Finding a deterministic way to split a secret into 6 shards

6 replies 57 views
the_hashMember
Posts: 8 · Reputation: 168
#1Mar 9, 2021, 08:37 PM
I'm looking for the best deterministic approach to split a secret into 6 pieces. No Shamir Secret Sharing Scheme, please. The goal is to be able to rebuild the original secret using any 2 of those pieces.
5 Reply Quote Share
pixel2014Hero Member
Posts: 857 · Reputation: 4132
#2Mar 11, 2021, 03:11 PM
The best is still the Slip39 mnemonic shares. It is available on wallet like Electrum and Trezor hardware wallet.
6 Reply Quote Share
paul2017Senior Member
Posts: 218 · Reputation: 1426
#3Mar 11, 2021, 05:09 PM
Please forgive my ignorance, but what's wrong with SSS and what do you mean by "*deterministic*"?
0 Reply Quote Share
HyperRavenFull Member
Posts: 175 · Reputation: 633
#4Mar 11, 2021, 05:56 PM
I'm assuming you're more concerned with having multiple entities hold shares of a secret, and in the context of Bitcoin to sign transactions. SLIP39 which is an implementation of SSS is a good one, though it has never been formalized into a Bitcoin standard. Blakley's scheme is another form of secret sharing, though I prefer Shamir's, partly because of SLIP39's existence and also that it doesn't offer additional benefits. As far as information-theoric secure, I believe these are the two main ones. Recently, there's a scheme called Multi-Party Computation as well. If you're going with that then might as well use Multi-sig since the latter is formalized into a standard.
4 Reply Quote Share
hash_bossLegendary
Posts: 1166 · Reputation: 5261
#5Mar 13, 2021, 07:06 AM
I believe he means the output is always same. AFAIK some SSS tool (such as https://iancoleman.io/shamir/) would generate different output if you use exactly same input on different occasion.
3 Reply Quote Share
im_apeHero Member
Posts: 629 · Reputation: 3824
#6Mar 13, 2021, 12:48 PM
Is deterministic Secret Sharing Scheme even possible? As far as I know these algorithms rely on randomness and if their randomness is "guessable" they lose their purpose.
1 Reply Quote Share
paul2017Senior Member
Posts: 218 · Reputation: 1426
#7Mar 13, 2021, 12:56 PM
Then the solution is simply to use non-random inputs, though as pooya87 pointed out, that would introduce weaknesses.
3 Reply Quote Share

Related topics