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]
Freenet, an anonymous distributed network

By ramses0 in News
Thu Mar 02, 2000 at 02:50:35 PM EST
Tags: etc (all tags)
/etc

http://freenet.sourceforge.net/

I stumbled across this link from a slashdot comment. The gist is that participants in this "free network" agree to host bundles 'data' attached to 'keys'. If one server doesn't have it, it can go ask another. If a particular freenet server recieves a lot of requests for a certain 'key', it can cache the 'data' locally, theoretically saving bandwidth.


ADVERTISEMENT
Sponsor: rusty
This space intentionally left blank
...because it's waiting for your ad. So why are you still reading this? Come on, get going. Read the story, and then get an ad. Alright stop it. I'm not going to say anything else. Now you're just being silly. STOP LOOKING AT ME! I'm done!
comments (24)
active | buy ad
ADVERTISEMENT

Since their site is a little tricky to navigate: features of freenet, how it works, and some philosophy about why something like this is necessary.

Personally, I'm impressed by this little slice of anarchy. First of all, it's a very impressive technological feat (if it works), sort of like the GPL for data.

Of course this relies on lots of users donating lots of space for something like this to work. With too many 'leeches' who just want to store data, the overall utility of this will be useless. But if you've got something important to say, and don't want it to be repressed, censored, or deleted, this seems like a good place to put it.

Any theorizing about what the future of a system like this would hold? Since it's all anonymous users and anonymous data, what happens if a person posts Microsoft Office? What if someone posts the DeCSS algorithm for DVD playback?

Like they mention, Freedom is a dangerous thing. Will the 'net ever be ready for something like this? And how is this different from newsgroups?

Sponsors

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

Login

Related Links
o Slashdot
o http://fre enet.sourceforge.net/
o features of freenet
o how it works
o philosophy
o Also by ramses0


Display: Sort:
Freenet, an anonymous distributed network | 19 comments (19 topical, editorial, 0 hidden)
A somewhat similar-sounding project... (2.00 / 1) (#1)
by rusty on Thu Mar 02, 2000 at 01:41:38 PM EST

rusty voted 1 on this story.

A somewhat similar-sounding project I was involved with a while ago is Ompages. Sounds like Freenet has more of a technological background than mpages. But both cool ideas.

____
Not the real rusty

I think this sounds really cool. An... (2.00 / 1) (#3)
by stimuli on Thu Mar 02, 2000 at 01:45:13 PM EST

stimuli voted 1 on this story.

I think this sounds really cool. And if its popular enough, new servers could pop up faster than the suits could get court orders to shut them down. That would be a good thing.
-- Jeffrey Straszheim

I'd like to see what would happen i... (2.00 / 1) (#2)
by asad on Thu Mar 02, 2000 at 01:46:13 PM EST

asad voted 0 on this story.

I'd like to see what would happen if they distributed a confidential document. I like the idea but the potential for abuse is very high.

Re: I'd like to see what would happen i... (5.00 / 2) (#7)
by ramses0 on Thu Mar 02, 2000 at 05:02:10 PM EST

I thought about this myself... and thought to myself- hey... if I were running one of these servers, I wouldn't want any child pornography on my ~node~.

My immediate next thought was "if I want to store something which somebody else finds offensive, then they should just lighten up and not worry about it"...

Which I then realized was hypocritical. I think that if you end up running one of these servers, you -really- have to believe that what people do and what data they are storing is their own damned business. It's like: if you have something important enough to want to put on a freenet, then you have to realize that you are supporting -everyone else- who has stuff they think is important enough or ~illegal~ enough to warrant putting it on a freenet.

It's a very interesting and thought-provoking topic.

--Robert
[ rate all comments , for great justice | sell.com ]
[ Parent ]
Interesting moral situation (3.00 / 1) (#15)
by joeyo on Thu Mar 02, 2000 at 11:37:13 PM EST

[argh, I wrote a big long reply to this comment and I must have only previewed it and not posted it- I'll try again with a thumbnail sketch...]

