Method generating threshold vault and implemented by computer
Decentralized threshold cryptography using Elliptic Curve Cryptography distributes private keys among multiple nodes, addressing security vulnerabilities in blockchain systems by eliminating central control and trust requirements, ensuring robust and resilient key management.
Patent Information
- Application Number
- JP2024150482
- Authority / Receiving Office
- JP · JP
- Patent Type
- Applications
- Current Assignee / Owner
- Priority Date
- 2018-07-26
- Filing Date
- 2024-09-02
- Publication Date
- 2024-11-21
- Estimated Expiration
- Not applicable · inactive patent
AI Technical Summary
Existing blockchain systems face security challenges due to the need for centralized control of private keys, which can lead to single points of failure and require trust in third parties, making them vulnerable to attacks and fraud.
A decentralized system using threshold cryptography and Elliptic Curve Cryptography (ECC) to distribute private keys among multiple nodes, ensuring no single node can generate the private key independently, with methods for secure key generation and signature creation without a central authority.
This approach enhances security by eliminating the need for trust in third parties, providing fault-tolerant key management, and preventing unauthorized access, while maintaining compatibility with existing protocols like Bitcoin, thus enhancing resilience against attacks.
Smart Images

Figure 2024163180000001_ABST
Abstract
Description
[Technical field]
[0001] This application relates generally to security of data and computer-based resources. More specifically, this application relates to Elliptic Curve Cryptography, Elliptic Curve Digital Signature Algorithm (ECDSA) applications, and Threshold Cryptography. For example, this application describes an example process for securely and privately cooperatively generating public-private key pairs such that a threshold number of nodes are allowed to generate digital signatures that comply with ECDSA requirements, while no node or third party can generate the private key itself. [Background technology]
[0002] In this document, the term "blockchain" is used to include all forms of electronic, computer-based distributed ledgers. These include consensus-based blockchain and transaction chain technologies, permissioned and un-permissioned ledgers, shared ledgers, and variations thereof. The most widely known application of blockchain technology is the Bitcoin ledger, although other blockchain implementations have been proposed and developed. Although Bitcoin may be referred to herein for convenience and for illustrative purposes only, it should be noted that the present invention is not limited to use with the Bitcoin blockchain, and alternative blockchain implementations and protocols are within the scope of this application.
[0003] A blockchain is a peer-to-peer electronic ledger, implemented as a computer-based decentralized system made up of blocks of transactions. Each transaction is a data structure that encodes the transfer of control of a digital asset between addresses in the blockchain system and contains at least one input and at least one output. Each block contains a hash of the previous block, such that blocks are chained together to form a permanent, immutable record of all transactions that have been written to the blockchain since its inception.
[0004] The concept of decentralisation is fundamental in blockchain systems. Decentralised systems have the advantage that, unlike distributed or centralised systems, there is no single point of failure. They therefore offer an improved level of security and resilience. This security is further enhanced by the use of known encryption techniques such as Elliptic Curve Cryptography and ECDSA.
[0005] Ownership and control over digital assets (which may represent currency, tangible assets, intellectual property, computing resources, or any other goods or services) in a blockchain system is controlled by public-private key pairs. An unspent transaction output (UTXO) address is akin to a public key, and control over the assets assigned to that UTXO is proven by use of the corresponding private key. In order to transfer digital assets by using a UTXO as an input to another transaction, the input must be digitally signed by the corresponding private key. Control over the security and confidentiality of private keys is therefore fundamental to the security and reliability of a blockchain. Summary of the Invention
[0006] Thus, there exists a need for a solution that further enhances the security of such systems. The present application provides, among other things, such an advantage.
[0007] The present application provides methods and systems as defined in the accompanying claims.
[0008] According to an aspect of the present application, there is provided a computer-implemented method of securely generating a threshold vault, the threshold vault having a public key address, a private key, and a threshold k, each node in a plurality of nodes having a respective first secret and a respective k-1 degree polynomial function with each of the nodes' respective first secrets set as free terms, the plurality of nodes including a first node having a first polynomial function. The method performed by the first node includes, for each other node in the plurality of nodes, setting a variable in the first polynomial function to a value associated with the other node, determining a result of each of the first polynomial functions for that value, and securely transmitting each result to the other node. The method includes: receiving respective second secrets from at least a threshold number of the other nodes, each of the respective second secrets being a result of a polynomial function of each of the other nodes on a first value associated with the first node; sending to the other nodes a respective elliptic curve point multiplication of each result; and receiving from at least the threshold number of the other nodes each of the second secret elliptic curve point multiplications shared by the other nodes with any of the plurality of nodes; forming and storing a first secret share of the secret key based on a sum of the respective second secrets received from the other nodes; determining a public key corresponding to the private key based on the sum of the received elliptic curve point multiplications of each second secret and using polynomial interpolation; Further includes:
[0009] In some implementations, each polynomial function is independently selected by its respective node, with each polynomial function having a different set of coefficients.
number
[0010] In some implementations, the method includes sending to the other nodes elliptic curve point multiplications of each coefficient of the first polynomial function. The method may further include receiving from at least the threshold number of the other nodes elliptic curve point multiplications of each coefficient of the polynomial function for each of them. The method may also include validating the second secrets received from the other nodes.
