G.Web.Cache improvement
 

Gnutella.Web.Cache is a nice innovation, sense it is just a list of ip addresses it is not like you can sue it. Now to improve the gnutella network it is necessary to extend what g.web.cache does.

Along with the ip address of a user display bandwidth, how may calculations the computer can do and how many peers it can handle.
This information would be gathered by the compatible clients and sent to g.web.cache and updated when the necessary. Now when a user opens up a gnutella client It will attempt to connect to the best clients available cutting down on the time it takes to connect to gnutella, which is a common complaint. Also the g.web.cache could say what gnutella protocol it uses maybe ban .6 considering every gnutella client that is still in development uses g1. I read somewhere that one problem with gnutella is that people are still using old clients I don't know if that is true or not. I am concerned that this function could be modified and ban a client or a protocol like G2 which would be damaging.

There is a phenomenon with gnutella I believe you call it "island effect" where there may be 1 million people on gnutella you only connect to 200 or so people. This is a little more advanced and it would require a more powerful server to handle g.web.cache but it would drastically improve the gnutella network.
On top having information about each user it should also tell who is connected to who, so lets say some one with a fairly high speed DSL with a new computer logs on, the client will look at g.web.cache and make share that it will connect to at lest one another high speed user(I believe you call them ultra peers or hubs or what ever) so eventually it will link together all the high speed users and help eliminate the "island effect" in the network. Now when some dial up user logs on it will connect to what ever it can.
This is an the best architecture that I can think of for gnutella. I am share that some one can figure out a better one if g.web.cache listed what users each user is connected to.
naturally I think all the major clients should work together on this project (BearShare, LimeWire, xolox, gnucles, morpheus and shareaza)

I have seen a program that turns a computer into a "pong server" I assume it is a program that turns your computer into a ulta peer. I think this idea should be expanded. I am sure that the majority of gnutella users know nothing about pong servers which is a shame. Like if you say "I run a DC hub" or a hotline or even a eDonkey you are pretty much a p2p god. But if you say "I run a gnutella pong server" people won't know what you are talking about if they even care. I think all the major clients on there front page in big letters saying "Do you have high speed internet? If so have you thought about setting up a pong server?" The software should be compiled for Win, Mac, and as many *nix as possible. I know for a fact I can get it on a computer on a T3 line that is on 24/7 and I am sure that there is at least one person at ever univercity in the US if not most of the world that can do the same.
You should have a stelth version. Where you open it and it just uses idle resorces and no one can tell it is running there should be some command to open it and edit the options but you should be able to change the command. As well as open every time the computer is turned on. (play dumb say something like "in case you don't was to hassle with a pong server use this version")
G.web,cache should know the difference between a pong server and a regular user. Maybe the stelth version will just do calculations for g.web.chach.

I support all of p2p however as far as decentralized networks go I feel that eDonekey has much more potential.

stealth version
 

I am Not Claiming to know all the things your article spoke of in detail but it would seem to be tricky to keep this secure and hack free when it continously searches everyones computer again i don't know much about security or hacking but is just concerned me as a gigantic potential thread maybe I am overreacting out of lack of knowledge ?

Re: G.Web.Cache improvement
 

Hi, I myself am a GWC developer. I'm coding Lynn.
I have thought about caching more information like speed, file and library size. But after a short discussion with Hauke (the 'inventor) we both decided to remove it.
GWC should be simple and fast. That's why it was invented.
Also these infos could be faked by some wicked corporations :)

G2 is not a Gnutella Protocol. Only Shareaza supports it.

A Pongserver is not an ultrapeer. A pong server only gives other people IPs and info about other gnutella node it can connect to, like a GWC. An ultrapeer brings a good structure into the gnutella net.

Maybe you should read up some info about gnutella (specs)
http://groups.yahoo.com/group/the_gdf/files

Re: Re: G.Web.Cache improvement
 

Well I am glade to see that the developers a gnutella are smarter than me :)

I am aware that g2 is only supported by shareza and such groups as limewire and BearShare oppose it. However if you look on www.zeropaid.com it is the number one files sharing program. (Well it was for like a long time and it will be again in a day or so currently emule is #1) I disagree, I think G2 is a gnutella protocol in the since that it uses GWC and acts very similar. Calling G2 something outer than gnutella would be like calling g1 something other than gnutella form the standpoint of g.6. Of cores g1 may serve the community better than G2 for the argument is that G2 is not scaleable, which it may not be. Regardless of who is right or wrong G2 is showing grater results now if the results continue that is up to debate and only time will tell. However I feel that the developers that don’t agree with shareaza’s G2 are hurting them selves for at the very least they should look at the code and incorporate it with there current projects to improve gnutella and make G3 (out of spite of cores). My only hope is that the users benefit from this quarrel

I was simply tiring to think of a way to organize gnutella, which seems to be gnutellas biggest weakness. The more server based a p2p network gets the faster and more reliable it gets or so it seems. However the centralized a networks is the easer it is to be shutdown by the AA's. So my solution was to create a server that helped a network as much as possible but be indirect. I feel that improving GWC could serve that purpose.

I also feel that gnutella is failed to harness the power of idle speed on highspeed peers and the willingness that highspeed or people with access to highspeed to contribute to gnutella. If that failure is because of the lack of notification to the users about such program as a pong server than that is easily rectified. Such networks as eDonkey, Hotline, and DC are great examples of the willingness to contribute. I know that the developers of gnutella can tackle this problem better than I. But I feel the need to give back in whatever way I can.

I also under stand the malevolence of corporations, I also understand the damage that could result in such a case, and why you ceased the project for that reason. However that should not stop development of a way to organize gnutella for that is its greatest problem.

As far as GWC interrogating each peer that would be absurd, the client would have to do all the work and report the information to GWC. Of cores staying constant with the good will of the gnutella people it would not be mandatory however it should be strongly recommended and the source to do so should be made available.

well you could read some more papers about gnutella,
keep up with the gdf discussions and maybe help improving
clients. :) some of them are opensource, so you can download
it and make some changes.

OMG I was about to pm you the link and you already replied to what I said.
I must say I am impressed.
Yes, I think I will read more into gnutella. Like I said I feel that eDonkey has more potential but eDonkey dose depend on the goodwill of highspeed users, which would make it less scaleable than gnutella. I guess the name of the game is scalability so I will start to read more into gnutella and maybe even help in development. Also eDonkey doesn’t seem to be interested in what there users have to say.