copy legitimate IDs from existing products.
Attack 2 - copy legitimate IDs from existing products
Attack 3 - Steal bulk IDs from the database
Attack 4 - Subverting the Channel
Attack 5- Subverting the Database
Attack 6 - Subverting the Server
This attack is exclusively against the anti-counterfeit protocol.
It can be seen that even with as little as 1% registration and no chain of custody at all (i.e. the Supply chain isn't wired up or isn't co-operating) the maximum possible sale before detection becomes a statistical certainty is 10,000. (1% registration among the buyers of the 10,000 fakes means 100 registrations. If a similar proportion of consumers register the matching legitimate products, then every fake registration has a 1% probability of collision with a real product registration - which will raise the alarm. A hundred fake registrations is thus, hypothetically, 100% probable to experience at least one collision.)
Already this places an unacceptable constraint on many of the counterfeit markets which rely on - or at least expect - volumes sometimes in millions. If Microsoft used this system for Windows XP, and could rely on as little as 1% registration, they would detect the first counterfeits within seconds. And if we could restrict piracy of Microsoft products to a mere 10,000 they would be ecstatic.
Of course, for Microsoft, things aren't that simple. Their product is selling vastly more in the CC market than the FD market, so voluntary registrations aren't going to happen. Which is why they're trying to implement a credible mandatory activation protocol as an alternative to registration. Our prediction is that it will achieve a certain level of damage limitation but that, worldwide, there will still be at least 2 counterfeits on the desktop for every legitimate copy. (The activation procedure was "cracked" 3 months before the product even hit the streets.)
In the FD market though, registration at 1% limits counterfeit sales to a volume dependent on whatever risk the counterfeiter is prepared to run, up to and including 100%. If he's that stupid he'll shift 10,000 and be caught. If he's more cautious, he'll limit his sales to a very quick 1000 and give himself a 90% chance of escape.
At 2% registration, the certainty of detection is reached at just 2,500 sales, while keeping the risk to acceptable levels probably reduces actual sales to a couple of hundred. This is well below the viable volume for any serious counterfeiting operation except for very high value products (which, by virtue of their high value, will probably enjoy registration rates of 90% and thus detect counterfeits on a near one to one basis)
Throw in the full co-operation of the supply chain, generous marketing incentives for such co-operation and for consumer registration and you can see that it very quickly becomes impossible to sell more than a handful of even relatively low value branded goods without detection.
The beauty of the system is that this result is independent of how good the copy is. It doesn't require sophisticated, expensive to implement and difficult to emulate, labelling or design. The unique ID can be printed on in plain text or even written on in crayon! All that matters is that it is a unique id and that it passes or fails the VR tests.
