Monday, November 20, 2017

Intel software guard plus bulletproofs equal verification

Stanford is developing short proofs of knowledge from some known number algebra.  (Crypto is really applied abstract algebra, bullet proofs and two color pits have a lot in common).

But the idea is simple.  Let me use generator terminology.    Normally they use finite log, but so do the pits, use finite log with some bound and conserved error.

Assume I have the secret generator, and can generate a sequence very similar to my secret sequence.  The other party can look at the member of a set of sequences, the one I send it. The other party verifies that I know the generator. Both parties then know the correct sequence will be transferred on request.

Now from  coin point of view, this is great. When the user signs up for the app, the smart cash card can get a secret generator.  This will be part of the 'token', or perhaps contract is a better word, but the theory will use protocol.    The user signs up and can verify he has the proper generator, the contract is valid.

Then it can verify account and protocols without revealing the actual bits.

This will find uses just above the hardware protection,  and bullet proofs can be an available function, within the protected element, hopefully commonly used.

Here is the coolest thing about the tech.  Short proofs mean there is not a lot of room for text and laundry, the code lengths are minimal, it happens on the independent security queues and is available to cybercops.  It is mostly verifying that some smart card is configured for the various trading protocols in the app.

No comments: