some clients attached do not answer at all (which could mean that they disconnected but logs show they are still there) some returne 1500-2500 pongs which I also consider as too much ?!?

Here are some entries of my log:


Dez 10 20:01:34 export qtella logging[3329]: Servent 128.103.189.195 ## 0:1 ## 1:3 ## 2:1 ## 3:8 ## 4:1 ## 5:50 ## 6:41 ## 7:0 ## 8:0 ## 9:0 ## 10:0 ## 11:0 ## 12:0 ## 13:0 ## 14:0 ## 15:0 ## 16:0 ## 17:0 ## 18:0 ## 19:0 ## 20:0 ## total: 105
Dez 10 20:01:34 export qtella logging[3329]: Servent 24.253.133.117 ## 0:0 ## 1:1 ## 2:5 ## 3:4 ## 4:9 ## 5:9 ## 6:8 ## 7:53 ## 8:0 ## 9:0 ## 10:0 ## 11:0 ## 12:0 ## 13:0 ## 14:0 ## 15:0 ## 16:0 ## 17:0 ## 18:0 ## 19:0 ## 20:0 ## total: 89
Dez 10 20:01:34 export qtella logging[3329]: Servent 80.19.204.186 ## 0:1 ## 1:4 ## 2:6 ## 3:19 ## 4:20 ## 5:92 ## 6:119 ## 7:0 ## 8:0 ## 9:0 ## 10:0 ## 11:0 ## 12:0 ## 13:0 ## 14:0 ## 15:0 ## 16:0 ## 17:0 ## 18:0 ## 19:0 ## 20:0 ## total: 261
Dez 10 20:01:34 export qtella logging[3329]: Servent 158.252.215.47 ## 0:1 ## 1:0 ## 2:0 ## 3:0 ## 4:0 ## 5:0 ## 6:0 ## 7:0 ## 8:0 ## 9:0 ## 10:0 ## 11:0 ## 12:0 ## 13:0 ## 14:0 ## 15:0 ## 16:0 ## 17:0 ## 18:0 ## 19:0 ## 20:0 ## total: 1
Dez 10 20:01:34 export qtella logging[3329]: Servent 12.89.79.21 ## 0:1 ## 1:2 ## 2:1 ## 3:32 ## 4:1 ## 5:8 ## 6:37 ## 7:0 ## 8:0 ## 9:0 ## 10:0 ## 11:0 ## 12:0 ## 13:0 ## 14:0 ## 15:0 ## 16:0 ## 17:0 ## 18:0 ## 19:0 ## 20:0 ## total: 82
Dez 10 20:01:34 export qtella logging[3329]: Servent 128.119.246.197 ## 0:0 ## 1:0 ## 2:0 ## 3:0 ## 4:0 ## 5:0 ## 6:0 ## 7:0 ## 8:0 ## 9:0 ## 10:0 ## 11:0 ## 12:0 ## 13:0 ## 14:0 ## 15:0 ## 16:0 ## 17:0 ## 18:0 ## 19:0 ## 20:0 ## total: 0
Dez 10 20:01:34 export qtella logging[3329]: Servent 24.81.77.60 ## 0:0 ## 1:2 ## 2:2 ## 3:7 ## 4:4 ## 5:61 ## 6:49 ## 7:24 ## 8:0 ## 9:0 ## 10:0 ## 11:0 ## 12:0 ## 13:0 ## 14:0 ## 15:0 ## 16:0 ## 17:0 ## 18:0 ## 19:0 ## 20:0 ## total: 149
Dez 10 20:01:34 export qtella logging[3329]: Servent 24.49.92.170 ## 0:0 ## 1:0 ## 2:0 ## 3:0 ## 4:0 ## 5:0 ## 6:0 ## 7:0 ## 8:0 ## 9:0 ## 10:0 ## 11:0 ## 12:0 ## 13:0 ## 14:0 ## 15:0 ## 16:0 ## 17:0 ## 18:0 ## 19:0 ## 20:0 ## total: 0
Dez 10 20:01:34 export qtella logging[3329]: Servent 4.61.240.48 ## 0:0 ## 1:0 ## 2:0 ## 3:0 ## 4:0 ## 5:0 ## 6:0 ## 7:0 ## 8:0 ## 9:0 ## 10:0 ## 11:0 ## 12:0 ## 13:0 ## 14:0 ## 15:0 ## 16:0 ## 17:0 ## 18:0 ## 19:0 ## 20:0 ## total: 0
Dez 10 20:01:34 export qtella logging[3329]: Servent 12.255.135.14 ## 0:0 ## 1:0 ## 2:0 ## 3:0 ## 4:0 ## 5:0 ## 6:0 ## 7:0 ## 8:0 ## 9:0 ## 10:0 ## 11:0 ## 12:0 ## 13:0 ## 14:0 ## 15:0 ## 16:0 ## 17:0 ## 18:0 ## 19:0 ## 20:0 ## total: 0
Dez 10 20:01:34 export qtella logging[3329]: Servents logged: 10 ## TTL was set to: 20


