Continued from >>51442732
GNUnet is a fully decentralised p2p framework for a number of applications, the most commonly used one (probably) is filesharing
>Why use GNUnet over private trackers?
>have to keep up a seed ratio or you're thrown out
>have to attach an account to your searches and requests
>have to trust that the private tracker's owner won't leak your details anywhere
>uploading must be approved
>no seeding requirement
>no account needed
>no trusted 3rd parties needed
>no way to trace any traffic back to you, unless you set anonymity to 0
>even just searching for content can be done via GNUnet's anonymous protocol
>anyone can publish anything
>no way for publishers to know who's downloading
>no way for downloaders to know who's publishing (except via optional namespaces)
>Sounds cool, where do I start?
/g/ GNUnet QUICKSTART GUIDE FOR FILE SHARING
### INSTALLATION ###
download and install using the package manager of your distro
### POST-INSTALLATION/USAGE ###
*Before doing anything else, run following command in terminal: gnunet-arm -s
To use the GUI, run gnunet-gtk and for setup/settings run gnunet-setup
For help on uploading/searching/downloading files via the GUI, see https://gnunet.org/first-steps-file-sharing
GNUnet can also be used in the terminal
To upload a file: gnunet-publish [-n] [-k KEYWORDS]* [-m TYPE:VALUE] FILENAME
*Note: see https://gnunet.org/gnunet-publish for more information
To search a file: gnunet-search "[SEARCH TERM]"
>example: gnunet-search "/g/"
To download a file: gnunet-download -o FILENAME -- GNUNETURL
*Note: the gnunet-search command will output the command needed to download the desired file
To shut down GNUnet: gnunet-arm -e
### ADDITIONAL RESOURCES ###
I'm keeping an eye on these threads,Sir. I think a major issue with getting people interested is the lack of stuff to download.Maybe we should set up a scavenger hunt or something. The /g/ search only goes so far when nobody wants to upload anything.
Doubt it, simply because ipfs is so fast you can already stream 720p videos over it no problem, while on gnu-net you can't even download sakura-fish unless you can dedicate the entire day to it.
That's just another issue i have with this. I guess pic related or iptraf but honestly you still have no idea what is going on with the connections or what you are transmitting or even if you've seeded a single file to completion.
also i hope for the love of god someone realizes that 5000B/s is about the worst starting speed for a file sharing service. You start it high, people download much faster and like the service enough to edit the settings to their liking. You start it low like gnunet did and nobody wants to stick around (pic related again - green+yellow is some douche uploading at 5kbs)
oh well though ded thred is ded
pains me that I cant get it to work on a VM. my laptop is too slow to seed plus I dont have it on everyday.
Freenet has a hard limit as to how much the network can serve. As a node, you don't manually host content so you can't just buy a new hdd and keep going; the network itself has to serve the content and there is a capacity limit because of this.
How are they gonna serve all that child porn then? This is why Tor is superior to freenet, Tor has unlimited capacity (for all intents and purposes) for serving child porn.
Or is this another one of those useless "anonimity" services that isn't meant for hosting child porn? Btw, is there child porn on gnunet?
(I've had a good laugh while writing this, pls don't van me NSA.)
But seriously though, tripsdubsguy, lack of decentralisation is the number one problem with the internet, and tor does nothing to aleviate this. That's basically how they could take down the silk road, and how they sensor clearnet sites: Find the server, and the site is down.
tor and i2p _are_ decentralized (modulo the DHT seed hosts, just like bittorrent, ipfs, and gnunet), but you're right that they're different than ipfs/gnunet/freenet in that they provide anonymous point-to-point communication, not distributed data storage. They're decent platforms to build p2p storage on though. i2p has a bittorrent client built-in, and there are some large tahoe-LAFS pools running on it.
Oh that makes sense. It makes me wonder what would the world be like had the silk road guy not fucked up (I mean had he configured the server properly, the real ip wouldn't have leaked).
Tor essentially gives you an anonymous tunnel between you and another host (be it the clearnet through an exit node, or an onion site). To do so, the network finds 3 different hosts running tor nodes, sets up non-anonymous but encrypted tunnels between them (3 layers of an onion). Those nodes are run by random people and coordinated through a DHT. That's the decentralized part. Attackers can compromise or take down a percentage the nodes with out affecting the privacy or availability of the network. If you go to the tor website, you'll find plenty of information for running your own node, and assuming you're not the NSA, that's one more node that would have to be compromised to take tor down.
There's a lot of misinformation flying around the ipfs and gnunet threads. I should really write a general FAQ at some point.
It's not decentralised just because the traffic goes through different nodes, the content disappears if one of the nodes taken down happens to be the host you're connecting to. The site is served from one server, hence tor is centralized.
If it's decentralized, the information would be spread around on different servers, if one goes down you're routed to another one.
So please don't write that FAQ.
Tor doesn't serve data, it serves anonymous tunnels which are decentralized for all useful definitions of the term; spelled out, you have to take down approximately 50% of the nodes running tor before you can no longer form an anonymous tunnel. If that's not what you want to mean by the word "decentralized", fine, but that's not actually what matters.
Well, obviously we're talking about serving of data, and obviously that actually matters (I mean, I've already mentioned the silk road bust, but the examples are endless).
Oh, and much more than 50% of the nodes can be taken down with tor still functioning well. But no more than 50% of the nodes can be owned by a single provider (aka NSA) before the anonymity is compromised.
Getting 5 people for free now is better than losing the chance when they get huge and the network grows with them. Though I agree that it's unlikely to be compromised yet in practice.
Okay, we're in agreement then. There's nothing inherent in tor that prevents data distribution, just that it's not built in, so you have to run, say, bit torrent over it. The reverse is also true, there's very little privacy built in to bit torrent (or ipfs).
Re: the silk road, I think that's more of a problem of centralized trust rather than data. They could have distributed the servers like the pirate bay and told people to use a different onion address, but since there was only one guy keeping the escrow keys and managing the accounts, the feds could take him out and kill the whole thing, even if another admin restarted the server with the same data elsewhere. https://openbazaar.org/ is looking nice though, hopefully that goes somewhere.