[0011] The first private key share is expressed by the formula:
number
[0012] In some implementations, the method includes modifying the first private key share through cooperative participation with the other nodes in a dealerless secret distribution of a joint zero secret share, whereby the first node receives a first zero secret share and modifies the first private key share by adding the first zero secret share to form and store a modified first private key share.
[0013] Determining the public key may include determining the polynomial function by determining coefficients of the polynomial function from a plurality of known values of the received elliptic curve point multiplications of each of the second secrets.
[0014] Determining the polynomial function may include performing an error correction algorithm.
[0015] Determining the polynomial function may include performing a Berlekamp-Welch decoding algorithm.
[0016] Determining the polynomial function comprises: defining an error locator polynomial function and a second polynomial function that is a product of the error locator polynomial function and the error locator polynomial function; determining coefficients of the second polynomial function and the error locator polynomial function from a plurality of known values of the received elliptic curve point multiplications of each of the second secrets; determining the polynomial function from the second polynomial function and the error locator polynomial function; may have
[0017] In another aspect, the present disclosure discloses a computer-implemented method for securely generating a digital signature for a message on a threshold vault, the threshold vault having a public key address, a private key, and a threshold, and each node in a plurality of nodes having a respective private key share, a temporary key share, a multiplication mask share, a first additive mask share, and a second additive mask share. The method, performed by a first node in the plurality of nodes, comprises: generating and sharing a double masked key share from the temporary key share masked by the multiplication mask share and by the first addition mask share, and an elliptic curve point multiplication of the multiplication mask share; receiving from at least a threshold number of other nodes their respective double masked key shares and elliptic curve point multiplications of their respective multiplication mask shares; determining a first signature component using polynomial interpolation on the double masked key shares and polynomial interpolation on the elliptic curve point multiplication of the multiplication mask shares; determining and publishing a second signature component share based on the message masked by the second additive mask share, the private key share, the first signature component, and the ephemeral key share; A second signature component is obtained from a second signature component share of at least the threshold number of each of the other nodes.
[0018] In some implementations, the method includes generating a private key share, a temporary key share, and a multiplication mask share for the first node using a dealerless secret distribution. The method further includes generating a first additive mask share and a second additive mask share for the first node using a joint zero secret share. In some implementations, generating the double masked key share includes: θ i =D k(i) α i +β i modn determining θ i is the double masked key share, and D k(i) is the ephemeral key share of the first node, and α i is the multiplication mask share of the first node, and β i is the first additive mask share of the first node, and n is the integer order of the elliptic curve.
[0019] In some implementations, the polynomial interpolation on the double masked key shares results in a multiplicatively masked ephemeral key, and the polynomial interpolation on the elliptic curve point multiplication of the multiplication mask shares results in an elliptic curve point multiplication of a multiplication mask, and determining the first signature component includes determining a product of an elliptic curve point multiplication of the multiplication mask and an inverse of the multiplicatively masked ephemeral key.
[0020] In some implementations, the second signature component share s i To determine s i =Dk(i) (e+d A(i) r)+c i modn D k(i) is the first node's ephemeral key share, e is the hash of the message, and d A(i) is the first node's private key share, r is the first signature component, and c i is the second additive mask share, and n is the integer order of the elliptic curve.
[0021] Additionally, in another implementation, the method includes determining the second signature component by polynomial interpolation on the second signature component shares of at least the threshold number of the other nodes. In some cases, the method includes appending the first signature component and the second signature component to a transaction including the message and submitting the transaction to a blockchain network.
[0022] In a further aspect, the present application provides a computer-implemented method of multi-party computing, comprising: distributing a plurality of shares of a secret among a plurality of first nodes, each said secret being a free term of a respective polynomial function, each said share being a respective value of said polynomial function associated with said first node, whereby said secret is accessible by polynomial interpolation of a threshold number of said shares and is inaccessible in the absence of said threshold number of shares; receiving, at a second node, respective shares of the computed secret from a plurality of said first nodes, each said first node computing its share of the computed secret by performing at least one predetermined computation on each of the plurality of shares of the secret assigned to said first node; determining, at the second node, the computed secret by polynomial interpolation of at least a threshold number of the shares of the computed secret; The method comprises:
[0023] This provides the advantage of allowing the computations to be performed in such a way that the first node cannot determine the individual secrets or the computed secrets.
[0024] For each polynomial function, the following formula
number
[0025] The method may further comprise sending, to a plurality of nodes, elliptic curve point multiplications of each coefficient of a plurality of the polynomial functions.
[0026] The method may further comprise sending to a plurality of nodes an elliptic curve point multiplication of at least one of the polynomial function values associated with each of the plurality of nodes.
[0027] The method may further comprise verifying consistency between an elliptic curve point multiplication of at least one of the polynomial function values associated with at least one of the nodes and a sum of elliptic curve point multiplications of the coefficients of the polynomial function.
[0028] This provides the advantage of allowing dishonest participants to be more quickly identified and excluded from contributing to the polynomial interpolation.
[0029] Determining the computed secret share comprises: μ i =x i y i modn determining x i is the share of the first secret, and y i is the share of the second secret, and n is the integer order of the elliptic curve.
[0030] Determining the computed secret share comprises: x i +y i modn determining x i is the share of the first secret, and y i is the share of the second secret, and n is the integer order of the elliptic curve.
