In my last post I was firstly trying to address the issue of whether good anonymity was possible over a P2P network, and in terms of strict possibility bandwidth isn't relevent. But, I'm glad to see you agree that it is at least possible. Even with the state of bandwidth as it is today, though, I still think it could be implemented. Just put a bandwidth cap on the proxy at say 20% of the machine's total bandwidth. The main incentive for doing this, in my mind, is not to allow illegal distributing of copyrighted materials (not that I favor copyright), but to allow people censored by their governments access to content they can't view directly. Simple web page downloading is not an enormous hit on most hosts, and if some people want to download or share music and video through proxies, they will have to trade bandwidth for anonymity (but the network in general shouldn't suffer). Also, the mere fact that the network allows proxies offers protection in itself since it would be difficult to prove that a particular person is sharing illicit material and not a person proxying through him/her, even though it may seem very obvious.

The assumption is that most people won't dump their TCP stream (either because they lack the sophistication or they just don't care). And encryption *is* still then an issue. You can't decrypt public-key encrpyted data without the *private* key. Even the person who does the encryption can't decrypt without the private key. The public key encrypts, the private key decrypts. The public key might be returned with the query so if the proxy computer is just for file transfers the proxy computer wouldn't see it. But, presumably, it could just ask for it. It still then has to crack the encryption to know the content. You could have system that disposes of each public/private key pair after each use, which might prevent the proxy from even getting a public key.

As for whether you could be shut down for proxying, the DMCA isn't quite clear and it might depend on the exact implementation. For it to be contributory infrigment there has to be financial benefit in letting the transfer procede (as there was with Napster). For it to be vicarious infrigement there has to be knowledge of the violation, and if the data is encrypted no one should be expected to decrypt it to find out. And there are substantial legitiment uses of the technology (such as giving censored peoples access to the free world, whistle blowing, etc.). Let me refer you to
http://www.eff.org/IP/P2P/Napster/20...ite_paper.html
and you can judge for yourself.

As I said, cap the proxy bandwidth at 20% and only let high bandwidth hosts proxy. Those who want anonymity will have to sacrifice bandwidth, but those who don't care shouldn't be affected.

The GQ rating could be made to work possibly by making it a more complicated formula based on connection speed, and number of files downloaded/shared. Modem users might get a class/rating of their own (say muGQ) and be exempt, while cable users or aothe broadband users could all be rated on the same level.

Also, it would be hacked and tinkered with within days of its release, and that could not be helped. Clients could be made wich would give out 100% GQ ratings all the time, as well as report being a different client (so you wouldn't know of these false ratings). So the GQ system is mostly plausible as a good faith / honor system / voluntary system. But, the only people who would cheat it are malicious broadband users, and why!?

Just my thoughts.

Thanks for straigtening that out ... let me restate what I *meant* to say, with this different (and less convoluted) understanding of which key goes where for what purpose. The policing computer sends the request for files, along with the newly generated public key that its client has generated, retaining the private key. The file gets sent, encrypted, by the direct infringer. The proxy passes it along. Policing computer gets the file, decrypts it, and discovers it is a file that its company owns the rights to.

See the problem? You can't create an open network that also excludes the record companies and software developers. If Joe Blow can get online and get a copy of Office XP, so can Microsoft. Step one of the sucussful action against the proxy is complete -- we've proven direct infringment.

No, for vicarious infringement, the person either has to know, or *should have known* the infringement was occurring. All you have to do is fire up a gnutella client to demonstrate that the person running the proxy *should have known* that it was being used to infringe on copyrights. The vast majority of the material is copyrighted. You know it, I know it, they know it. And if you don't know it, you should have known it.

Finally, the person running the proxy clearly contributed to the infringement by providing the proxy and thereby providing annonymity and connectivity to the other parties involved.

All it takes is one lawsuit. You don't have to sue everyone. Once you've succeeded against one poor schmuck, you could use the precident to get several proxies per day shut down.

Who would be stupid enough to run one of these proxies? Would you do it?

I'll say it again: The only thing that can save you from the big bad wolf is the law of large numbers. That means you need as many people actually sharing files as possible, and you need to do it in as decentralized a manner as possible. Proxies would centralize the network and give great big red targets for the record companies to shoot at.

The moral of the story is: SHARE YOUR FILES!

The other moral is : GIVE US A TOOL TO DETERMINE WHETHER OR NOT PEOPLE DOWNLOADING FROM US ARE SHARING THEIR FILES!

(y'all remember that this thread was originally about freeloaders, right?)

