Kuro5hin.org: technology and culture, from the trenches
create account | help/FAQ | contact | links | search | IRC | site news
[ Everything | Diaries | Technology | Science | Culture | Politics | Media | News | Internet | Op-Ed | Fiction | Meta | MLP ]
We need your support: buy an ad | premium membership

[P]
Another Filesharing App Borked

By Dlugar in News
Tue Oct 02, 2001 at 10:11:35 PM EST
Tags: Internet (all tags)
Internet

One of the best peer-to-peer filesharing networks out there is the FastTrack system, used by KaZaA, Grokster, and Morpheus. However, these are all quite proprietary, and severely dependent on Microsoft Windows and Internet Explorer.

giFT, a "Generic Interface to Fast Track," was an [admittedly crappy] interface that worked under Linux--that is, until last weekend.


The FastTrack network is particularly awesome for two reasons. First, meta-data is stored and searchable apart from the filename. For example, when I do my search for Simpsons episodes, if I find one that's not approximately 20 minutes long, I know it's not complete. The artist, title, album, bitrate, genre, and keywords are all among this extra meta-data, making searches much easier and more valuable.

Second, the use of a two-level "supernode" network makes the network much more scalable than the Gnutella network. On the FastTrack network,
Initially, every peer joins the network as a node. If it reports that it has enough bandwidth, low latency, etc., it may get automatically promoted to a supernode. Each non-supernode on the network has its supernode that it talks to. Nodes tell their supernodes what files they are willing to share, and send their search queries to them. Supernodes communicate amongst themselves, handling search queries (each supernode knows the collected shared file lists for all of its nodes). This self-organizing two level hierarchy allows the network to scale much better than, for example, Gnutella. [from the giFT README]

In short, all the benefits of the decentralized Gnutella network, with none of the scaling problems--until last weekend, that is. I had been trying to connect to the network rather unsuccessfully, but only today did I go to the giFT site to see what had happened. This is what they say:
The latest version of Kazaa and Morpheus, version 1.3.3 has broken giFT. We can no longer connect to the network and communicate with other supernodes or nodes. We believe that the protocol was changed in such a way that you must now log into a central server to get a new "key" for generating the cipher state for encryption and decryption. This was a bad move by FastTrack, as it now makes it's network reliant on a centralized server, and possibly puts them in a situation similar to Napster.

Even though the makers of the FastTrack network seemed interested in an open-source Linux version at first, they supplied the giFT developers with no specifications and no assistance at all. Some "anonymous donors" apparently gave the giFT people what they needed to get the project off the ground, helping them to reverse-engineer the protocol. But now, without warning, all their work has been flushed down the drain. "Back to square one," they say. A truly disappointing move on FastTrack's part to centralize the servers, and perhaps one that will lead to the network's ultimate demise.

Sponsors

Voxel dot net
o Managed Hosting
o VoxCAST Content Delivery
o Raw Infrastructure

Login

Poll
For my filesharing needs, I use:
o Gnutella 32%
o A FastTrack client (KaZaA/Grokster/Morpheus) 21%
o Napster/OpenNap 1%
o WinMX 10%
o AudioGalaxy 10%
o eDonkey2000 1%
o Other 21%

Votes: 65
Results | Other Polls

Related Links
o KaZaA
o Grokster
o Morpheus
o giFT
o last weekend
o Simpsons
o Gnutella
o they say
o Also by Dlugar


Display: Sort:
Another Filesharing App Borked | 29 comments (26 topical, 3 editorial, 0 hidden)
I don't get it (4.00 / 1) (#3)
by roiem on Tue Oct 02, 2001 at 06:42:36 PM EST

What about people using older versions of these Windows programs? Can't they connect either? Or am I missing something important?
90% of all projects out there are basically glorified interfaces to relational databases.
So what? (3.50 / 2) (#5)
by greenrd on Tue Oct 02, 2001 at 06:48:09 PM EST

The client is free as in beer, so why should the company care?


