#dusk $DUSK @Dusk _Foundation
Someone asked me how a “private” chain can still check your balance if the balance is hidden, and I realised I couldn’t actually explain it.
So I went digging into how Hedger does it, and honestly, it still feels a little weird.
Normally, if a system needs to do something with a value, it has to read the value. Add two balances, check whether something is over a limit, whatever it is. The actual number is usually sitting there somewhere.
That’s where Hedger gets interesting.
Dusk’s privacy system for the EVM side uses homomorphic encryption — ElGamal over elliptic curves, specifically. The basic idea is pretty strange: you can perform certain computations on encrypted values without first decrypting them. You get the correct result while the underlying values stay encrypted.
Then you add zero-knowledge proofs, so you can also prove the computation followed the rules without exposing the private inputs.
The part that really caught me is this:
the plaintext balance doesn’t have to be exposed to the contract for the rule to be enforced.
For regulated finance, that could be useful for things like eligibility checks, balance limits, transfer restrictions, and other compliance logic where the sensitive numbers shouldn’t be sitting in public view.
But there’s an obvious question.
Homomorphic encryption is computationally expensive. Being able to compute on ciphertext doesn’t mean every kind of computation is cheap, fast, or practical.
So I’m genuinely curious about this part:
**where is the performance line today? Is HE actually practical for meaningful financial logic, or are we still mostly talking about narrower operations?**
$PORTAL $GPS
Someone asked me how a “private” chain can still check your balance if the balance is hidden, and I realised I couldn’t actually explain it.
So I went digging into how Hedger does it, and honestly, it still feels a little weird.
Normally, if a system needs to do something with a value, it has to read the value. Add two balances, check whether something is over a limit, whatever it is. The actual number is usually sitting there somewhere.
That’s where Hedger gets interesting.
Dusk’s privacy system for the EVM side uses homomorphic encryption — ElGamal over elliptic curves, specifically. The basic idea is pretty strange: you can perform certain computations on encrypted values without first decrypting them. You get the correct result while the underlying values stay encrypted.
Then you add zero-knowledge proofs, so you can also prove the computation followed the rules without exposing the private inputs.
The part that really caught me is this:
the plaintext balance doesn’t have to be exposed to the contract for the rule to be enforced.
For regulated finance, that could be useful for things like eligibility checks, balance limits, transfer restrictions, and other compliance logic where the sensitive numbers shouldn’t be sitting in public view.
But there’s an obvious question.
Homomorphic encryption is computationally expensive. Being able to compute on ciphertext doesn’t mean every kind of computation is cheap, fast, or practical.
So I’m genuinely curious about this part:
**where is the performance line today? Is HE actually practical for meaningful financial logic, or are we still mostly talking about narrower operations?**
$PORTAL $GPS
