>>>>> "SGM" == Scott Gregory Miller <[EMAIL PROTECTED]> writes:
OS> We have no solution to the problem of the honest cancer (not
OS> even Mr. Bad's "lets break freenet" proposal really
OS> helps).
Me> En garde! Defend that statement, sir.
SGM> On the contrary, the burden of proof lies with you. You must
SGM> prove that a system does work, not rely on others to prove
SGM> that it doesnt.
Yes, that's why I drew out all the diagrams and gave the long-winded
explanation of isolating hostile nodes.
However, I think I might have misunderstood Oskar's statement the
first time around -- I missed the "honest" part. The difference
between the hostile node that I described and what he's calling an
"honest cancer" is that the hostile node returns verifiably "bad"
data.
The "honest cancer," on the other hand, is just -too- good: a
well-connected, large-cached node. Its "routing gravity" is high
enough that more and more routing entries point to it over
time. Effectively it becomes a central point of failure that can be
compromised.
I agree with Oskar that what I described wouldn't solve this
problem. The only thing I can think of that would is having the
routing algorithms try to preserve diversity in the routing table --
which would probably be hard to do right.
~Mr. Bad
--
~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
/\____/\ Mr. Bad <[EMAIL PROTECTED]>
\ / Pigdog Journal | http://pigdog.org/ | *Stay*Real*Bad*
| (X \x)
( ((**) "If it's not bad, don't do it.
\ <vvv> If it's not crazy, don't say it." - Ben Franklin
~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
_______________________________________________
Freenet-dev mailing list
[EMAIL PROTECTED]
http://lists.sourceforge.net/mailman/listinfo/freenet-dev
