On Tue, Aug 15, 2006 at 11:52:58AM -0700, Ian Clarke wrote:
> 
> The opennet proposal is far from perfect, but there are measures that  
> can be taken to make harvesting more difficult,

On 0.7 opennet, we won't send a reference after a successful request; we
will offer an extra stage for connect-to-the-source. This is necessary
to deal with NATs. The cost is that the nodes in the middle know whether
you connected to the source, but this is irrelevant as they don't know
the source. (Except perhaps a hash of its identity to prevent connecting
to nodes you already know).

This means that harvesting 0.7 will require actually connecting to the
harvested nodes. This is however good practice anyway for a harvester;
they will give you access to more of the network, and the connects use
up no more bandwidth than the requests that preceded them.

I don't think this is particularly more secure than 0.5. We can
artificially reduce the rate of connection churn, but we have to reckon
with node churn; more node churn will require more connection churn.

Also it must be possible to connect to a node in the seednodes list
without prior communication.

> and its decentralized  
> nature makes it harder to compromise. 

Compared to #freenet-refs ? Sure. Compared to a true darknet, no. I'm
not saying don't do opennet, I'm saying don't forget darknet, and don't
do opennet until we have some reason to think it will work: Simulations,
and a decent effort at solving the larger problems facing the network
today.

> And of course, lets not forget  
> the issues of user convenience and actually achieving a small world  
> network topology, neither of which is provided by #freenet-refs or  
> the other forums you mention.
> 
> Ian.
-- 
Matthew J Toseland - toad at amphibian.dyndns.org
Freenet Project Official Codemonkey - http://freenetproject.org/
ICTHUS - Nothing is impossible. Our Boss says so.
-------------- next part --------------
A non-text attachment was scrubbed...
Name: signature.asc
Type: application/pgp-signature
Size: 189 bytes
Desc: Digital signature
URL: 
<https://emu.freenetproject.org/pipermail/devl/attachments/20060815/e9e392c0/attachment.pgp>

Reply via email to