[0031] Determining the computed secret share comprises: x i -1 modn determining x i -1 is the inverse of the first secret share modulo n, where n is the integer order of the elliptic curve.
[0032] The method may further comprise providing modified shares of the computed secret by adding zero secret shares, which are free terms of a polynomial function equal to zero, such that the computed secret is accessible by polynomial interpolation of a threshold number of the modified shares and is inaccessible in the absence of the threshold number of shares.
[0033] This has the advantage of adding extra safety to the process.
[0034] Determining the calculated secret may comprise determining a polynomial function corresponding to the calculated secret share by determining coefficients of the polynomial function from a plurality of known values of the calculated secret share.
[0035] Determining the polynomial function may include performing an error correction algorithm.
[0036] Determining the polynomial function may include performing a Berlekamp-Welch decoding algorithm.
[0037] Determining the polynomial function comprises: defining an error locator polynomial function and a second polynomial function which is the product of said polynomial function and said error locator polynomial function associated with said calculated secret shares; determining coefficients of the second polynomial function and the error locator polynomial function from a plurality of known values of the calculated secret share; determining the polynomial function associated with the calculated secret share from the second polynomial function and the error locator polynomial function; may have
[0038] In a further aspect, the present application provides a computer-implemented method of distributing shares of a private key of a public-private key cryptography system among a plurality of nodes, each of the plurality of nodes having a respective first secret, a respective first polynomial function with the respective first secret of the node set as a free term, and a respective first share of a private key, the private key being a sum of the first secrets and a free term of a second polynomial function, the private key being accessible by polynomial interpolation of at least a threshold number of the first shares and less than the threshold number of the first shares being inaccessible, the method being performed by a first node of the plurality of nodes, the method comprising: selecting each third polynomial function having a zero set as its respective free term; for each other node of the plurality of nodes, setting a variable in the third polynomial function to a value associated with that other node, determining a result of each of the third polynomial functions for that value, and securely transmitting each result to that other node; receiving from each other node a respective share of a respective second secret, said share being a result of a third polynomial function of each of the other nodes on a first value associated with the first node; forming and storing a second share of each of said secret keys, said second share being the sum of said respective first shares and the sum of said received share of said second secret; The method comprises:
[0039] This provides the advantage of allowing shares of the private key to be updated without having to update the private key itself, thereby allowing malicious or unresponsive participants to be excluded.
[0040] The method may further comprise sending to the other nodes respective elliptic curve point multiplications of each of the results, and sending to the other nodes elliptic curve point multiplications of coefficients of the third polynomial function associated with the first node.
[0041] The method may further include receiving from each other node an elliptic curve point multiplication of the result associated with that node, and receiving from each other node an elliptic curve point multiplication of coefficients of the third polynomial function associated with that node.
[0042] This provides the advantage of allowing malicious or unresponsive participants to be more easily identified.
[0043] The method may further comprise checking the consistency of the result with respect to at least one other node with coefficients of the third polynomial function with respect to that other node.
[0044] The method may further comprise generating a share of a digital signature based on the message and the second share of the private key.
[0045] The method may further include receiving, from at least a threshold number of nodes, respective shares of digital signatures based on the message and a second share of each of the private keys, and generating digital signatures by polynomial interpolation of at least the threshold number of shares of the digital signatures.
[0046] According to a further aspect of the present application, there is provided a computer implemented system for carrying out the above method.
[0047] These and other aspects of the present application will be apparent from and elucidated with reference to the embodiments described herein.
[0048] Embodiments of the present application will now be described, by way of example only, with reference to the accompanying drawings. [Brief description of the drawings]
[0049] [Figure 1] 1 represents parts of an example elliptic curve digital signature process. [Diagram 2] 1 illustrates a further portion of an example elliptic curve digital signature process. [Diagram 3] 1 illustrates a reconfiguration hierarchy of shares distributed according to an exemplary method. [Figure 4] 1 illustrates diagrammatically an example of the distribution of shares relevant to a threshold vault implementation. [Diagram 5] 1 illustrates, in flow chart form, one example of a method for receiving and distributing digital assets using a threshold vault. [Figure 6] An example method for generating a threshold vault is presented in flow chart form. [Figure 7] 1 illustrates, in flow chart form, an example method for generating a digital signature for distributing an asset from a threshold vault. [Figure 8] 1 illustrates a block diagram of a blockchain node. [Figure 9] 1 illustrates, in flow chart form, an example method for performing multi-party computation using a threshold vault according to the present invention. [Figure 10] 1 shows a Berlekamp-Welch decoder that performs polynomial interpolation in a manner embodying the present invention. DETAILED DESCRIPTION OF THE PREFERRED EMBODIMENTS
[0050] <Summary> The present invention provides novel and inventive techniques that can be used to enhance the security of computer systems and resources that need to be protected from unauthorized access. The following description provides implementations, use cases and examples that are relevant to systems related to cryptocurrencies, including Bitcoin. However, it is important to note that the present invention has broader applicability and may be used to secure other types of systems and computer-based resources, and the present invention is not limited in this respect.
