On Fri, 12 Sep 2003, Edgar Friendly wrote:

> Thomas Leske <[EMAIL PROTECTED]> writes:
> 
> > The intension of probalistic caching is to keep data longer
> > available in the network and improve routing. It works by
> > giving data source nodes a better clue on the global popularity
> > of the data, because the reloads increase its local popularity.
> > 
> <SNIP>
> 
> let me get this clear, you're proposing to have nodes pretend they
> don't have data?  This seems to have the downside of non-probabilistic

... unless it comes back as DNF, in which case they answer "Yes".
Basically a "I have this, but let's let the 'owner' know it's popular"
scheme.  Better to translate it to an InsertRequest, though.  That way,
if you get a collision, great, the data is still there.  If not, you
send the data both ways, up to the "best" node and down to the node
requesting it.

Even with those fixes the proposal loses.  It solves the
"popular-to-death" problem but opens up new flooding attacks.

Popular-to-death is a very difficult attack where you request the same
key from lots of nodes that should NOT be specalized in it.  When they
find the key, you've pulled it to the edges of the network... now
requests for it never reach the "best" node so it dies of LRU.  Quit
requesting it and many of the edge nodes will begin dropping it... and
eventually the content is gone.  (Note: Edge nodes are the nodes most
unspecialized for a particular part of keyspace.  They should have the
most hops to reach it.)

The fact that freenet routing is so dynamic makes this attack neigh
impossible.  Flooding is a much more valid concern.

--Dan

Attachment: pgp00000.pgp
Description: PGP signature

_______________________________________________
Devl mailing list
[EMAIL PROTECTED]
http://dodo.freenetproject.org/cgi-bin/mailman/listinfo/devl

Reply via email to