Dez 10 20:06:34 export qtella logging[3329]: Servent 128.103.189.195 ## 0:1 ## 1:4 ## 2:1 ## 3:13 ## 4:2 ## 5:46 ## 6:59 ## 7:0 ## 8:0 ## 9:0 ## 10:0 ## 11:0 ## 12:0 ## 13:0 ## 14:0 ## 15:0 ## 16:0 ## 17:0 ## 18:0 ## 19:0 ## 20:0 ## total: 126
Dez 10 20:06:34 export qtella logging[3329]: Servent 24.253.133.117 ## 0:0 ## 1:0 ## 2:0 ## 3:0 ## 4:0 ## 5:0 ## 6:0 ## 7:72 ## 8:0 ## 9:0 ## 10:0 ## 11:0 ## 12:0 ## 13:0 ## 14:0 ## 15:0 ## 16:0 ## 17:0 ## 18:0 ## 19:0 ## 20:0 ## total: 72
Dez 10 20:06:34 export qtella logging[3329]: Servent 80.19.204.186 ## 0:1 ## 1:0 ## 2:6 ## 3:1 ## 4:8 ## 5:6 ## 6:34 ## 7:0 ## 8:0 ## 9:0 ## 10:0 ## 11:0 ## 12:0 ## 13:0 ## 14:0 ## 15:0 ## 16:0 ## 17:0 ## 18:0 ## 19:0 ## 20:0 ## total: 56
Dez 10 20:06:34 export qtella logging[3329]: Servent 158.252.215.47 ## 0:1 ## 1:0 ## 2:0 ## 3:0 ## 4:0 ## 5:0 ## 6:0 ## 7:0 ## 8:0 ## 9:0 ## 10:0 ## 11:0 ## 12:0 ## 13:0 ## 14:0 ## 15:0 ## 16:0 ## 17:0 ## 18:0 ## 19:0 ## 20:0 ## total: 1
Dez 10 20:06:34 export qtella logging[3329]: Servent 12.89.79.21 ## 0:1 ## 1:1 ## 2:2 ## 3:16 ## 4:4 ## 5:24 ## 6:54 ## 7:0 ## 8:0 ## 9:0 ## 10:0 ## 11:0 ## 12:0 ## 13:0 ## 14:0 ## 15:0 ## 16:0 ## 17:0 ## 18:0 ## 19:0 ## 20:0 ## total: 102
Dez 10 20:06:34 export qtella logging[3329]: Servent 128.119.246.197 ## 0:0 ## 1:0 ## 2:0 ## 3:0 ## 4:0 ## 5:0 ## 6:0 ## 7:0 ## 8:0 ## 9:0 ## 10:0 ## 11:0 ## 12:0 ## 13:0 ## 14:0 ## 15:0 ## 16:0 ## 17:0 ## 18:0 ## 19:0 ## 20:0 ## total: 0
Dez 10 20:06:34 export qtella logging[3329]: Servent 24.81.77.60 ## 0:0 ## 1:2 ## 2:2 ## 3:7 ## 4:3 ## 5:26 ## 6:25 ## 7:31 ## 8:0 ## 9:0 ## 10:0 ## 11:0 ## 12:0 ## 13:0 ## 14:0 ## 15:0 ## 16:0 ## 17:0 ## 18:0 ## 19:0 ## 20:0 ## total: 96
Dez 10 20:06:34 export qtella logging[3329]: Servent 24.49.92.170 ## 0:0 ## 1:0 ## 2:0 ## 3:0 ## 4:0 ## 5:0 ## 6:0 ## 7:0 ## 8:0 ## 9:0 ## 10:0 ## 11:0 ## 12:0 ## 13:0 ## 14:0 ## 15:0 ## 16:0 ## 17:0 ## 18:0 ## 19:0 ## 20:0 ## total: 0
Dez 10 20:06:34 export qtella logging[3329]: Servent 4.61.240.48 ## 0:0 ## 1:0 ## 2:0 ## 3:0 ## 4:0 ## 5:0 ## 6:0 ## 7:0 ## 8:0 ## 9:0 ## 10:0 ## 11:0 ## 12:0 ## 13:0 ## 14:0 ## 15:0 ## 16:0 ## 17:0 ## 18:0 ## 19:0 ## 20:0 ## total: 0
Dez 10 20:06:34 export qtella logging[3329]: Servent 12.255.135.14 ## 0:0 ## 1:0 ## 2:0 ## 3:4 ## 4:3 ## 5:18 ## 6:12 ## 7:0 ## 8:0 ## 9:0 ## 10:0 ## 11:0 ## 12:0 ## 13:0 ## 14:0 ## 15:0 ## 16:0 ## 17:0 ## 18:0 ## 19:0 ## 20:0 ## total: 37
Dez 10 20:06:34 export qtella logging[3329]: Servents logged: 10 ## TTL was set to: 20