[0051] Embodiments of the present invention provide improved security by allowing the addition of group-based threshold encryption combined with the ability for dealerless deployment. Embodiments also support non-interactive signing of messages and provide for the division of private keys into shares that can be distributed to individuals and groups. Furthermore, the present invention provides a solution to create a distributed key generation system that eliminates the need for any centralized control while minimizing any threat of fraud or attack. In applying DSA's threshold-based solution to ECDSA, the present invention provides a fully distributive signature system that mitigates any single point of failure.
[0052] When used in conjunction with Bitcoin or alternatives, it can be combined with a search scheme that relates CLTV and multisig wallets to an infinitely scalable and secure means of deploying cryptocurrencies. With a group and ring based system, the present invention can be used to implement blind signatures on issued transactions.
[0053] This application discloses a threshold-based, dealerless private key distribution system that is fully compatible with Bitcoin. The system builds on a group signature scheme that moves away from traditional individual signature systems deployed within Bitcoin wallets. When deployed, the system is scalable and robust, resistant to errors and malicious adversaries. The system supports deployment in both dealer and dealerless systems and highly flexible distribution combinations.
[0054] Individual parties, either as sole participants or in collaboration, can act as dealers distributing slices of their protected key slices across machines for security and resilience, or in groups for deployment based on roles and voting thresholds in access control lists.
[0055] There is no limit to how deep the slices can be divided. This complexity issue needs to be weighed against each individual deployment. In this way, the signature and transaction records can be hidden from all outside participants, even from within the group, with the extensions presented in subsequent documents, thus embodiments of the present invention introduce a degree of anonymity and plausible deniability into Bitcoin transactions while raising the standard of pseudonymous protection for users.
[0056] Ibrahim et al.
[2003] developed an initial robust threshold ECDSA scheme. The following protocol is a further extension of the elliptic curve formulation of threshold DSS introduced by Gennaro et al.
[1996] .
[0057] [Table 1] The use of group mathematics allows the construction of a verifiable secret sharing scheme (VSS) that extends the work of Shamir
[1979] and Feldman
[1987] and Pedersen
[1992] in secret hiding from RSA and DSA schemes, so that it can be used in ECC and ECDSA based signature systems, such as Bitcoin [Koblitz, 1998]. Embodiments of the invention are resistant to malicious adversaries, halting, and robust against eavesdropping.
[0058] This disclosure begins by presenting a method that allows for co-signing of ECDSA signatures when neither party knows the private key. Moreover, it allows for private key pairs to be updated and refreshed without the need to change the private key, which is a significant technical advantage.
[0059] All existing solutions require a trusted party. Utilizing the present invention also makes it possible to extend the work of Chaum
[1983] from a centralized system to a truly decentralized method of issuing electronic nodes that can be solved directly on the Bitcoin blockchain, obsoleting the requirement for an alternative blockchain or sidechain.
[0060] For the above generator G, Bitcoin uses secp256k1, which defines the parameters of the ECDSA curve used in Bitcoin and can be referenced from the Standards for Efficient Cryptography (SEC) (Certicom Research, http: / / www.secg.org / sec2-v2.pdf).
[0061] A question of trust All existing systems require some degree of trust. Until now, Bitcoin has required the protection of private keys using a secure system isolated from the world, which has proven difficult to achieve. Notably, the systems in which Bitcoins may be exchanged or stored require trust in a centralised authority. The present invention completely changes this requirement, decentralising and decentralising the key generation and message signing process within Bitcoin without changing any of the core requirements of the protocol. The methods mentioned herein can be implemented without modifying the Bitcoin protocol, and in fact there is no way to determine whether this process has been deployed through analysis of signed messages.
[0062] In building a distributed signature scheme for Bitcoin, the present invention allows a group of people or systems to securely hold keys such that no individual can generate signatures on their own. When extended, the scheme also allows for the secure recovery of Bitcoin's private keys themselves and each of its shares. Signatures generated by a group are indistinguishable from those generated from existing protocols. As such, signature verification remains as if it had been established by a single signer using a standard transaction.
[0063] This increased reliability is achieved by having a private key shared by a group of n participants or a group of m participants. A threshold number of participants is required to sign a transaction, and any federation of participants or groups of participants that meet the minimum threshold can perform the signing operation. Importantly, the protocol can be established as a batch process or synchronously, where individuals or groups can attempt to create a federation of participants.
[0064] <Background research> First, Shamir
[1979] introduced a dealer-based secret sharing scheme that allows for distributed management of keys. The problems with this scheme stem from the necessity to trust an unverifiable dealer. Schemes of this type are fully compatible with the present invention and can be used for group distribution of individual key slices generated through the process described herein.
[0065] <Joint Random Secret Sharing (JRSS) [Pedersen, 1992]> The stated goal of this procedure is to construct a way in which a group of participants can collaboratively share a secret without any of them knowing the secret. Each participant chooses a random value as their local secret and distributes values derived from this random value using Shamir's secret sharing scheme (SSSS) among the group. Each participant then adds together all the shares received from participants, including itself. This sum is the joint random secret share. The randomness provided by one honest participant is sufficient to keep the combined secret value secret. This condition still holds even if all the (n-1) other participants intentionally choose non-random secret values.
[0066] <Joint Zero Secret Sharing (JZSS) [Ben-Or, 1988]> Joint Zero Secret Sharing (JZSS) is similar to Joint Random Secret Sharing (JRSS), except that each participant shares zeros instead of random values. The shares generated by this technique help to eliminate any potential weaknesses in the JRSS algorithm.