-Lurker

The method I was suggesting is a little different, but what you are describing should work as well. Queries on gnutella are already anonymous, it is only file transfers that are not (unless by random luck you get a hit after only one hop). So the proxy is only for file transfers and has nothing to with the passing of queries or public keys. My thought was to send a public key with the response rather than the query (a little easier on bandwidth it is ). But it doesn't matter who's public key it is so long as it can get encrypted on one end and decrypted on the other. So, the policing computer gets the file and finds that it's copyrighted. For all they know the proxy computer is serving the file. Suppose by dumb luck they own both the proxy and the requesting computers. They still have to prove that the computer seen by the proxy is the one serving the file and is not a proxy itself. For a large network like gnutella, owning two computers involved in an infringing transfer is going to be rare. Given the unlikelyhood that they could even prosecute in that case, I don't see the RIAA (or whoever) investing so much in computing power and bandwidth for such a futile cause. It's much more likely in my opinion that they would try to have passed new laws that would make the technology illegal.

I don't follow you. How could a proxy be expected to know that there is an infringment if the stream is strongly encrypted? ISPs aren't held responsible for the data they forward and most of that data *isn't* encrypted. Just recently eBay won a case concerning culpabilty for copyright infringement over their system (first time the DMCA has been thwarted so far in court in its short history).


He/she may be contributing, but contributing by itself isn't illegal. Under the DMCA a contributor can be found guilty of either contributory or vicarious infringement, as you're aware. But to be found guilty of either you have to violate *all*, not any, of the elements of either case. A requirement of contributory infringement is knowledge of a direct infringement (by no stretch can a user be expected to know this from an encrypted stream), and a requirement of vicarious infringement is financial benefit. I don't see at all how a proxy, using the network layed out above, could be found guilty of infringement under the DMCA. The law of course may change, but that isn't the issue. And, again, there are substantial non-infringing uses.



Who would be stupid enough to share copyrighted materials over Gnutella? Would you?
Personally, I think I'd have a much better legal case, if not a perfectly sound case, with the proxy.


And yes, on the original point, everthing possible should be done to see that users share more files

I'm new to these forums but I wanted to share some of my views on sharing files.

For me, I set my client to download many files at once, knowing full well that if I'm lucky, I might get one or two per day fully completed. I leave the service on almost all the time and have it set for 3-4 simultaneous uploads at 1 file each.

My problem is that I have one computer (notebook with 18Gb capacity) pretty much dedicated to this function while I'm away. When I'm home, it also serves as my main workstation. Most of the files I download are video-related, however I do have several games and utilities as well.

FILE SIZES & LIMITED STORAGE
With the size of files I'm working with, and the limited storage I have in relation to those file sizes, I can't keep increasing my shared collection size indefinitely without running out of space and crippling the performance of my computer. Eventually, some of those shared files will have to be relocated or deleted.

VALUE OF FILES IN A COLLECTION
Depending on the nature of a file's contents, it's value to me may differ greatly. Files which I wish to keep for a long time will stay in my collection, however again we come to the argument about limited storage space -- it can't go on forever without consequences. Usually these files may end up on CD-R. Files which I have a moderate interest in (say a song or a movie), depending on the subject may get one listen or viewing. If it's crap, I'll delete it (hence removing it from the shared collection) -- again, I can't afford to overload my storage. If it's something I might want to keep, perhaps CD-R. A consumers we need to realize that the value of any particular file is not static - it's fluid. It changes over time, and that value may directly impact how long it stays in a shared collection.

PARTICIPACTION
One thing I try to do is leave as many seeming popular files in the shared collection as long as possible. I even keep mental notes on which files get uploaded the most and try to give people as much opportunity to share what I have available, as long as it doesn't impact too much on my own utility of my computer. As the storage fills up, I have to make judgement calls on what to keep and what to turf from the collection to make room for new downloads. Sometimes that means removing a very popular file. There have even been the odd occasion where I've had to halt someone's upload midstream (and remove one or more shared files) because my new downloaded files were so numerous that I again was close to overloading my harddrive. Sure I felt bad about it, but what can I do (I use Gnotella)? I wouldn't like that if someone did it to me (and I'm sure they have on numerous occasions for whatever reasons).

DEFRAGMENTATION
Recently I discovered that my harddrive was severly fragmented and the notebook's overall performance was drastically getting worse. The harddrive was so full, there was not enough empty space to use as temporary working space -- hence the defragmentation process never completed. I had to relocate and remove massive amounts of files elsewhere temporarily just to get the defrag to complete. I also took that opportunity to cull several large files from my collection to make room for potential new downloads.

All I'm saying is that the issue of sharing files is not cut & dry. It goes beyond the issue of some people not wanting to share.

~Artedium

I don´t think all freeloader are bad!

This discussion between Moak and me shows a few ideas against freeloading, and why people "free" load:

http://www.gnutellaforums.com/showth...&threadid=4613

Morgwen