Skip to main content

Split a secret

Before you can split any secret, you have to configure a cyclic group to be used for Pedersen operations as is shown in Configure a cyclic group.

This example shows how to split a secret with a (t,n)(t, n)-threshold scheme of (t,n)=((t, n) = ( schemeThreshold , schemeParts )).

import (
    "github.com/matteoarella/pedersen"
)

group := /* cyclic group */
schemeParts := 5
schemeThreshold := 3
secret := /* secret to split */

p, err := pedersen.NewPedersen(schemeParts, schemeThreshold, pedersen.CyclicGroup(group))
if err != nil {
    panic(err)
}

shares, err = p.Split(secret, nil)
if err != nil {
    panic(err)
}

/* ...
   send shares parts to the corresponding shareholder and broadcast the commitments
   ...
*/

Secret shares structure​

When a secret is splitted with pedersen.Split a pedersen.Shares object is returned. This object contains every information that the dealer has to transfer to the shareholders.

If the secret that has to be split is not representable in the cyclic group (this is the case if the secret is bigger than the order of the cyclic group), the secret is split into chunks, and each chunk is split into secret parts according to Pedersen verifiable secret sharing.

Abscissae vector​

Shares.Abscissae is the abscissae vector used for computing the ordinate values of the secret parts. There is one abscissa for each shareholder, so if shareholderIdx represents the index of one shareholder, Abscissae[shareholderIdx] is the abscissa related to that shareholder.

Secret parts​

Shares.Parts is the matrix of secret parts.

The first index of Shares.Parts represents the shareholder index, while the second index represents the chunk index (Parts[shareholderIdx][chunkIdx]).

Secret commitments​

Shares.Commitments is the matrix of commitments. The first index of Shares.Commitments represents the chunk index so Commitments[chunkIdx] is the vector of commitments related to the chunk with index chunkIdx.