Slashdot Mirror
New File Sharing Networks
An anonymous reader sends in: "Most readers of slashdot have been following the exploits of the RIAA and their attempts to shut down Napster, KaZaA, Morpheus, etc. In response, it appears some live music fans have taken things into their own hands and started new file sharing networks made exclusively for trading live recordings of bands that allow that sort of thing. The main player, RNL has reached version 1.0, features a distributed architecture, supports linux, and is even GPLed. Another peice of interesting software is Furthur. Though still only in beta, Furthur has cool features like allowing a user to piggy-back another user's download to reduce the load of the uploader."
Haven't these guys heard of etree.org? Etree has been around for a few years, and exists to allow the trading of lossless recordings of live shows from bands who allow trading.
Its not p2p, mostly ftps and burn + post cds, but it has been there for some time. Loads of good shows too :)
This post will enter the public domain 70 years after my death, unless Disney buys another extension.
as a means on facilitating online transfers instead of mail-trades. all .shn's are md5 verified before being allowed to share.
the history of the world
Another good one is WinMX
:-(
I get very good results with this. I have extensively subjected it to my "Matmos" test of file sharing programs, and shown it to be as good as kazaa/morpheus.
The main strength of the program is that it has its own p2p protocol, but also allows you to connect to multiple OpenNap servers at once (unlike napigator). If you are patient, you can log onto a sh*t load of servers and get excellent results.
However, there are some drawbacks. The interface is a little buggy (but a little more for the "power" user than morpheus). Also, getting a good list of OpenNap servers into the program can be a real bastard. I strongly advise looking here and here for solutions to this problem. Also, as the name suggests, there is no linux version
This sort of thing has been going on for a while via FTP and loosely based networks like etree.org trading lossless quality shn files.
It even works for Hillbilly music. Check out www.bluegrassbox.com for an unbelieveable (hundreds of gigs) resource of extreemly high quality audio files in shn format.
And remember, Friends don't let Friends use MP3!
Though I am no byte-level expert, this isn't really anything new or its misleading. What this seems to refer to is how the traffic would be routed. IE, if "Joe" has live Pearl Jam and 3 people request it, the network is smart enough to take bytes from people farther along in the download. Even then, thats more load-balancing.
Otherwise, this is no different from any other P2P filesharing mechanism where files naturally propagte from a source and are eventually downloaded from other nodes. Still, if your network were *smart* enough to resend packets as little as possible (IE, if the network would multicast concurrently-requested packets) then this would be leaps-and-bounds above current P2P.
Furthur is being developed by some of the etree.org crew. Etree.org has an outstanding track record so far. The group trades only "taper-friendly" recordings, and if anyone asks for a recording that isn't taper-friendly, members are sure to jump on the request and tell him to look elsewhere (some are more polite than others).
/.) would be chaos. However, the group is pretty well self-policed.
You'd think that a group of over 13,000 (that was last I heard a while ago, and with the second related story in a week, there are sure to be many more directed from
The fact is, if someone wants an illegal bootleg, there are plenty of other places to look rather than the etree.org lists. Rather than get flamed and endanger etree.org, those people just go elsewhere for those needs.
Check out etree.org's legal page for more information about policy. After 3 years, there still hasn't been an RIAA shakedown.
No, let's say you and 4 friends all have adsl, 640/128. The four friends all have a file you want to download. Getting it from one of them will give you a 128k download, due to their upstream cap. If you grab chunks from all of them at once, you can get it at 512k.
Is another one decentralized and open source. It works through HTTP proxies and firewalls.
http://go.to/netmess
Presumably the record companies defer to the bands on this point.
Well, the way I understand recording contracts is that the recording company owns the recording, but the band still owns the music. The band rarely gets enough of a cut from the contract that sales hurt their personal bottom line. Huge bands like Pearl Jam, Metallica et al are big enough to negotiate that sort of deal. This is why Lars from Metallica was one of the few artists who cared about Napster. If you'll notice the RIAA, not the artists, is suing people. They claim they are protecting the artists, but they are really looking out for their own pockets.
