On Monday 05 January 2009 18:46, Gregory Maxwell wrote:
> 2009/1/5 3BUIb3S50i 3BUIb3S50i <3buib3s50i at gmail.com>:
> > The exchange of USB keys seems very far from the initial project of
> > Freenet... Too slow, not anonymous. TrueCrypt is a better choice in this
> > case!
> 
> USB keys should be no less anonymous than darknet freenet is normally:
> You know who your neighbors are in the network, but not further, and
> you don't know who a request came from.  It may even be more anonymous
> than normal freenet because the batching would create bigger 'mixes'
> and some kinds of timing attacks may become impossible.
> 
> You have it a bit backward: Truecrypt would have zero anonymity and it
> would not really be a 'network' without a lot of manual labor.
> 
> ... though I suspect a "batch-mode" for freenet might ultimately find
> more utility for opportunistic exchanges with ad-hoc wireless.

Opportunistic ad-hoc wireless isn't freenet, it's Haggle. It's very far from 
meeting our standards for security, but against some threat models, and given 
some hardware assumptions, it may be useful. The basic principle is to have 
your cellphone announce to everyone on the same train that you want a 
specific, probably illegal, file, and hope that nobody can trace your request 
before you leave. It ignores the fact that cellphones are probably the most 
locked-down widely used hardware in existence apart from cable/satellite 
decoders and maybe games consoles, and it doesn't fit into freenet's current 
architecture either for darknet or opennet.

However, ad-hoc networking with known friends, over some short-range 
high-bandwidth low-power system, maybe UWB-based, is a good idea and 
hopefully will happen one day, but it would need a huge amount of burst 
bandwidth to get a reasonable overall transfer rate... exchanging USB keys 
could get around 1Mbps per peer equivalent, assuming 8GB sticks swapped 
daily.

Anyway, all this delay tolerant stuff relies on figuring out a way to do 
delay-tolerant location swapping, which is far from trivial, and may not be 
possible.
-------------- next part --------------
A non-text attachment was scrubbed...
Name: not available
Type: application/pgp-signature
Size: 827 bytes
Desc: not available
URL: 
<https://emu.freenetproject.org/pipermail/tech/attachments/20090105/1176fa7e/attachment.pgp>

Reply via email to