Much the same as you, practically my first thought when I was reading about how this thing works was "I don't want someone to store their kiddie pr0n on *my* box". And you know, there are some interesting issues here:

Let's suppose the freeweb system decides to put something horribly illegal and immoral on my node. Am I morally responsible? I'm thinking that I could live with freeweb if 1) I had no idea what I was hosting on my node 2) No one else has any idea what I'm hosting on my node :) But is this a cop-out?

Besides, I find it hard to believe that freenet could be robustly anonymous for very long. At the very least someone could hack their server (open source right?) so that they know what is on their own box and they could probably get some interesting data on what their neighboring boxes have/request.

Well, I guess you could at least post and read files off freeweb without hosting stuff...

/joeyo

--
"Give me enough variables to work with, and I can probably do away with the notion of human free will." -- demi
[ Parent ]

Re: Interesting moral situation (4.00 / 1) (#17)
by ramses0 on Fri Mar 03, 2000 at 12:55:48 AM EST

This doesn't seem like a problem they could even use Large(tm) cryptography to avoid since the whole point is to have as large a distribution as possible.

Assuming it is only possible to attack your own node, and that you have a list of data which is stored on your own node. The system might be able to support data/key signing, meaning that tampered files are immediately noticed when they are seen by a different server. Then, if a user requests a file from a freenet, and finds that it has been tampered with (ie, they notice that an internal MS memo has been replaced with a promotional brochure) they could "search again", but disallow that document signature from returning in results.

This would effectively limit the distribution of tainted documents, and a good way for the network to protect itself is. Nodes which somehow determine they have received tainted data could lower their "trust level" of either the node which originated the data packet, or the neighboring node which it received the data from, and thusly not request data from those nodes as frequently.

This way, a node can only affect it's -own- routing/network path, untrustworthy (ie, hacked) nodes don't receive requests for data, and if a hacked node decides to forge a lot of "this file's tampered" packets, they are only further ostracizing themselves from the network.

Whew! Complicated stuff! But a -very- interesting research project.

--Robert
[ rate all comments , for great justice | sell.com ]
[ Parent ]
A very powerful idea... (3.00 / 2) (#4)
by joeyo on Thu Mar 02, 2000 at 04:02:40 PM EST

It seems that no one would really knows where the data is located, even the owners of a particular node would not know what they are hosting. This, in-and-of-itself is kinda scarey. But even better, the more that a file is requested, the more that it is passed around. So trying to track down a file actually spreads it around.

As I understand it however, it must pass data one node at a time (ie if a file is located 10 nodes downstream from you, that 10th node will not send it to you directly but rather will send it upstream node by node. This would be good for privacy and provide lots of redundancy but it would get dar slow. I'm also left wondering how the servers know about eachother. Obviously, there can not be a master server list.

--
"Give me enough variables to work with, and I can probably do away with the notion of human free will." -- demi

Re: A very powerful idea... (2.50 / 2) (#5)
by stimuli on Thu Mar 02, 2000 at 04:19:02 PM EST

True. If you ask for something far away, it will take some time for you to get it. On the other hand, the system does appear to dynamically structure itself to match points of high demand. Obviously a master server list would be a real bad idea, but the density of the server graph could be tweaked quite a bit as such a system evolves.

At first glance the system seems powerful. I'll be interested in seeing what steps the suits take to try to stop it.
-- Jeffrey Straszheim
[ Parent ]

Freenet, Ompages -- and Eternity (4.00 / 1) (#6)
by Demona on Thu Mar 02, 2000 at 04:49:16 PM EST

Just to remind any Cypherpunks of Adam Shostack's Eternity Server. (Note: That pages' links to replay.com are dead since ReplayTV took over that domain.)

I'd like to see a comparison of Freenet, Ompages and Eternity, along with some suggestions on how to, as Ken Kesey says, "Take what you can use, and let the rest go by." Eternity appears to be a bit stagnant compared to the others, but based on author reputation alone I'd tend to trust its cryptography. So what are the best features of each, from both a design and end-user standpoint, and how can they best be combined? Or are the designs different enough that, like Lynx and w3m, it's just best to keep them separate and look for convergence a bit further down the road?

