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

Ian Clarke wrote:
> Even if we can't find a way to incorporate inserts into the "tit-for-
> tat" mechanism, there is still much value in implementing such a 
> mechanism for requests.

Maybe. I'm worried that if there's an incentive to forward requests but
no incentive to forward inserts, selfish nodes will have even more
reason to drop inserts.

> Even if C is faking responses, that will only lead  to
> B incurring a debt to C, which will make B less likely to forward 
> requests to C in future.

Sorry, why will B be less likely to forward requests to C? Perhaps we're
talking at cross purposes - what kind of mechanism do you have in mind?

> Essentially I think the question is whether it is really such a bad 
> thing if C can mislead its neighbors into thinking that it is  inserting
> when it really isn't.  What is the worst case scenario here?

C saves bandwidth by dropping inserts, so it appears to have lots of
bandwidth, and sends fake 'insert succeeded' responses, so it appears to
be super-reliable. C's neighbours give it lots of bandwidth in return. C
can mount a more powerful attack than it could have done without
reciprocation, because its neighbours allow it to use a larger share of
their bandwidth.

Cheers,
Michael
-----BEGIN PGP SIGNATURE-----
Version: GnuPG v1.4.3 (GNU/Linux)

iD8DBQFEdg/ayua14OQlJ3sRAnEwAKDL+u1bjU9YcBxciwkAjdf8hPeumQCgx5J/
el2XLfC3CUcTo6kCcwWH9Lg=
=dMSU
-----END PGP SIGNATURE-----

Reply via email to