search

dbosk / crocus

Securely and privately verifiable protests
Other
0 stars 0 forks source link

Security definitions of individual and universal verifiability #31

Open dbosk opened 6 years ago

dbosk commented 6 years ago

We need adversarial games for individual and universal verifiability. In general I think we must show soundness and completeness, i.e. that we can prove what we want.

dbosk commented 6 years ago

If we assume that the storage is immutable and unconditionally available, we should be fine. We simply assume that someone else solved the problem. However, I suppose we have the CAP theorem to consider. And then I'm not sure if it's a realistic assumption. If Eve creates a partition, then we will have neither individual nor universal verifiability. On the other hand, eventually Eve's partition will likely go away. Then we'll be fine if the system can recover to a consistent state.

dbosk commented 6 years ago

Individual verifiability: a valid but missing proof. Why is this a problem when we only provide a lower bound on number of participants?

Because you want as high a lower bound as possible. In the extreme case, zero is also a lower bound :-)

Maybe we should call it infimum (the greatest lower bound) instead of a lower bound.

dbosk commented 6 years ago

Maybe there are formal definitions that are useful for us in the Delaune papers, but I don't think this is important for the submission.