-----BEGIN PGP SIGNED MESSAGE-----
Hash: SHA1

Matthew Toseland wrote:
> We can prevent flooding with useless content by for example having
> inserts not start at the top of the LRU.

I was thinking about bandwidth/request flooding rather than
datastore/cache flooding, but I guess that's an issue too.

> Can we add in insert verification later on?

Are we even sure it can be done? If not, adding a tit-for-tat mechanism
based on the number of requests answered will just create an incentive
to drop inserts (and possibly lie about it) in order to answer more
requests.

> The basic problem is that we need to tune things so that it is cheaper
> to keep using an existing node, and build up some trust with it, than to
> connect to a new node.

I agree this is crucial. The problem's made more difficult by the fact
that new nodes aren't very good at answering requests (they have empty
stores and few connections), so we can't expect them to perform as well
as our existing neighbours to start with.

> This isn't necessarily insoluble either: Although
> most successful requests will likely come from a new node (on a large
> network), we won't necessarily succeed in our attempt to connect to
> them.

Sorry, I don't follow - are you saying the difficulty of finding and
connecting to nodes will be high enough to discourage people from
exploiting them and moving on?

Cheers,
Michael
-----BEGIN PGP SIGNATURE-----
Version: GnuPG v1.4.5 (GNU/Linux)

iD8DBQFE62/tyua14OQlJ3sRAp3xAJ9E61xGXOXkb3pyzbeWRcuX9V+oWgCgngnG
VGW1fhzt0uUpH2VleyY7ias=
=dFl1
-----END PGP SIGNATURE-----

Reply via email to