"Capitalism is the absurd belief that the worst of men, for the worst of reasons, will somehow work for the benefit of us all." -- John Maynard Keynes
[ Parent ]

Auto-updates... (nt) (5.00 / 1) (#15)
by slaytanic killer on Tue Oct 02, 2001 at 08:16:01 PM EST

(no text)

[ Parent ]
A story (4.07 / 13) (#6)
by babylago on Tue Oct 02, 2001 at 06:59:26 PM EST

Two children are sitting at a table. On the table is a chess set, arranged in the classic starting position-white to the right, queen on her color, and all that. Child One, who owns the chess set, makes her first move, a Queen's Pawn opening. At this point, Child Two rearranges the pieces on his side to resemble the classic starting position for checkers. His first move is to diagonally move a knight one square up and forward, according to the rules of checkers. Child One says, "You can't play like that." Child Two says, "It's just a board, I can play however I want." Child One packs up her chess set and goes off to play with someone else. Child Two begins complaining about his right to play whatever he wants with someone else's chess set.

The incredibly lame story I just told is a substandard illustration of parasitic open-source projects. Child One represents the owners of a proprietary network and software who have worked hard and expect a reward for their services. Child Two represents the Open Source developers who believe that they have a right to use Child One's network for whatever they want, using whatever interface they want. Then, when Child One packs it up and goes home (or in this case changes the proprietary protocol information), Child Two goes on kuro5hin.org and complains about how unfair it is, and how this spells the end of Child One's network.

Same story could apply to Open Source messaging clients, to this whole giFT thing, and probably in the future to the Mono Project. There's nothing new about this scenario.

---
[ Blog | Hunnh ]

Better analogy (4.20 / 5) (#8)
by J'raxis on Tue Oct 02, 2001 at 07:10:14 PM EST

Two children are sitting at a table. On the table is a chess set, arranged in the classic starting position — white to the right, queen on her color, and all that. Child One, who owns the chess set, makes her first move, a Queen’s Pawn opening. Child Two doesn’t know how to play chess, but slowly figures it out. Child One wins repeatedly. They continue playing until Child Two, having learned the game, checkmates Child One. Child One suddenly decides to change the rules of the entire game so Child Two once again has no idea what he’s doing and starts losing again. Naturally, Child Two is now going to start complaining.

— The Raxis

[ J’raxis·Com | Liberty in your lifetime ]
[ Parent ]

Illustrative drift (none / 0) (#14)
by babylago on Tue Oct 02, 2001 at 07:58:05 PM EST

Humour me. What constitutes "winning" in your example?

---
[ Blog | Hunnh ]
[ Parent ]
Chess (none / 0) (#21)
by J'raxis on Wed Oct 03, 2001 at 04:41:56 PM EST

Whoever was “winning” in my analogy would be the side (either the company selling the Kazaa client or the OSS developers making their own client) getting a working client off the ground and getting users to use it.

To finish clarifying, “figuring out the game” meant reverse-engineering the protocol, “changing the rules” meant rewriting the protocol, “losing” should be obvious, and “checkmate” meant when the OSS developers produced a client as attractive as, or more than, the Kazaa client itself.

HTH.

— The clarifying Raxis

[ J’raxis·Com | Liberty in your lifetime ]
[ Parent ]

Very good, but one change ... (5.00 / 5) (#10)
by Dlugar on Tue Oct 02, 2001 at 07:26:05 PM EST

Child One owns the chess set and all the pieces, and refuses to allow Child Two to play with them. "You can play, certainly," Child One says, "if you supply your own pieces." Child Two works hard at the lathe constructing her own pieces, and comes back to play.

They play for a while, at which point Child One pulls out a new set of pieces and says, "Well, ok, now if you want to play you have to have pieces that look like this," and Child Two goes back to the lathe.

My big problem with this isn't that Child Two doesn't want Child One playing on his board--that's his prerogative. What I find distasteful is that, instead of simply saying, "Look, we don't want you here, go somewhere else," he instead wants to play his sadistic little games, forcing Child Two "back to the lathe" for ridiculous reasons.