[0067] Desmedt
[1987] introduced the concept of group-oriented encryption. This process allows participants to send messages to a group of people in such a way that only a select subset of participants can decrypt the message. In the system, members are known if the sender needs to know them using a public key, and a group is anonymous if there is only a single public key for the group that is kept independent of the members. The present invention integrates both methods, allowing known and anonymous senders and signers to exist simultaneously in a group.
[0068] <The present invention> Any elliptic curve with large primes (CURVE), and the prime field Z p A base point G∈CURVE(Z p ), a system can be constructed that allows for the secure distribution and use of ECC private keys to key shares such that no participant can recover the original private key from less than a threshold number of shares.
[0069] unknown integer d A For 1≦d A If ≦(n-1), then Q A =d A Assuming ×G, d A It is known to be extremely difficult to calculate [Kapoor, 2008].
[0070] The basic technique of the present invention is obtained by applying threshold cryptography, in which the ECDSA private key exists only implicitly and does not need to be reproduced in any system. Each of these multiple shares is shared by multiple participants [p (i) ]. Thus, the signature process differs from the one deployed in Bitcoin. In this process, the coordinating participants p (c)generates transaction and message signatures that are distributed to the group. Each participant can vote on the use of its private key share by computing or passing a partial signature.
[0071] In fact, a pass is equivalent to a no vote. (c) verifies the responses and, if a minimum threshold number of partial signatures are received, combines them to form a complete signature.
[0072] Coordination participant p (c) can either accept negative votes and perform calculations based on null values from other parties, or lobby parties to convince them to sign messages. The protocol can be implemented with a configured coordinator, or any individual or group can form this role and propose transactions to a threshold group to be signed. The present invention extends the work of Ibrahim et al.
[2003] to provide a fully distributed ECDSA private key generation algorithm. This application also presents a distributed key redistribution algorithm and a distributed ECDSA signature algorithm for use with Bitcoin. The key redistribution algorithm may be used to invalidate all currently existing private key shares in favor of new ones, or to reallocate private key shares to new participants. The protocol extends to the distribution of private key shares as well as ECDSA private keys. This result means that shares can be constructed and voted on as a group process.
[0073] The present invention removes all requirements for the existence of a trusted third party. As a result, it is possible to build new overlays and wallets for Bitcoin that are fully compatible with existing protocols and eliminate all remaining single points of failure while allowing further scalability. The present invention can also be extended to allow the introduction of blind signatures.
[0074] The present invention not only removes the need for a trusted third party, since private keys do not need to be loaded into memory, but also eliminates a wide range of common attacks. The protocol is extensible to allow the number of shares required and the distribution of shares to be determined by the use case, economic scenario and risk requirements.
[0075] The present invention mitigates all side-channel attacks, and therefore any cache timing attacks. The system incorporates the work of Gennaro et al.
[1996] and extends it from DSS so that it can be used successfully with any ECDSA-based application.
[0076] <ecdsa> Bitcoin uses ECDSA based on the secp256k1 curve. ECDSA was initially standardized by NIST in 2003 to modify the requirements for key exchange based on Diffie-Hellman using elliptic curve Cryptography (ECC) [NIST]. The generation of ECC was particularly important due to the reduced key size and processing power when compared to other public / private key systems. No sub-exponential time algorithms have been found for ECDLP. ECDLP is known to be intractable and refers to the elliptic curve discrete logarithm problem [Johnson, 2001].
[0077] The parameters used throughout this application are provided in Table 1 above.
[0078] <Safety concerns> The system is limited by the security of ECDSA, which is currently limited in bits. Currently, ECDSA remains secure if the private key can be safely unfolded. The present invention mitigates side channel and memory disclosure attacks up to a threshold that requires a threshold number of participants to be compromised before a rekey event. Furthermore, a majority of uncompromised participants can identify compromised participants below the threshold.
[0079] Halting problem Service disruption is a type of attack that can be undertaken by a malicious adversary attempting to launch a denial of service attack against a participant, forcing the participant to receive invalid signatures that require processing time to analyze, or a large number of network messages that are then dropped.
[0080] The requirement to encrypt messages to participants using either ECC or ECC based on signcryption mitigates such an attack vector. Since an attacker would need to have already compromised the participants in order to send a message that is an invalid, partially signed message, such an attack is no longer necessary.
[0081] Randomness Algorithm 2 provides a scenario where sufficient randomness is introduced even if (n-1) participants are unable to choose random values. What can be added to this protocol is the introduction of a group oracle that is specifically designed for the introduction of random values into the signing and key regeneration processes. In any such scenario, each of the key slices can be generated using the same protocol. For example, if there is an m of n primary slice requirement, each of the underlying key slices can also be generated and managed using an m’ of n’ threshold condition.
[0082] Participants using this system can add an external oracle that does nothing other than inject randomness into the protocol. Users with m’ key slices (where m’ < n-1) can choose to reproduce and process their signature solution based on the key slices they hold, or they can introduce an external oracle that is unnecessary except for the introduction of randomness.
[0083] Each slice can be similarly divided for robustness and security. The key slices can be distributed such that a user has slices in an external device such as a mobile phone or smart card and in a software program executed on a computer, such that a combination of sources is required for them to generate a partial signature.
