GNU/consensus/whistle/012013-08
Contents
GNU/consensus Whistle
Volume I, Number 1
Editorial: Internet End-to-End
Stuff I want to talk about for August... The P2P / cypherpunk side of socialnet (minus Bitcoin, which will come in an upcoming issue about p2p-economy). Make this issue more attractive to potential contributors to the projects...
You Broke the Internet. We're Making Ourselves a GNU One
Keywords: GNunet, SecuShare, Internet Censorship, and Free Software.
https://gnunet.org/internetistschuld
This is the video from the talks given by Christian Grothoff, Carlo von lynX, Jacob Appelbaum and Richard Stallman in Berlin on August 1st. The talks are in English, even though the welcoming words are in German.
Christian Grothoff's talk summarized the recent revelations about PRISM and their implications for non-American citizens, industries and governments. It then presented technical solutions towards a secure and fully decentralized future Internet, which would address key challenges for self-determined life created by the world-wide police state.
Interesting details on this:
- A new cryptographic method for a privacy-capable DNS/DNSSEC replacement, called GADS.
- A faster and smarter extensible messaging syntax than XML and JSON, called PSYC. (WIP--add link to tg's text before launch)
- A strategy for distributed and liberated Internet search, called RegEx.
Carlo von lynX gave a presentation on how secushare intends to provide messaging and Facebook-like functionality on top of GNUnet. Keywords: Scalability by multicast; Social graph vs. Onion routing; Unsafety of your own server.
Richard Stallman and Jacob Appelbaum added closing notes of free software and free hardware and responded to questions.
It's not about how much you want to make believe you got nothing to hide. It's about your civic duty to not be a predictable populace. -- lynX
Big thanks to the Pirate Party for providing the venue and the recording technology.
Failure to Synergize
by lynX
Eleanor asked for cooperation again, although we already concluded at last year's meeting that we should stop developing horizontally, each project its own complete new Internet stack, and rather work vertically: DHT hackers focus on the P2P backbone, others focus on the distributed messaging capabilities and ultimately all the GUI and usability hackers build on top of it. Thus, there should be only one core P2P engine and only one multicast implementation that allows for long-term scalability – or at least we should focus on having just one of each until we have at least one complete functional stack. Then we can always fork into all directions and improve on this or that.
The distributed routing core
The de-facto leading core is Tor. Hidden services are for the win. Issues with exit nodes or javascript-based traps are not relevant for our purposes, or if they are, then they will affect us on any routing core. Actual problems with Tor are the necessity of directory servers and, if I'm not mistaken, the unflexible number of hops - too low for some requirements, too high for real-time streaming. Not sure if it makes sense to work on Tor or any other routing engine as GNUnet seems to be conceptually way ahead. That's why I don't understand why projects like Briar, Retroshare and Tox are still spinning their own wheel - some of them not even offering onion routing. I thought we had sorted that out at the last meeting, but maybe somebody can explain to me what I am missing. Maybe we should all learn the tools first.
The message distribution layer
Since our application focus is social we have the most massive scalability problem there is, therefore if we don't want to do applications limited for small groups we must solve the multicast distribution issue. It's still rocket science, but it has been done before. Even Facebook and Google have multicast distribution strategies somewhere in their cloud backbone. We're not alone. Payload extensibility is also a plus. This is what secushare is currently working on and I don't see other projects providing this part of the puzzle in any form, yet. Consider that even if you think servers are okay and federation architectures aren't all bad, you still need a distribution strategy – and that just hasn't been dealt with. At least from my point of view. Not sure if yours is sufficiently different to not consider this the bottleneck of ten years of attempts to redo social networking in a free way.
The GUIs
This seems to be the area that has seen most work. Some like Retroshare have developed native GUIs while the majority has done web-based interfaces. As long as the address in the browser is something like localhost, even web-based can do, so there is plenty of possibility to try out different user interface approaches and application focuses on top of the same backbone.
Alpha
The Briar Project
The Briar project is building secure communication tools to enable journalists, activists and civil society groups to communicate safely without fear of government interference.
cjdns
cjdns is a networking protocol and reference implementation, founded on the ideology that networks should be easy to set up, protocols should scale up smoothly, and security should be ubiquitous.
GNUnet + Secushare
Not just obfuscated P2P with a DHT, we also address scalability by use of stateful multicast subscription channels, addressability by a new cryptographic name resolution strategy called GADS, application flexibility by letting the app choose the number of hops for each message and extensibility by means of the PSYC syntax which performs almost as fast as binary protocols while providing comparable semantic richness to XML.
https://gnunet.org/ and http://secushare.org/
See also https://secushare.org/comparison
Pond: Like Email But Not Quite
- state: alpha (do not use it)
- ephemeral email-like messaging with perfect-forward-secrecy
https://pond.imperialviolet.org/
Tox Is Not Skype
Issue #58
<lynX> From what i gather in the roadmap they chose to redo yet another DHT backend instead of building upon existing ones. There is no sign of onion routing thus I presume Tox is trying to achieve what Retroshare already provides today: end-to-end encrypted chat, VoIP and more without meta-data obfuscation.
Beta
GNUnet
Powerful engine, but not much application for it yet. See also https://gnunet.org/compare
I²P: Invisible Internet Project
I²P (Invisible Internet Project) is a computer network layer that allows applications to send messages to each other pseudonymously and securely. Uses include anonymous web surfing, chatting, blogging and file transfers.
Read Christian's paper on ISP: http://grothoff.org/christian/i2p.pdf
RetroShare
<lynX> ... does the full show: messaging, web sites, forums, VoIP, social. The GUI even looks end-user compatible. Apparent weaknesses: Onion routing only on demand, social graph visible by direct links, no multicasting strategy, no forward secrecy(?), ...?
Tor Needs Relays
Relay the (permanent) call for relays...
<lynX> then again it is quite performant these days.. how is that possible? Hidden services show how to use Tor for real. torchat is just a simple example of how they perform, but really everyone should operate their personal crypto mailbox on a hidden service address. It's the first step in re-inventing the Internet, do your old-fashioned Internet services behind a .onion. We can gain some experience from this and work it into our new developments.
Events
- Internet Ist Schuld, Berlin 01 August
- FedSocWeb, OHM2013 02 August
- Socialnet_3.0, Berlin 24-25 August
