chris783
join:2004-04-01
1 edit | new kinds of anonymous p2p software ...
I am aware that this is not for the mainstream, but i saw p2p software lists in this forum that are a bit outdated.
If you are willing to try new beta software these 2 candidates might be worth a shot.
GNUnet is my personal favorite but it comes only for linux/unix until now and mute has a windows release available but more technical problems.
You have been warned 
home: »www.ovmj.org/GNUnet/
forum: »gnunet.cryss.net/forum/index.php
home: »mute-net.sourceforge.net/
(One is sure. These are 100% [add/spy/whatever]-ware free) |
|
jap
Premium
join:2003-08-10
038xx
 ·RoadRunner Cable
| How can they be anon unless they go through some "washing" node - a machine that would be first positively overwhelmed with traffic loads (huge $) and then be targetted by enforcement efforts & quickly shutdown? There's no way to ship the data if there's not an IP embedded in each packet. |
|
salahx
join:2001-12-03
Saint Louis, MO
| Don't forget Freenet and Frost. also Check out the Invisible IRC Project and The Invisible Internet (I2P) Project.
said by jap  :
How can they be anon unless they go through some "washing" node - a machine that would be first positively overwhelmed with traffic loads (huge $) and then be targetted by enforcement efforts & quickly shutdown? There's no way to ship the data if there's not an IP embedded in each packet.
The general idea behind most anonymity features is each "node" on the network doesn't know every other node, just a few "neighbors". The node hands the request off to the "neighbor" who it thinks has the data. If the neighbor has the data, it returns it (possibly after a random delay). If not, the neighbor hands off the request to one of ITS neighbors who it think have the data. If the neighbor of the neighbor has the data, it returns it (possibly after a random delay). The original neighbor the hands it off to the original node. If not, the process continues (until some limiting factor is reached - a hop count, timeout, etc). Of course, all the data is encrypted, so node has any idea what is passing or storing (and renders sniffers useless). Essentially, each node looks like its proxying the data for another node, so no one can be sure where the origin or the destination is. (Look up "garlic routing" for examples of this.)
You can get an explanation of how Freenet works at The Freenet Help Site.
These projects are all understaffed and underfunded, so they are slow going. |
|
sorne guy
@milwwi.ameritech
| reply to chris783
anonymity vs. efficiency
do you want to hide, or do you want your files in a timely manner?
anonymity brings hurdles with it--the passing of data through other computers (which may lack processing power or bandwidth)can slow down your transfers, and there is really nothing you can do about it
freenet brings up the question of whether or not you can live with the idea that the files you cache and transfer might be kiddy porn, snuff, or other abhorrent things you may disagree with--freenet is based on absolute free speech, so whether you agree with the bad stuff or not, you may be distributing it anyway
do you want to be a mule for the local pedophile? |
|
netscape 6
join:2002-03-07
Constantine, MI
| said by sorne guy:
do you want to be a mule for the local pedophile?
Do you want to help a blood thirsty torching dictator censor the critics in his country? Don't run freenet if you do. Do you want to help the church of $cientology censor it's critics through lawsuits and harassment? Don't run freenet if you do.
That argument works both ways. |
|
chris783
join:2004-04-01
| reply to sorne guy
quote:
do you want to hide, or do you want your files in a timely manner?
I think that is answered by the title of this thread. If you want to exchange data using the internet like buying apples in a supermarket you should not share anything critical to avoid attacks on you. If not you need to "buy" that security with something.
quote:
the passing of data through other computers (which may lack processing power or bandwidth) can slow down your transfers, and there is really nothing you can do about it
Thats outdated or never was true. You just need to break down data into blocks (and now dont try to tell me that costs something else e.g. bandwidth) that are suitable for small systems/slow connections. GNUnet is definitely not like common filesharing software.
If you really want to understand why these are facts, then read more here
»www.ovmj.org/GNUnet/faq.php3?xla···#compare
and here
»www.ovmj.org/GNUnet/download/esed.ps
Chris |
|
sorne guy
@milwwi.ameritech
| reply to netscape 6
quote:
Do you want to help a blood thirsty torching dictator censor the critics in his country? Don't run freenet if you do. Do you want to help the church of $cientology censor it's critics through lawsuits and harassment? Don't run freenet if you do.