[0084] A unique random ephemeral key D k is generated, or the secret key d A It is important that the information can be used to reconstruct the
[0085] <Public Signature> The main purpose of transaction signing in this protocol is to allow distributed signing of Bitcoin transactions. Any transaction that has never been published on the blockchain can be kept private by the participants. Thus, the coordinating participant p (c) On each occasion that a single Bitcoin user fails to achieve the required level of votes to successfully sign a transaction, there is no need to generate a new Bitcoin transaction. If the key slice is itself secure against the threshold, the ownership of any settled transaction remains secure.
[0086] If the system is properly deployed, the ability to compromise up to (k-1) participants leaves the system secure against sub-threshold attacks. When combined with a periodic rekeying protocol (Algorithm 2), embodiments of the present invention can withstand side-channel attacks and memory disclosure.
[0087] Methods and Practice of the Invention Because embodiments of the present invention encrypt any secret information that needs to be sent between participants using ECC based on hierarchical derivation [Wright, 2016], it is possible and desirable to package all messages into a single packet that is sent to all users, so that validation can be performed against potentially compromised or hostile participants, if necessary.
[0088] Signature generation is performed by coordinating participant p (c) By default, any key slice can act as a coordinating participant, and the requirements come down to the individual implementation of the protocol. The algorithms used are described below, and later sections provide details regarding their deployment.
[0089] <Algorithm 1: Key generation> Domain parameters (CURVE, cardinality n, generator G) Input: NA Output: Public key Q A Private key share A(1) ,d A(2) , ,d A(j) For a threshold of k slices from (j) participants, a key segment d A(i) is constructed by associating participant (i) with (j-1) participants, designated as participants (h), who are the remaining parties with which participant (i) exchanges secrets in order to sign keys (and thus Bitcoin transactions).
[0090] In the scheme, j is the total number of participants, where k≦j, and therefore h=j-1. ·Therefore, there exists a (k,j)-threshold distribution scheme.
[0091] The method for Algorithm 1 is as follows: 1) For 1≦i≦j, for each participant p in (j), (i) Exchanges ECC public keys (i.e., in this implementation, a Bitcoin address) with all other participants. This address is the group's identifying address and does not need to be used for any other purposes.
[0092] It should be noted that this is an address and key derived based on shared values between each of the participants from the process of International Patent Application WO2017 / 145016.
[0093] 2) Each participant p (i) is a polynomial f of degree (k-1) with random coefficients that are secret from all other parties. i Select (x).
[0094] This function is the participant's secret a0, which is chosen as a polynomial free term. (i) The first secret value is of the form:
[0095] f i (h) is the value at point (x=h) of participant p (i) The function f selected by (x) is defined as the result of participant p (i) The basic formula is the function:
number
[0096] In this formula, a0 is the (i) are confidential and will not be shared.
[0097] Therefore, each participant p (i) teeth:
number
[0098] 3) Each participant p (i) As described above, as disclosed in WO2017 / 145016, P (h) Participant P uses the public key of (h) The first share f to ∀h={1, ,(i-1),(i+1), ,j} i (h) is encrypted and P (h) exchanges values to be decrypted.
[0099] Any base point G∈E(Z p It should be noted that for all n×G, n×G=φ. In the case of Bitcoin, the values are: Elliptic curve formula: y 2 =x 3 +7 Prime numbers modulo: 2 256 -2 32 -2 9 -2 8 -2 7 -2 6 -2 4 -1 =FFFFFFFF FFFFFFFF FFFFFFFF FFFFFFFF FFFFFFFF FFFFFFFF FFFFFFFE FFFFFC2F Base point=04 79BE667E F9DCBBAC 55A06295 CE870B07 029BFCDB 2DCE28D9 59F2815B 16F81798 483ADA77 26A3C465 5DA4FBFC 0E1108A8 FD17B448 A6855419 9C47D08F FB10D4B8 Order=FFFFFFFF FFFFFFFF FFFFFFFF FFFFFFFE BAAEDCE6 AF48A03B BFD25E8C D0364141 It is.
[0100] As such, for any set B of integers expressible as (b, b1, b2, . . .), i ∈Z n }, if bG=[b1G+b2G+···]modp then b=[b1+b2+···]modn. Furthermore, if bG=[b1b2···]Gmodp then b=[b1b2···]modn.
[0101] Z n If Z is a field and Lagrange interpolation modulo n can be usefully performed on the value chosen for the ECC private key, then Shamir's Secret Sharing Scheme (SSSS) [5] can be used to n There are conditions that lead to the conclusion that it can be implemented.
[0102] 4) Each participant p (i) is the following value
number
[0103] The value associated with the variable h in the above formula is (h) If P corresponds to the third participant in the scheme, then h=3. (h) or may equally represent the values of the ECC public keys used by the participants as integers. Use cases and scenarios exist for either implementation. In the latter implementation, the value h={1,...,j} is replaced by an array of values that map to the public keys used by the individual participants.
[0104] 5) Each participant P (h≠i) verifies the consistency of the received share with that received from each other participant, i.e.:
number
[0105] 6) Each participant P (h≠i) The participant (P (h≠i) ) matches other received shares:
number
[0106] 7) At this time, participant p (i) each has their share d A(i) of,
number
number
[0107] At most (k-1) values are null, and the remaining values are G×α i Let the form be {ω1, ,ω j }(j≧(2k-1)) is set, and each α i exists in some (k-1)th degree polynomial H(·), then θ=G×H(0).
