Well, as long as it is only search entries being "unrequsted" the attack is not
of that big a concern, seeing as search entries are small (bandwidth wise, and
this is a bandwidth attack).
Speaking of floods, maybe you covered this before Ian, but how to do you want to
terminate search requests? As I read your description ("In the event of a
timeout, the best score will be noted, and...") it seems you want for all
searches to go on until they timeout. But wouldn't that cause the same overflow
of nodes in popular parts of the searchspace as you were concerned with in the
keyspace if requests for SVKs sometimes fall through?
A possible solution would be to place a limit (x hits with result greater y)
but it is still possible that a lot of people would be searching for things
with that threshhold set higher then anything they will find...
On Fri, 19 May 2000, Scott G. Miller wrote:
>
> > I still think that the same problem is solved by filtering hits based on
> > signatures and I prefer filtering to unrequesting.
> In light of the Request, Unrequest, Request (RUR) attack, I'm going to
> have to agree.
>
>
----------------------------------------
Content-Type: application/pgp-signature; name="unnamed"
Content-Transfer-Encoding: 7bit
Content-Description:
----------------------------------------
--
Oskar Sandberg
md98-osa at nada.kth.se
_______________________________________________
Freenet-dev mailing list
Freenet-dev at lists.sourceforge.net
http://lists.sourceforge.net/mailman/listinfo/freenet-dev