Re: Freenet, Ompages -- and Eternity (4.00 / 1) (#13)
by Anonymous Hero on Thu Mar 02, 2000 at 08:20:58 PM EST

Eternity is what I call "author centric" - it is designed in the interests of information authors, particularly wealthy information authors. Freenet is "reader centric", it is designed to deliver the best quality content to the most people as reliably as possible. That is the fundamental difference. Author-centric is a bit like a weblog with no ratings (author centric, if you want to be heard you need to be first/influential), reader centric is a bit like a weblog with ratings (if you want to be heard, you need to be good!).

[ Parent ]
Re: Freenet, Ompages -- and Eternity (none / 0) (#14)
by Anonymous Hero on Thu Mar 02, 2000 at 10:12:50 PM EST

Haven't looked at Freenet and Eternity, but Ompages is currently undergoing a change of focus; it appears that it will be turning into a commercial project. What exactly will be the services that end up being offered, I don't know.

[ Parent ]
Re: Freenet, an anonymous distributed network (2.67 / 3) (#8)
by rongen on Thu Mar 02, 2000 at 06:08:11 PM EST

Wow, stuff like this really floats my boat. This is the kind of thing
that will keep the net from turning into cable-TV in a few years. One
of the greatest things about the Internet is the way it empowers
people to get their thoughts out to others. Most people are passionate
about one thing or another that doesn't really interest their friends
or colleagues, the Internet (USENET, WWW, etc) can be a great way to
explore these interests (okay, /some/ people like bomb making... not
I).

This Freenet gives us a chance to redefine the way ideas get passed
around, and in a good way. People say Usenet is dying out. That bugs
me. I like the idea of anonymously re-mailing something to a newsgroup
and being able to watch that idea discussed, flamed, agreed to (I
don't post anonymously but I like knowing how, and knowing I can, if I
want to). It's tough to do that on the Web. If you want to talk about
client/server you need some kind of way to tie the browser to an IP,
etc (yes, I know about anonomizing proxy servers, they don't assuage
my paranoia enough!). But with Usenet you just send your message
off... great. Especially if you want to discuss something
"controversial". I bet "Coolio" is wishing he had been using a system
like this instead of IRC to reveal (or take unearned credit for) the
DoS attacks...

Meanwhile I have a question... One thing I didn't get from the Freenet
pages is this: Having unleashed my anonymous document onto Freenet,
how do I edit it? Would they use keys for this, and hopefully a
timestamp (for versioning?)... Because I am thinking it wouldn't be
too hard to archive this stuff and use "editing" landmarks as a way to
trace a document back to a set of servers. Touch a document three
times and you've narrowed the search down considerably. Does this make
sense?

PS> If anyone is wondering, I am not a privacy freak and don't
practice anonymous posting, or even read USENET all that frequently!
:) or do I?


read/write http://www.prosebush.com
Re: Freenet, an anonymous distributed network (4.00 / 1) (#11)
by Anonymous Hero on Thu Mar 02, 2000 at 07:53:48 PM EST

Hi, I am the guy who came up with Freenet. Regarding your question about editing data, we have several proposals about how to do this, although it won't be added before our second release. Basically you will have the option, when you create some data, to create a public/private key and attach the public key to the data. If you want to edit the data you just sign the instructions using your private key and send those instructions to the various servers hosting your data.

[ Parent ]
Re: Freenet, an anonymous distributed network (3.00 / 1) (#16)
by rongen on Thu Mar 02, 2000 at 11:57:17 PM EST

So you could maybe implement the standard UNIX file perms (well: read, read-write anyway) using public key cryptography? Slick... This idea is sounding better all the time. Actually having data as officially unchangeable by anyone but it's author would give the system a great deal more integrity than a mirrored set of documents on the web has. Also, if people were unable to alter information moving across their node they are, in a way, less responsible for it (no editorial power and thus less chance of being sued/prosecuted for the data).