[0108] This value is θ=Σ i∈V ω i ×λ i =Σ i∈V (G×H(i))×λ i β. In this case, V can be calculated by the correct ω i is a (k)-subset of values, and further, λ i represent the resulting Lagrangian interpolation coefficients. The polynomials can be computed by using a Berlekamp-Welch decoder. Return (d A(i) ,Q A ) At this time, d A(i) is the third share of the third secret value.
[0109] participant p (i) Then, p uses its share in computing the signature. This role is played by any participant or party p who acts as a coordinator in the process of collecting signatures. (c) This can be done by the participant p (c) may vary and need not be the same party in each attempt to collect enough shares to sign a transaction.
[0110] Therefore, the secret key share d A(i) ←Z n * was generated without knowledge of the other participants' shares.
[0111] <Algorithm 2: Private key update> Input: d A(i) The private key d is expressed as A Participant P i Share Output: Participant P i New private key share d A(i) Algorithm 2 can be used both to update the secret key and to add randomness to the protocol.
[0112] Using keys in the format of [Wright, 2016], this process can result in the recomputation of hierarchical subkeys without the reconstruction or even the computed existence of the private key. In this way, it is possible to construct a hierarchy of Bitcoin addresses and private key slices that, if deployed correctly, precludes any large-scale fraud or database theft that has occurred in the past.
[0113] 1) Each participant chooses a random polynomial of degree (k-1) with zero as its free term. This is similar to Algorithm 1, but participants must validate that all other participants' chosen secrets are zero.
[0114] It should be noted that φG=nG=0, where 0 is the point at infinity on the elliptic curve.
[0115] Using this formula, all active participants are calculated using the function:
number
[0116] Generate zero share: z i ←Z n * 2)d A(i) ′=d A(i) +z i 3)Return:D A(i) ' The result of this algorithm is a new key share associated with the original private key. Variations of this algorithm can engage in redistribution exercises resulting in new key slices without the need to change potential Bitcoin addresses, or increase the randomness of the first algorithm. In this way, the present invention allows a group to additionally mask private key shares without changing the underlying private key. This process can be used to minimize any potential key leakage associated with the continued use and deployment of individual key shares without changing the underlying Bitcoin addresses and private keys.
[0117] <Algorithm 3: Signature Generation> Domain parameters: CURVE, cardinality n, generator G Input: Message to be signed e=H(m) Private key share d A(i) ∈Z n * Output: Signature e=H(m) for (r,s)∈Z n * A) Distributed key generation 1) Generate ephemeral key shares using Algorithm 1: D k(i) ←Z n * 2) Generate mask shares using Algorithm 1: α i ←Z n 3) Generate mask shares using Algorithm 2: β i ,c i ←Z n 2 By running Algorithm 2 twice using a polynomial of degree 2(k-1), the shares generated by these protocols are
number
[0118] The shares of β and c are then kept secret by the participants.
[0119] B) Signature generation 4) e=H(m) Validate the hash of message m.
[0120] 5) θ i =D k(i) α i +β i Broadcast modn and ωi = G × αi.
[0121] 6)
number
[0122] Let {θ1, ,θ(.) be the set of all the residuals such that most (k-1) are null and all the residuals are in the (k-1) degree polynomial F(.). n }(j ≥ (2k-1)) form a set, then μ = F(0).
[0123] 7)
number
number
[0124] 10)s i =D k(i) (e+d A(i) r)+c i Broadcast modn.
[0125] 11) S = Interpolate(s i , ,s n )modn If s=0, start Algorithm 3 over again (A.1).
[0126] 12) Return (r,s) 13) In Bitcoin, the (r,s) pair is reconstructed to form a standard transaction.
[0127] FIG. 10 illustrates a novel use of a conventional Berlekamp-Welch decoder 70 to obtain a polynomial function that represents a partial signature of a digital message.
[0128] In a conventional use of the Berlekamp-Welch algorithm to correct errors in transmitted data, a message m is divided by an encoder 72 into k consecutive bytes, each of which is denoted c0, c1, . . . , c k-1 is encoded as an integer modulo p. The message is then encoded as a polynomial function:
number
[0129] The data M received at the receiver 76 (i.e., the received message) is represented as pairs (a1, b1, . . . , a n ,b n )
number
[0130] (a i ,b i For n known values of E(a i ) is a polynomial function of degree e, and P(a i ) is a polynomial function of degree (k-1), so Q(a i ) is a polynomial function of degree (e+k-1). Therefore, (a i ,b i ) can be expressed as a linear system:
number
[0131] Thus, the Berlekamp-Welch decoder 70 takes as input pairs representing points on a polynomial function and outputs the polynomial function, and thus may be used in the present invention as an alternative to Lagrange interpolation to determine a polynomial from a threshold number of shares represented by that polynomial.
[0132] <Model - Threshold ECDSA(T,ECDSA)> According to an embodiment of the present invention, a system of n groups or individuals designated as participants is permitted. Each player can be an individual, as a single participant or group, or a combination thereof. Participant p (i) may be mapped to an identity using a public key computation derived in general, or (i) may remain as a pseudonymous entity with the participant's public key used only for this protocol, without being mapped back to an individual.
