On Friday 21 Dec 2012 21:54:54 Arne Babenhauserheide wrote: > Am Freitag, 21. Dezember 2012, 19:46:00 schrieb Matthew Toseland: > > Without any way to make it hard to run lots of nodes, AFAICS our only option > > for really strong security is darknet. > > How much would it cost to run 10.000 freenet nodes for one year? > > Let’s try to answer that: > > >From memory consumption I would guess that you can have 10 to 20 nodes on > >one > computer, if you avoid the WoT. My small homeserver here eats 20W, which > translates to about 20€ a year when you use industry-power (instead of my eco- > home-contract). With 20 nodes per computer, you need 500 computers to run > 10.000 freenet nodes. Which translates to 10.000€ a year (and an initial > investment of about 100.000€).
I am skeptical that you could fit 20 nodes on a 20W system; that's using what, a low-end laptop processor? > > Darn, 10.000€ is just what the BKA (german central police) spends yearly to > surveill skype… > > So they could easily afford making every second freenet node a police-node. Yes, that's the bottom line. The maths is maybe a bit better, but the basic costs are energy, hardware, and bandwidth. The bigger the system the more scope for optimising by 1) economies of scale (especially for bandwidth) and 2) custom code. We can try to make it a little more difficult based on scarcity of IP addresses or CAPTCHAs, but this will not add significantly IMHO as both are fairly cheap. We might try to limit the number of IPs in a given range but this will be problematic with real users being excluded... We need some source of scarcity that is expensive for an attacker but cheap for a normal user. Unfortunately it does not exist. > > And I just saw that simple relation: You need 1€ per year to run a freenet > node, plus an initial investment of 10€ - maybe less if you purchase in high > quantity. > > So we really need millions of users to make it hard to impede free speech - > or > we need widespread Darknet. Right. And I'm skeptical about "millions of users" even being enough, because a big target attracts more funding. Having said that, I think it should be possible to make it rather expensive to trace inserts, at a moderate performance cost.
signature.asc
Description: This is a digitally signed message part.
_______________________________________________ Devl mailing list [email protected] https://emu.freenetproject.org/cgi-bin/mailman/listinfo/devl
