On Thu, Dec 28, 2000 at 06:18:28PM -0800, Ian Clarke wrote:
> > The "shadowing" makes sense should the most important and scarce commodity
> > on the network be disk space
>
> Shadowing is purely a tool to aid in protecting people's nodes from "Media
> Enforcer"-style attacks, you are correct in saying that it will eat up
> more bandwidth, but this isn't the point. Also, I am not so sure that I
> agree that bandwidth is truely a scarce commodity since, at least not the
> "last mile", since if you aggragate most people's bandwidth usage over
> time (say 24 hours) it will only be a fraction of their potential
> bandwidth usage.
I think that the abundance of bandwidth you are seeing on the broadband
connections today is mostly a farce, that the companies can offer it
precisely because it isn't being used, and that should people start using
it the terms would have to change radically. I'm not very well read on the
topic though, so I could be completely wrong.
> > but I think it's more likely that the most
> > scarce commodity will be bandwidth, to which the "shadowed" nodes would
> > give no contribution.
>
> Er - "shadowed nodes" are not intended to be a benefit in terms of
> bandwidth, or diskspace usage, but in terms of helping people to reduce
> the probability that their node can be identified by someone like "Media
> Enforcer" or the Chinese government. I am therefore not really sure why
> you are bringing bandwidth into the discussion.
Just like you noted with the clusters, the question is whether operator of
the "shadowed" node is providing anything to the network that makes it
worthwhile (from the perspective of the network) for him to run a node at
all. Since his node is serving data he is obviously providing disk space,
but all the bandwidth his node is providing to the network costs the
gateway node equally much. So whether the existence of such a node is
worthwhile depends on which resource is more scarce.
> > But it's not a bad idea and since it is harmless when not used it may be
> > worth adding at some point.
>
> Do you think that it at-least partially addresses the "Media
> Enforcer" problem?
Well, it protects an individual node, but the practically viable "Media
Enforcer" attack (after we implement PKI) is simply to randomly attack all
public nodes in an attempt to stifle the entire network, so it seems to me
that any solution that depends on the existence of any public nodes is
only a slight improvement.
>
> Ian.
--
'DeCSS would be fine. Where is it?'
'Here,' Montag touched his head.
'Ah,' Granger smiled and nodded.
Oskar Sandberg
[EMAIL PROTECTED]
_______________________________________________
Freenet-dev mailing list
[EMAIL PROTECTED]
http://lists.sourceforge.net/mailman/listinfo/freenet-dev