that argument doesn't work at all
scientology crap is all over newsgroups and the internet (but who cares anyway?) what will necessitate the need for anonymous transfers? it isn't going to be grandmas cookie recipes, it's going to be the most illegal, vile crap being transferred between the sickest, most perverted freaks on earth, hiding themselves behind a veil of anonymity
people that want to share something that they think is important will do it in public, and risk getting sued, if they really believe in it. the "censorship" arguement in the case of freenet is a huge cop out--it makes a system with largely illegitimate uses look like it's crusading for human rights
some types of "speech" are illegal for good reason--child porn is exploiting somebody's kids, there is no good reason for racist speech/hate speech, or bomb-making plans, but i guarantee those 3 things make a good percentage of what's transferred on freenet--and if you use it, you are an accessory to the crime because you are enabling these freaks |
|
P2P Dude
@sympatico.ca
| quote:
what will necessitate the need for anonymous transfers? it isn't going to be grandmas cookie recipes, it's going to be the most illegal, vile crap being transferred between the sickest, most perverted freaks on earth, hiding themselves behind a veil of anonymity
You couldn't be more wrong. Greedy corporations that are suing file sharers are the reason why anonymous P2P applications have become a necessity. Many P2P users have already been extorted by the RIAA and since common sense and justice no longer mean anything in justice systems, we need to do something. You can already find perverted things on the WWW or P2P systems. You don't need anonymous P2P to trade that.
quote:
people that want to share something that they think is important will do it in public, and risk getting sued, if they really believe in it. the "censorship" arguement in the case of freenet is a huge cop out--it makes a system with largely illegitimate uses look like it's crusading for human rights
Really? Are you sure? Do regular people really would rather risk paying the RIAA or MPAA $5000-$10000 for something that shouldn't be against the law? Or do you think they would rather trade file anonymously?
The censorship argument is hardly a cop out. Do you even know why it was created in the first place? I doubt it.
quote:
some types of "speech" are illegal for good reason--child porn is exploiting somebody's kids, there is no good reason for racist speech/hate speech, or bomb-making plans, but i guarantee those 3 things make a good percentage of what's transferred on freenet--and if you use it, you are an accessory to the crime because you are enabling these freaks
I use Freenet all the time and I don't partake in any of the 3 things you mentioned. Claiming that Freenet users are enabling these freaks to do anything is moronic. It's like saying ISP's are enabling their customers to download pornographic images off the WWW. Get a clue. |
|
sorne guy
@milwwi.ameritech
| quote:
I use Freenet all the time and I don't partake in any of the 3 things you mentioned.
the crap you are cacheing for other users is encrypted, so how do you know it isn't one of those 3 things?
if an ISP is hosting illegal stuff, they have to take it down--they are responsible for what is stored on their servers
how are you any different? |
|
P2P Dude
@sympatico.ca
| quote:
the crap you are cacheing for other users is encrypted, so how do you know it isn't one of those 3 things?
if an ISP is hosting illegal stuff, they have to take it down--they are responsible for what is stored on their servers
how are you any different?
I don't know and neither do you. So neither of us can claim that any particular Freenet user is, or isn't hosting those 3 things. All I know is that I have no interest in those things and use Freenet for other things.
Freenet users have no relationship with other Freenet users. ISP's have customers that pay them to have internet access.
Do a search for any of those 3 things you mentioned on google and you will probably find a bunch of relavent links. Pretty much everything you find on Freenet can be found on the WWW.
Why should I have to stop using Freenet just because there are some sickos who use it to download disgusting things? If I felt I should stop using Freenet, why shouldn't I feel like getting offline completely because of the disgusting stuff on the WWW? |
|
jap
Premium
join:2003-08-10
038xx
·RoadRunner Cable
| reply to salahx
salahx -- Got it. Thx for the explanation. I trust, then, that at each hand-off the previous node's IP is over-writ so there is no persistent record of who handled it more than one node back .... right? That works. Also means that requesting machine must complete the file and re-request parts that fail to arrive. How well does it work in practice? And, in such a model, how (and how well) are searches handled? |
|
chris783
join:2004-04-01
| reply to sorne guy
It is/will be a fact that criminals will make use of anonymous p2p software. But if this software doesnt exist, it does not mean that there are no other ways for them. They can allready use all kinds of freely available steganographic and cryptic methods to hide their content even more secure than p2p networks can be.
They also might not be interested in free distribution (even at that small risk of an anonymous p2p network) anyway.
Finally most of them need to make money .. and thats another point a p2p system can not fulfill ... fortunately |
|
salahx
join:2001-12-03
Saint Louis, MO
| reply to jap
said by jap :
salahx -- Got it. Thx for the explanation. I trust, then, that at each hand-off the previous node's IP is over-writ so there is no persistent record of who handled it more than one node back .... right? That works. Also means that requesting machine must complete the file and re-request parts that fail to arrive. How well does it work in practice? And, in such a model, how (and how well) are searches handled?