If AOL had told the AIM clones, "Please do not connect to our network," that would have been fine. Not very nice, but fine. Then Jabber or whatever other open protocol happened to be out there would get a considerable boost and the resources of those developers who might have otherwise been working on Gaim and so forth. Similarly, if FastTrack would have told giFT, "We don't want you on our network, get off or we'll do some nasty encryption stuff," that would have been fine too. Not very nice, but fine. Then those developers would be spending their time on something more productive.

However, what all these companies seem to be doing is giving these developers false hopes that compatibility will be ensured, while breaking the protocols behind their backs--very likely a ploy to attempt to keep a large market share (by sucking in all the developers with their schemes), while breaking, every once in a while, anything that doesn't make them money.

Please, just tell us you don't want your proprietary networks and/or protocols! We'll go home and stop complaining. But when you dangle the carrot just beneath our nose to further your own ends, I consider that parasitic, not our attempts to provide you with clients on other platforms.

Dlugar

[ Parent ]
Asking politely (5.00 / 5) (#13)
by babylago on Tue Oct 02, 2001 at 07:55:31 PM EST

Two points:

First, no matter how the example is phrased, there are two underlying important assumptions. One is that Child One owns the chessboard and that the rules that Child One establishes are the rules of the game. Two is that Child Two can walk away at any time, and is not required to play. Unfortunately, many parasitic Open Source projects (and I don't mean this in a negative sense, just that they piggyback onto proprietary foundations) confuse the ability to do something with the mandate to actually do it. That, I think, is where your argument in your post is coming from, that Child One is somehow "wrong" or "unfair" for maliciously changing the rules, when it is fully within her rights as owner to do just that, whether or not it is convenient for an independent Open Source effort or not.

Second, telling me how much you're doing for me by developing Open Source projects when a perfectly good commercial product is reasonably available in no way supports the argument that Child One is doing something "wrong." It does demonstrate that you have time and the willingness to do something, but it does not demonstrate that I am any way obligated to you for doing it.

---
[ Blog | Hunnh ]
[ Parent ]

You miss my point. (4.66 / 3) (#16)
by Dlugar on Tue Oct 02, 2001 at 08:16:17 PM EST

That, I think, is where your argument in your post is coming from, that Child One is somehow "wrong" or "unfair" for maliciously changing the rules, when it is fully within her rights as owner to do just that, whether or not it is convenient for an independent Open Source effort or not.
I fully understand that. My point is only that if owners of proprietary protocols don't want open source projects "piggybacking", then when these projects come to them for help, they should say, "No, we don't want you using our network," instead of throwing a few bones their way. It is wrong and unfair and malicious, but it is still within their rights as owners of those protocols [and, in some cases, networks].

it does not demonstrate that I am any way obligated to you for doing it.
No, but if you act as if you're pleased with what I'm doing, and then go and lock me out without warning, I do consider that sleazy and underhanded, regardless of the legality of it.

Dlugar

[ Parent ]
Perfectly good? Not likely. (4.40 / 5) (#17)
by Ialdabaoth on Tue Oct 02, 2001 at 08:32:57 PM EST

Babylago, if the commercial product only runs on one platform, the I couldn't call it perfectly good. If I have to run a certain defective operating system to use the commercial product, then I have two choices: do without or write a work-alike.

This is not to say that the commercial developers are obligated to help me write my work-alike -- especially if doing so eats away at their sales. However, if you are going to oppose me, have enough pride to do it openly and say, "If you want to write your own tool, go ahead. But don't come to us for help."

Child One should get his ass off the fence: either aid or oppose. But don't pretend to aid while actually opposing. Nobody has time for that sort of passive-aggressive bullshit.
*******
"Act upon thy thoughts shall be the whole of the Law."

--paraphrase of Aleister Crowley
[ Parent ]

Belongs... (none / 0) (#24)
by karjala on Sat Oct 06, 2001 at 04:25:14 PM EST

The network shouldn't "belong" (and therefore doesn't, according to me) to the creators of the programs. It's like saying the WWW network belongs to the CERN institute or that the internet belongs to some US agencies.

The community... Does the community of AIM users belong to AOL? After a big period time, should they still be able to control what software we eat, how much security we can have, and so on?