Most bands dont make money on the record, they make money on the concerts and appearances. Granted, they do make money, but it's usually small compared to what they make from other sources...why else would they tour? The recordings basically serve them as advertisements.
The bands typically own all rights to the music itself (not the recording) and they have the right to allow or disallow fans to record concerts.
Usual dislaimers apply...IANAL, esp a contract lawyer, and have never seen a 'typical' recording contract and am just make observations on what I've heard and read.
They that can give up essential liberty to obtain a little temporary safety deserve neither liberty nor safety.-Franklin
I have been using Etree for three or four years now. For those not aware here, etree site-op's release their server content on the etree-announce mailing list periodically. Users can download, through FTP, high quality concerts from folks like The Grateful Dead and other microphone friendly bands. I think what keeps etree pure currently, is that with only a (relatively) few site-ops, control over content, is easily implemented. This will undoubtedly collapse under the massive abuse inherent in peer-to-peer networks.
.shn's of Brittney Spears 2/18/02 Cleveland Show being traded alongside some of music's most influential live bands. For what it's worth check out what is being traded on etree at their database site
It would be quite sad to see
Getting a good list of OpenNap servers into the program can be a real bastard. I strongly advise looking here and here for solutions to this problem.
a good overview of different p2p architectures is over here at openp2p.com.
One system the author fails to mention is Circle, which uses a decentralized hashtable system., more about it at his system is in a pdf slideshow he'll be giving at linux.conf.au
My favorite quote from his page: "FastTrack (aka Kazza/Morpheus) is kind of like trying to optimize a bublesort", which leads me to believe he has a regular quicksort at hand. (actually he does claim O(n log n) seachs, so its about right)
Also to note are Chord and GISP which seem to use simular schemes, where Chord is pure acadamia (someones masters thesis). GISP is an implementation of something from JXTA, suns p2p framework.
this is my sig.
from furthur's website:
All shows on the Furthur Network are 100% MD5 verified! When users initially share a show, Furthur will run an MD5 verification check, to ensure the file integrity. If any file doesn't check out, Furthur won't allow it on the network
This gives them a lot of control over network content. Don't want something on the network? Pull it's MD5 sum from your database.
do not read this line twice.
For Windows users:
There's an Open Source project hosted on Sourceforge called Gnucleus. Here is the project page.
It supports multiple hosts download, so if you were an user of Xolox, but want a client that development still continues and you want to get those large files using multiple connections, get it now. Sadly, download of partial files from other hosts is still not possible (since there's no consensus from the Gnutella protocol developers about how this should work).
Gnucleus even has a LAN mode, so you may run it to share files over your network that has locked ports or net access blocked (great for colleges!).
Buy a Nintendo DS Lite
many may not have heard of SHN vs. mp3 (debates for or against these 2 can cause a war), but SHN is a lossless compression of a WAV file, and it compresses the wav file approximately 50%. This is compared to mp3's where they are lossly compressed about 90%, but it throws out information in the original wav.
For one thing, FLAC performs a few percent better than SHN and has a more free license. For another, tests performed by r3mix.net have shown that it's possible to encode MP3 at a variable bit rate centered about 192 kbps and lose nothing audible. (Whether this is legal under the Fraunhofer patents is a different story.) MP3 and Ogg Vorbis produce significant quality loss in only the following situations: 1. low bit-rate operation, 2. crappy encoders, and 3. repeated conversions of wav -> compressed -> wav.
A lot of the hard-core collectors of the live music refuse to collect mp3's due to the loss in quality from original wav->mp3
What about the loss in quality from analog->wav? It's negligible, but it's still a measurable loss.
Will I retire or break 10K?
So instead of letting go of your Windows apps, you let go of your money. VMware Workstation costs $300, and Windows 2000 Pro for VMware costs $200 [vmware.com] plus the connect time to download the service packs.
Actually, the VMware Express package is still available for download, and the license is a lot cheaper (around 50$, IIRC). Although it will only run Windows98 (and only one virtual machine at a time), that's enough for most Windows apps - and who doesn't have a copy of Windows98 lying around...
According to CodeWeavers' app database, most features of WinMX work
Still, I haven't talked to anyone who got it to work...I'd be interested!