Each node only gets the IP address of who handed off the request - if A originated the request, A hands it to B (which it thinks is the "destination"), B hands it to C - but C only knows about B, no information about "A" (except maybe the virtual node ID, which can't be mapped to an IP) is in there. So as far as "C" is concerned, "B" originated the request (which it didn't). Likewise, as far as "B" is concerned, "A" originated the request (which it did). When the data is found (or isn't found, or timeouts), it goes back through the chain, each node sending it back to what it thinks the "origin". So when "B" gets (or doesn't get) the data from "C", it thinks "C" is the destination, and when "A"gets it back from "B", "A" thinks "B" is the destination. As you can see, there no way to tell who's the "real" origin or destination. (much less the CONTENTS of what is begin relayed).
As for "searching", there are lots of different models, but they pretty much work through a "smart stumbling in the dark algorithm". Nodes start sending requests randomly. Based on which ones succeed and fail, and possibly how long each success and failure takes, the node gets an idea which path are ideal to find the data. On Freenet, its possible to "spider" freesites - typically, you'd send an anonymous messages (via NIM, or mixmaster remailers, or IIP, or via Frost), with your "key". The freesite would add that key to the spider, and harvest the links off your freesite, and add them to the spider, and then harvest the links to the freesites you linked to, and in turn.... (This is how people found data on the web before Google). There is a "FreeReactor" freesite on Freenet, which contains anonymous and non-anonymous links to various files and freesites. (As an added bonus, on Freenet, the act of spidering cause the data to propagate and be easier to find). The most active (and maintained one) is "Dolphin's Freenet Index". Other interfaces exist too - one could spider Frost the same way (even though it more resembles Usenet).
On Freenet, files are encoded using "Forward Error Correction (FEC)" - the same technology uses on "Interleaved" DSL lines. The file is split into X change, and some redundant chunks (say 50% X) are added too. So if a file is 16 "blocks" longs, you actually insert 24 "blocks" (16 + 8). You client only need to get 16 - ANY 16 of the 24, and it can rebuild the whole file (including the redundant chunks).
Other schemes can vary . As to how it works - well, your mileage will vary - a lot. You need to leave you "node" on for while (usually 1 or 2 days) to "learn" about the network. On Freenet, performance varies depending on (1) How "smart" you node is (hint: Don't be a "transient" node, unless your behind an uncontrollable NAT or don't have a 24/7 connection); (2) how popular the data is; (3) you nodes contribution to the network (i.e. bandwidth and HD space); (4) whether you are on the "stable" or "unstable" branch; (5) the version you are using (always use the latest - if you're on the "unstable" branch, upgrade every day at least; (6) the whims of the developers (new feature make Freenet work better but they are usually full of bugs - especially on Unstable - so things usually slow down until the bugs get worked out, then everything moves faster than originally); (7) Most importantly: Financial resources of the project. For a little as $5 month (less than 1/2 the cost of the CD!), you too can help feed a starving developer. |
|
chris783
join:2004-04-01
| First of all thanx salahx for the great explanation .. couldnt have done better
Exept
quote:
Each node only gets the IP address of who handed off the request - if A originated the request, A hands it to B (which it thinks is the "destination"), B hands it to C - but C only knows about B, no information about "A" (except maybe the virtual node ID, which can't be mapped to an IP) is in there.
For MUTE you are correct. There are virtual IDs which are very hard to map but thats one of the minor weaknesses of MUTEs protocoll. With GNUnet that is not existent cause routing is done by the entry/packet/request itself as an identification. Each node keeps track of the source-connection of routed packets so it knows where to send the answers when they come in. |
|
lakino
Premium
join:2003-04-03
Campbell, CA
| reply to sorne guy
said by sorne guy:
quote:
Do you want to help a blood thirsty torching dictator censor the critics in his country? Don't run freenet if you do. Do you want to help the church of $cientology censor it's critics through lawsuits and harassment? Don't run freenet if you do.
that argument doesn't work at all
scientology crap is all over newsgroups and the internet (but who cares anyway?) what will necessitate the need for anonymous transfers? it isn't going to be grandmas cookie recipes, it's going to be the most illegal, vile crap being transferred between the sickest, most perverted freaks on earth, hiding themselves behind a veil of anonymity
people that want to share something that they think is important will do it in public, and risk getting sued, if they really believe in it. the "censorship" arguement in the case of freenet is a huge cop out--it makes a system with largely illegitimate uses look like it's crusading for human rights
some types of "speech" are illegal for good reason--child porn is exploiting somebody's kids, there is no good reason for racist speech/hate speech, or bomb-making plans, but i guarantee those 3 things make a good percentage of what's transferred on freenet--and if you use it, you are an accessory to the crime because you are enabling these freaks
LOL. This post wins for the IRONY of the CENTURY award. This coming from a self proclaimed anon poster, posting under Sore Dud for at least a year. Why hide behind this anonymity for so long?