[ Parent ]
Well (3.00 / 3) (#7)
by spacejack on Tue Oct 02, 2001 at 07:05:29 PM EST

A truly disappointing move on FastTrack's part to centralize the servers, and perhaps one that will lead to the network's ultimate demise.

If you mean because they'll miss out on all the hundreds of open-source Linux clients... probably not. If you mean because of all the open source Windows clients that won't show their ads... they probably don't appreciate those. If you mean because the authorities will be able to track copyright infringers down, well.. they'll have to deal with the RIAA/MPAA at some point once they get big enough anyways.

The last. (5.00 / 3) (#11)
by Dlugar on Tue Oct 02, 2001 at 07:30:59 PM EST

If you mean because the authorities will be able to track copyright infringers down, well.. they'll have to deal with the RIAA/MPAA at some point once they get big enough anyways.
Yes, that's what I was referring to. That's why I included the bit, "to centralize the servers"--I have no delusions that they would be ecstatic about a crappy pre-Alpha open source clone of their software.

If the servers were sufficiently distributed Gnutella-style (as they used to be), and the protocol even somewhat open (as it used to be), then it would likely survive whatever nasties the RIAA or MPAA might have for them. However, now there is a central server to shut down, which is quite easy for the Media Giants to do if they feel so inclined. And bye bye to the FastTrack network.

Dlugar

[ Parent ]
In other news.. (1.75 / 4) (#9)
by DeadBaby on Tue Oct 02, 2001 at 07:20:41 PM EST

Countless crappy open source clone programs are often only a fraction of the quality of beta releases of the Windows software they clone. Sometimes not even working at all.
"Our planet is a lonely speck in the great enveloping cosmic dark. In our obscurity -- in all this vastness -- there is no hint that help will come from elsewhere to save us from ourselves. It is up to us." - Carl Sagan
That's not the problem. (4.00 / 1) (#12)
by Dlugar on Tue Oct 02, 2001 at 07:33:23 PM EST

The problem lies in that to get rid of this crappy pre-Alpha Linux software, they moved from a decentralized Gnutella-server style network to a single-point-of-failure network which can be quite easily shut down--and likely will be.

Dlugar

[ Parent ]
Phex (none / 0) (#18)
by kiwipeso on Tue Oct 02, 2001 at 11:41:01 PM EST

Just about all my gnutella programs crash on startup, except for MacPhex.
Phex is a very stable Java app that gets into gnutella networks very quickly and has intelligent downloads built in.
It uses candidate file matching, if your current candidate goes offline, you can just get it to search for another or it will switch to another if it knows of others.

Kaos operating system creator.
Now go look at the W3C's proposal... (5.00 / 2) (#19)
by jason on Wed Oct 03, 2001 at 12:13:07 AM EST

This is what happens with proprietary communication protocols. Consider using this as a decent example in the debate over the W3C proposal to allow proprietary standards.

Now if some free software standard weds a scalable network with reliable, scalable multicast downloads, we might have a winner. Note that Digital Fountain is working on open standards for multicast. I'm not sure what their stance is on Tornado codes for multi-source downloads. Some aspects seem patented, but those aspects aren't as necessary in a non-streaming setting. I think you could also use this to implement a Freenet-like system and defeat some basic traffic analysis. When a user requests a file, other nodes that saw the query could download the file simultaneously. hm. A Crowds-like query system could defeat analysis of the requests, assuming there are enough requests.

With that and ad-hoc network pools set up through scoped multicasting (sigh, ipv6-only), you could replace the FastTrak (sp?) protocol with IETF standards...

FastTrack, Kazaa, MusicCity sued! (5.00 / 4) (#20)
by unixrulz on Wed Oct 03, 2001 at 03:49:15 AM EST

Breaking News: The RIAA/MPAA have sued the companies involved in the FastTrack network. Only 4 days after moving their clients to authenticate with them. Coincidence?
Here's the URL:
http://news.cnet.com/news/0-1005-200-7389552.html?tag=mn_hd

Fork the network (5.00 / 2) (#22)
by hardburn on Thu Oct 04, 2001 at 10:38:39 AM EST

So you have a perfectly good libre clone of the FastTrack network, except for that fact that it now cannot connect into the network the propriety areas of the network. Furthermore, the apparent reason for this is due to centralization.