Reminder: find a new sig
I don't think that production and distribution of CDs is the big expense that the recording industry bears, I think its the costs involved in recording and production that are so expensive.
Think of what studio time must cost to have a five-man band and all the people involved. Its not just the band, a recording engineer and a producer. There's loads of other engineers to mic everything, a couple of guys on the console, producer, roadies, catering, not to mention the time to rent the studio equipment. Then there's the guys involved in the mixdown (producer, engineers) and the equipment time. Mastering, etc etc. All expensive, equipment and people intensive.
Production and distribution, as has been stated elsewhere, is a buck or two per disc.
We did a significant amount of work in this vein with Swarmcast and have been building a lot of the (rather complex) technology necessary for Reliable Multicast
Multicast is not panacea though, because it is not very widely deployed on the Internet, and since there is no caching in Multicast, all of the receivers must be downloading near the same time to realize the bandwidth savings...So in many ways P2P caching has advantages over multicast, which is why we do both.
--
Justin Chapweske, Onion Networks
I would imagine that client communicates with a centralized server to check MD5 sums and also check filenames so the only way to actually put up a illegal file for sharing is change its name to something like 11.29.98-Phish-David-Bowie03.shn and post it as a new file so a MD5sum is created.
Why would one even need to look at the filename? Calculate the md5 signature, look it up in the database, and if you find a match, you're good to go, regardless of the filename. An md5 signature is 16 bytes long, that's the same length as an IPv6 address, the kind they describe as being sufficient as allowing every atom on earth to have its own I.P. address. Shouldn't it be vanishingly unlikely that someone could alter a piece of music so that its md5 matches with something previously registered?
That's assuming that someone is actually checking out these files before entering md5's into the database, I guess. Is the safety factor just based on the fact that you won't approve a filename that doesn't match the known list of acceptable bands? I suppose if you can't search for something illegal based on name that it doesn't matter in some sense whether it is in the system or not..
- jon
Ganymede, a GPL'ed metadirectory for UNIX
But, all these issues don't prevent me from running Freenet. I find it comforting to support anonymity on the internet, like in the old days. It's really not up to scratch for P2P file sharing, though. Never will be.
Frost, on the other hand, seems like a really slick attempt at totally anonymous newsgroups built on top of Freenet. I've run it a few times and like it. It's slow as molasses, but that's not Frost's fault (see above).
Any connection between your reality and mine is purely coincidental.
Furthur does md5 checking to group together the files, and once you start downloading it looks at md5's to figure out what it needs to download. AFAIK, All internal file identification is done by md5, and not by file name.
Of the ones that did multi-home download, none ever kept trying to find sources for the files in progress (enhancment/feature?).Furthur will look for other sources every 15 minutes (or whatever you set it to). If you start a download, and the guys you are downloading from all leave, you can just let it sit in the Partial tab. Come back in a few days (or whatever), and the show will probably be fully downloading. You never have to do anything after you click "download".
My biggest beef with all of them is that none could continue a d/l that had stopped. So every time it re-started it would start at the beginning.
Furthur can do this of course. It can piece together downloads from multiple people who already have the file, and even from others who are currently downloading. And if the download is interupted, it will pick up where it left off when a source becomes available.
Only a few of them could resume searching/downloading if the client died(or I killed it) (gtk-gnutella could save the d/l requests)
Furthur can do this, too.
Spammage - you could do a exact title serch, and get hacking info, or porno or ... well you know what I mean. Heavans forbid if any commercial company really got serious about it.
Well, furthur is still dependent on what people put in the file descriptions, so if someone wanted to put a bunch of spam in there, they could. I haven't seen that happen yet, though.
-Mike
PS I'm not afiliated with Furthur in any way, other than as a satisfied user
As far as I know, Furthur (which is closed source I believe) receives a list of appropriate artists/bands and limits you to searches for those bands. So as long as no one reverse engineers the protocol and writes another client they should maintain control of their network. There are already plenty of places for people to get their non-taper-friendly music (in addition to warez and whatever else) hopefully they will use them and leave Furthur alone.
Semper Ubi Sub Ubi
That won't happen unless someone does something stupid like post these URLs to slashdot.
oh....