Skip to Content.
Sympa Menu

bluesky - Re: weak secure IDs

bluesky AT lists.ibiblio.org

Subject: Global-Scale Distributed Storage Systems

List archive

Chronological Thread  
  • From: Wei Dai <weidai AT weidai.com>
  • To: Global-Scale Distributed Storage Systems <bluesky AT franklin.metalab.unc.edu>
  • Subject: Re: weak secure IDs
  • Date: Mon, 30 Sep 2002 14:12:04 -0400


On Mon, Sep 30, 2002 at 05:53:15AM -0400, Zooko wrote:
> I like your idea for incrementally more expensive IDs.
>
> However I don't like the idea, in general, of allowing nodes to cause
> damage
> but attempting to constrain the number of such nodes. Have you seen "The
> Sybil Attack" by John Doceur? [1]

The first idea seems only useful in the context of the second idea, so why
do you like the former without liking the latter?

Also, what's the alternative to the second idea? Are there systems that
can't be damaged by malicious nodes?

> I don't at all agree with the conclusions, but he makes a good argument.

One of his arguments is that all identities need to be validated
simultaneously, because otherwise an attacker can generate new IDs one
after another. My scheme does not require simultaneous validation, because
new IDs become exponentially more expensive for the attacker to generate.

Another is "When accepting identities that are not directly validated, the
required number of vouchers exceeds the number of systemwide failures."
In my scheme, all IDs can be directly validated at minimal cost, so this
is not a problem.

His only other argument is that all entities would have to operate under
nearly identical resource constraints. This does apply, in the sense that
an attacker can prevent nodes with slow CPUs from participating by raising
security levels, but I think it's an acceptable loss.




Archive powered by MHonArc 2.6.24.

Top of Page