I suggest the creaters of giFT don't worry about what the rest of FastTrack is doing and go off in a diffrent direction. Create a Windows version of giFT to get more market share if you must, but don't worry about Kazza and the rest of the lame Napster-wannabees. If done right, the giFT developers can make extentions to the existing, open network protocol and make it better then the propreitary version.

Of course, when Freenet becomes more stable, all this will be moot.


----
while($story = K5::Story->new()) { $story->vote(-1) if($story->section() == $POLITICS); }


Yes, but... (none / 0) (#28)
by dash2 on Tue Oct 09, 2001 at 07:25:49 AM EST

... it's about content, not technology. File sharing suffers from network effects: protocols get more useful if they have a large number of subscribers online. If you start your own FreeFastTrack, you'll need to get enough people on board to make it worthwhile. Of course, you could just go and hack Gnutella or Mojonation or Freenet... Gnutella might be interested in some of these ideas.
------------------------
If I speak with the tongues of men and of angels, but have not love, I am become sounding brass, or a clanging cymbal.
[ Parent ]
heh.. (none / 0) (#23)
by rebelcool on Sat Oct 06, 2001 at 03:39:23 PM EST

back in the day when i had more time on my hands, I wrote a quick decentralized p2p system that used metadata. I created a nice little search language that allowed quite a bit of customization in the search (basically a mini-sql)... I really wish I could get back to working on that :(

COG. Build your own community. Free, easy, powerful. Demo site

Then go do it (none / 0) (#29)
by hardburn on Wed Oct 10, 2001 at 09:22:58 AM EST

Then go ahead and do it. There are lots and lots of P2P projects trying to be "anoynomous" while still being efficent. I'd like to see some P2P projects that say "screw anonynomous" and just focus on being efficent. While I'm all for privacy and free speech and such (which is why I work on Freenet), there are some places where it's just not needed. Freenet, for instance, is already very efficent, but it won't give up anonynimity for efficency. A Freenet-like network without the anonynimity requirement would be down right amazing.


----
while($story = K5::Story->new()) { $story->vote(-1) if($story->section() == $POLITICS); }


[ Parent ]
Well. (none / 0) (#25)
by mindstrm on Sun Oct 07, 2001 at 12:31:17 PM EST

Is there any actual statement from the FastTrack people as to what happened? Or is everyone just guessing.



Questions. (none / 0) (#26)
by mindstrm on Sun Oct 07, 2001 at 12:37:25 PM EST

First.. you say 'Even though the makers of FastTrack network seemed interested in an open-source Linux version at first', they provided no support to the giFT project... well, then how exactly did they 'seem interested' in an open source version? If they didn't help from square one, didn't supply specs, and giFT used reverse-engineered specs.. how could they reasonably expect this NOT to happen?


See giFT's press release (none / 0) (#27)
by Dlugar on Sun Oct 07, 2001 at 11:55:59 PM EST

While Kazaa originally stated that they were interested in working with the open source community we quickly discovered that it was not the case. They asked us to develop a closed source Linux client, however we decided to go the open source route and start on our own.

I assume they simply figured that the FastTrack people simply couldn't be bothered by a project so small and insignificant, but wouldn't be annoyed at them for connecting to the network.

Dlugar

[ Parent ]
Another Filesharing App Borked | 29 comments (26 topical, 3 editorial, 0 hidden)
Display: Sort:

kuro5hin.org

[XML]
All trademarks and copyrights on this page are owned by their respective companies. The Rest 2000 - Present Kuro5hin.org Inc.
See our legalese page for copyright policies. Please also read our Privacy Policy.
Kuro5hin.org is powered by Free Software, including Apache, Perl, and Linux, The Scoop Engine that runs this site is freely available, under the terms of the GPL.
Need some help? Email help@kuro5hin.org.
My heart's the long stairs.

Powered by Scoop create account | help/FAQ | mission | links | search | IRC | YOU choose the stories!