Chip biometric part N

 My continuing series on the analog chip biometric.

It is a one way analog logic circuit, with unique delays.  On the card, this is epoxied, shielded and layered in the card itself. An attempt to reverse engineer it and get a measuring device on it for a 10 pico second analog transistor set? We have transistor transit times of that order, talking any few ten of these disturbed and the original functionality gone.  Yet very cheap to implement in mass production, then a short learning period under the master fab.

Where is it vulnerable? Low security over the Internet, but the links are secure links anyway. So the attacker has to know when I am online with my card, then record everything, then run a super computer against it just to grb a low security key from a consumer with a cash limit, and then you have to still run a high powered CPU so it can act like the original analog. And then the currency cops will close that card on spotting the unathorized accounts.  Consider the trade off for a attacker, it is not nice.  This is a very good idea, I think it should be standard for smart card.

The cards come out of fab.  They hae aout four or five critical shared system keys.  But, if they are stored as forward solution to the polynomial, then the attacker still needs the analog machine. It is impossible o steal keys from the card.  Even though the keys are encoded, they are in kernel memory, the attacker needs to infiltrate the kernel non destructively.  Security is variable over a range, by precision,  If communication with the fab is rare, make it a high precision encode, the attacker needs about 2^precision to reverse it, but the storage goes up as the square, likely. There is a very wide range.

This works, I think. Well enough, I cannot find an easy attack, but I can find attacks. In the attack you carry a specially programmed high performance cpu in your pocket, with a wired connection to the fake.  This card is having a $200 limit, it is not worth the time.  Two parties can verify to near perfection, the other is not a double spender.