Thanks for responding to my question. Best of luck with this, I'll be watching this project closely! (in a good way) :)
read/write http://www.prosebush.com
[ Parent ]
Re: Freenet, an anonymous distributed network (2.50 / 2) (#9)
by Matthew Guenther on Thu Mar 02, 2000 at 06:08:29 PM EST

I like this idea, it reminds me of the collective described in the Tad Williams book "Otherland". However, I didn't see anything on their page about error-checking... what's to prevent someone from modifying the data located on a server under their control? This could be a highly effective method of sabotaging the network... join Freenet, get a copy of the material you want to sabotage on your machine, modify it, and then go public saying you are trying to censor it. The high demand causes your modified copy to get spread, and the corrupt version will increase demand further as people try to the key repeatedly in hopes of getting a good copy.

Looking through the documents, Freenet doesn't seem to handle this. Am I confused, or does anybody know how they are planning on handling this?

MBG



Re: Freenet, an anonymous distributed network (3.00 / 1) (#10)
by stimuli on Thu Mar 02, 2000 at 06:39:48 PM EST

We could hope that the spirit of the thing would prevent abuses of this source. Also, it seems likely that one bad apple on a large network wouldn't be able to ruin it, because the document he messes up at least lives on enough other servers to keep the real one circulating. Someone needs to run some simulations to test this.
-- Jeffrey Straszheim
[ Parent ]

Re: Freenet, an anonymous distributed network (4.00 / 1) (#12)
by Anonymous Hero on Thu Mar 02, 2000 at 08:09:14 PM EST

Modifying a piece of data stored locally on your server will be a very limited way to annoy Freenet. Firstly, it won't be up to you what data is stored on your server, so if you have a specific agenda it will only be through extreme luck that you can progress it through modifying data on Freenet. Further, unless the data is really unpopular (and therefore really rare) any modifications you do make to a local copy will be a drop in the ocean. Thirdly, we are working on the idea of encrypting data (actually the idea has been fully thought out, but we won't implement it for the first release) which will make it impossible to modify, or even determine what it is, without turning it into gibberish.

[ Parent ]
Re: Freenet, an anonymous distributed network (none / 0) (#18)
by Matthew Guenther on Fri Mar 03, 2000 at 08:19:08 AM EST

Modifying a piece of data stored locally on your server will be a very limited way to annoy Freenet. Firstly, it won't be up to you what data is stored on your server, so if you have a specific agenda it will only be through extreme luck that you can progress it through modifying data on Freenet.

Couldn't I just request the data a whole lot from a server located on a machine under my control? As I understand it the data will migrate towards the greatest desire for it, would this eventually cause it to show up on the machine I'm targetting?

Further, unless the data is really unpopular (and therefore really rare) any modifications you do make to a local copy will be a drop in the ocean.

How much redundancy are you planning on building in? Too much and nobody will have enough storage space. Also, it seems to me that although the initial modification would, as you say, be a "drop in the ocean", it has the potential to grow exponentially as each retrieval of the flawed copy will result in another request for the same key in the hopes of getting a good copy. This demand would cause the spread of data associated with that key, be it the original or modified copy. Voila, you decrease people's ability to get the original data by:

  1. obscuring the actual data
  2. creating a large load on the freenet system.

Encryption sounds like the way to do things (digitally sign each copy), but good crypto is always trickier to implement than it appears, and it would probably sacrifice some of the anonymity.

Not trying to be a downer, I sure hope this succeeds, but this is something one would probably want to get this right on the first try.

MBG



[ Parent ]
Something I don't understand about freenet (4.00 / 1) (#19)
by Anonymous Hero on Sat Mar 04, 2000 at 05:13:04 AM EST

Mmm. Maybe it is too late to get a response about this.

There is something I don't understand in freenet design:

How can you get sure that a document don't get removed from the system because it is not asked often enough ?

(I think that you can't be sure a document don't disappear, but I may be wrong).

If you cannot get sure, then what prevent an organisation to delete meaningfull documents by "spamming" freenet (ie: adding documents and then using bots/monkeys to retreive it from a great number of nodes. Freenet will only keep information important for its users, so by perveting the user base, you can pervert the content)

Cheers,

--fred

Freenet, an anonymous distributed network | 19 comments (19 topical, 0 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!