Hey there,

The subject of my diploma thesis I am currently working on is the scalability of P2P networks. One part of it is to verify the mathematical results of the PhD I am working with that show that a Gnutella (I mean the protocol here) based P2P network IS scaling, not as sometimes mentioned in other articles on that issue. The idea to show the results in ar real network is the following: We will send Pings to our connected servants and wait (let's say a minute) for Pongs from the net. From the TTL parameter in the packets one could determine how many servants are how many hops away from ones client.

The thing is that I can think of some issues that may cause problems doing that: Is there something like an agreed-on TTL for Ping? And if so would it cause problems if I used a larger TTL for my Pings to reach deeper into the network? What I think of is that some Clients could/would not respond if the TTL is "too high" or that they will decrease the value to something they know as "normal".

Are there any of these problems I mentioned? I am currently hacking Qtella 2.3 (Linux) and integrate logging features to it.

I would be pleased if I could get some feedback if other clients could fudge the experiment in a way I mentioned ...

Thanks,
Felix

Ask the_GDF, it's a mailinglist where you find all client developers. Ask Google or write to the_gdf-subscribe@yahoogroups.com

I've commonly seen packets with a TTL of 7, but clients may arbitrarally lower TTL's over 5 to 5.

See "Ping & Pong? What for?"
http://www.gnutellaforums.com/showth...&threadid=5213