As you can see i.e. the marked servent does not responf but is still connected... I got some thousand log entries ... showing that he is still connected.

So what I try now is to see how some of the network protection mechanisms influence the results or where results like no replies at all come from and from that how I can use my logs for statistics (in a correct way).

Perhaps you wanna give a grafical visualisation of your log? Sorry, too tired to understand now (long night)... maybe tomorrow.

hmm, the good old "Is Gnutella scalable?" question.

There are some links but I never investigated that problem, because I personally thought Gnutella was technically never scalable. In reality it is or it seems to be, because of horizons and host caches. Perhaps someone can explain what Gnutella scalabilty means, no really. (We have TTLs and Horizons allready, which doesn't mean every soul in universe can't use Gnutella. Everyone can, in a horizon. Together with superpeers, flow controll, caching and other improvements horions can be increased, improved, dynamic or crosslinked... but we still have horizons, right?)

http://www.google.com/search?q=Gnutella+scalable
http://www.darkridge.com/~jpr5/doc/gnutella.html
http://www.gnutellameter.com/gnutella-editor.html
http://www.gnutella.com/forums/dev/20

I think it's time that the Gnutella protocol is officially updated. Or perhaps make a seperate 'Gnutella v0.x' protocol altogether? The network isn't as efficient as other types, and no matter how many subtle changes developers introduce with their clients, there will chaos, and everyone will be using a different idea or implementation. And seeing the number of new clients appearing everywhere, it would be best to have them adhere to a better standard than the out-dated 0.4 protocol.

About pong caching: It's used in Sparky beta version only.
Crohrs wrote: "LW 1.9 does not use pong-caching, mainly because we have enough new things in it to keep us busy. :-)" http://www.gnutellaforums.com/showth...2&pagenumber=3

Perhaps in the future LW will work on it. I know that right now, things are very hectic at LimePeer. They're got several bugs in the their UltraPeer system. They are all fixable, though.
:-D
A major problem they have too is the Mac version. It doesn't save files correctly, and people can't open them. Once these things are out of the way, perhaps they'll have time to do pong-caching.

:-)