Study of the amount of fees paid to discover how many time a secret contract is executed


#1

Do you guys already have an ideia how to tackle the amount of fees paid to run a computation of a certain secret contract and hide it from public eyes so you can’t do a study of how many times that secret contract was executed?

This question is relative to the last question made in this video

thanks!


#2

There are no plans currently to hide access patterns to a secret contract. It is doable, but requires quite a lot of orthogonal work and we haven’t been able to come up with serious use cases where this makes sense.

We did put a lot of thought into how to hide access patterns to state memory, because that could reveal a lot. That said, I’m curious - what kind of adverse leakage would you expect from simply observing how often a secret contract is used?


#3

Guy: it allows you to approximately estimate when new data enters the system, which makes it way easier to have needle-in-a-haystack deanonymization attacks on large datasets. With DP this isn’t an issue, but DP’s going to be (iirc) a later add-on to the network


#4

That’s still secondary in terms of priority.

Also, I don’t think DP is the right approach here. When we move to our own chain, we could treat ENG as another privacy token (as mentioned here). Until then, and while we rely on Ethereum for TXs - we aren’t trying to hide tx meta-data.