Is PrivateSend sybil resistant?

[Antti Kaikkonen]

The collateral fee which is paid randomly (10% chance) has been reduced from 0.01 to 0.0001 in the last year or so. The cheapness of mixing leads me to think that an attacker could run a lot of mixing clients to become the only mixing participant to some honest mixers. If you are only mixing with one person then the other person can know the source of your PrivateSend transactions.

 

[qwizzie]

The collateral fee which is paid randomly (10% chance) has been reduced from 0.01 to 0.0001 in the last year or so. The cheapness of mixing leads me to think that an attacker could run a lot of mixing clients to become the only mixing participant to some honest mixers. If you are only mixing with one person then the other person can know the source of your PrivateSend transactions.

The way i understand is that Dash always needs an minimum of 3 participants to mix, which can extend to 5 participants (i'm not totally sure about this last part)

 

[Antti Kaikkonen]

The way i understand is that Dash always needs an minimum of 3 participants to mix, which can extend to 5 participants

Do you know how the participant is defined here? How is it prevented that the the other 2, 3 or 4 participants are not the same entity?

 

[qwizzie]

Do you know how the participant is defined here? How is it prevented that the the other 2, 3 or 4 participants are not the same entity?

I found the github pull : https://github.com/dashpay/dash/pull/2612 ([PrivateSend] Allow more then 3 mixing participants#2612)

Make number of PS mix participants random, and make it known only to MN.

The number of mixing participants will now randomly range from 3-5 resulting in a average of 4. This increases the raw anonymity set by over 3 times (4 participants ^ 4 rounds) / (3 participants ^ 4 rounds) as well as removing a major metadata leakage.