--
In an uncertain world, there is absolutely no security in banding together. -- Robert X. Cringely |
|
jap
Premium
join:2003-08-10
038xx
·RoadRunner Cable
| reply to salahx
salaxhx- honey, I'm baking you a batch of cookies!
Thanks (again) for taking the time to propigate info on these networks. Previous to stumbling on this thread I was altogether ignorant of this budding network model and find it more than a little interesting.
said by salahx :
said by jap :
You need to leave your "node" on for while (usually 1 or 2 days) to "learn" about the network.
Does that then mean that dynamic, short lease IPs with a forced churn throw a wrench in the works? |
|
jap
Premium
join:2003-08-10
038xx
·RoadRunner Cable
| reply to chris783
Re: new kinds of anonymous p2p software ...
said by chris783 :
For MUTE you are correct. There are virtual IDs which are very hard to map but thats one of the minor weaknesses of MUTEs protocoll.
Why would that be a weakness? Can anyone point curious me to a place I can read/learn more?
said by chris783 :
With GNUnet that is not existent cause routing is done by the entry/packet/request itself as an identification. Each node keeps track of the source-connection of routed packets so it knows where to send the answers when they come in.
Chris, please clarify: if I read your post correctly that means GNUnet is not anon as each "chunk" contains a travel history (clear-text IPs or names) back to source .... just like TCP/IP right? |
|
sorne guy
@milwwi.ameritech
| reply to lakino
quote:
This coming from a self proclaimed anon poster, posting under Sore Dud for at least a year. Why hide behind this anonymity for so long?
i use the same name, you see my isp, so what is anonymous about it?
there are no real benefits to logging in (other than not having someone steal my arbitrary nick), so why would i bother? |
|
chris783
join:2004-04-01
1 edit | reply to jap
said by jap :
said by chris783 :
For MUTE you are correct. There are virtual IDs which are very hard to map but thats one of the minor weaknesses of MUTEs protocoll.
Why would that be a weakness?
Because this ID is known all over the network and even if the attacker is 5 hops away from you he can identify every packet by your id. He can gather information about your node even if he is not connected to you. So if he is able (of course thats difficult) to map that ID to an IP address the game is over.
said by chris783 :
With GNUnet that is not existent cause routing is done by the entry/packet/request itself as an identification. Each node keeps track of the source-connection of routed packets so it knows where to send the answers when they come in.
Chris, please clarify: if I read your post correctly that means GNUnet is not anon as each "chunk" contains a travel history (clear-text IPs or names) back to source .... just like TCP/IP right?
No, sorry if i have been unclear.
That history is kept by each node itself. The packet gets a new sender address (your address) and then handed to some other nodes (node-choice is influenced by randomness, your and the target nodes bandwidth, cpu usage and other criteria to balance the network). This way the next node only sees information it can allready guess. -> The packet which is coming over your connection carries (in nearly all cases) your address as its sender.
Lets say i get a request for a file "abba" from connection #5. I make a note on the list like [conn #5 requested "abba"]. I will try to find the answer for this request in my database and then send the request to almost all other nodes which i am connected to. When finally the answer comes in i will look for "abba" in the list and find out who asked for it. Now i can send the answer for "abba" back to connection #5 and delete the list entry. |
|
chris783
join:2004-04-01
1 edit | reply to chris783
»www.ovmj.org/GNUnet/papers.php3?···=English
For documentation in generally this is a good start.
Even if that is almost about GNUnet specific things it is still generic enough to address common anon-p2p issues as well. Also security of known p2p networks in comparison is documented well.
[An Encoding for Censorship-Resistant Sharing]
is a little bit math/tech paper about GNUnet
[The GNet Whitepaper]
addresses the way of GNUnet networking more general and compares with other technology
[gap -- practical anonymous networking]
discusses anonymity under GNUnet (something like how anonymous can it be and what kind of techniques are available to achiev that) |
|