[0133] The present invention introduces a dedicated broadcast channel that allows members in a group to remain unidentified while still allowing recognition of other participants as valid players and members of the scheme. (i) When a key is broadcast from a group, members of the group will recognize the message as coming from an authorized party, without necessarily being able to identify the end user or individual associated with the key, and it may even be possible to link the key's identifier to an individual, if such a system is warranted.
[0134] The process flow is summarized as follows:
[0135] In Figure 1: Step 1) Please refer to Figure 1.
[0136] Step 2) A Pc sends a raw transaction to the group, if it validates (i.e. the raw transaction matches the signed hash) participants vote by signing it.
[0137] Step 3) If the answer is yes, each participant returns a partially signed transaction.
[0138] Please refer to Figure 2.
[0139] Step 4) When a threshold number of partially signed transactions are received, Pc (or any other participant) reconstructs the complete signature.
[0140] Step 5) The PC broadcasts the transaction as a signed Bitcoin transaction.
[0141] The computation of the message signature can be initiated by a stationary individual or through a temporary broadcast party. The role of the protocol coordinator can be played by any participant or by a party p(c) that becomes the coordinator in the process of collecting signatures.
[0142] <Key generation> A modified ECDSA key generation algorithm is used to create a fully distributed signature scheme, in which private keys are chosen jointly by the distributed group using a combination of a hidden random secret.
[0143] The threshold key derivation algorithm is given in Algorithm 1.
[0144] The algorithm is scalable and each step of the algorithm can be performed simultaneously by any participant without a dealer, or in a group or individual or dealer. This implementation is fully compatible with the current Bitcoin protocol. Any signer appears to an outside observer or verifier as if they signed in a standard way. As a result, there is no way to tell if a key was generated in a standard format or using the advanced protocol of the present invention.
[0145] <Signature generation> The concept of threshold signature generation is described in [Shamir, 1979]. Algorithm 3 is related to a procedure reported in [Feldman, 1987] that is based on a DH-based system and modified to allow for ECDSA.
[0146] The present invention extends this process to be fully compatible with both processing and signing Bitcoin transactions. It also spans multisig transactions, allowing for distributed keys to be derived for each of the multiple signatures required.
[0147] <Redistribution of private keys> This process can be extended to introduce a fully decentralized key redistribution scheme. The redistribution is completed when the current participants run Algorithm 2 once and add the resulting zero shares to their private key shares. New shares are distributed randomly if one participant introduces a random value.
[0148] This process allows for additive masking of the private key shares without changing the actual private key.
[0149] <Derivation of threshold ECDSA signature> A threshold ECDSA signature generation system is derived using ideas for a threshold DSS signature generation protocol found in [Feldman, 1987] following the scheme developed in [Shamir, 1979].
[0150] <Verification> The present invention allows for offline signing and verification of messages before any value is transferred to a known Bitcoin address. Each party can independently calculate and validate the address using the process mentioned in Algorithm 1. Thus, all participants can know that their shares are valid before any execution that seeks to fund the Bitcoin address. In this process, validation schemes are possible, but they are not necessary. Any threshold participant who chooses to send an invalid signature slice is in effect casting a negative vote; that is, a vote to not sign the message and therefore not complete the transaction in Bitcoin is achieved through inactivity. The effect is as if they had not signed the message at all.
[0151] Algorithm 2 provides a way by which participants can verify the consistency of their shares. If a threshold number of non-malicious participants is kept, it is possible to exclude any known malicious participants during rekeying. Thus, key slices can be updated without allocating new slices to known malicious participants, allowing key refreshing in a way that still allows for slice reallocation.
[0152] When completing secure multiparty computations in environments where untrusted and malicious adversaries are to be expected as standard, it is possible to further enhance the robustness of the verification process, increasing its ability to defend against j / 2 passive and j / 3 active adversaries [Ben-Or, 1989; Rabin, 1988]. The robustness of the system can be enhanced using additional processes: 1. Let D be a polynomial A(x) of degree (k-1) shared among j participants. a Let us assume that. 2. Individual participants (i) D is made available to the group a Share D a(i) and D. a(i) We have G∀(i∈0, ,j). 3. All participants then share secret b by using Algorithm 2, so that each participant p (i) is a polynomial of degree (k-1) D b New hidden share of D b(i) has.
number
number
[0153] The system allows any participant to determine whether another participant is behaving maliciously.
[0154] <Distributed key generation> This scheme makes it possible to complete the implementation of both distributed autonomous corporations (DAC) and distributed autonomous social organisations (DASO) in a secure manner. It is shown that any k members can represent such a group through an identification scheme (including by means of a digital certificate signed and issued by a certificate authority), and that any k members can compose digital signatures on behalf of the organisation. This system is extended to Bitcoin transaction signatures providing value transfer and verification without any identification function. These authentication schemes are proven to be secure.
[0155] <Method and Implementation> The protocol is based on the technology disclosed in International Patent Application WO2017 / 145016, and uses ECC to encrypt any secret information that needs to be sent between participants, so that it is possible and desirable to package all messages into a single packet that is sent to all users, so that validation can be performed, if necessary, against potentially compromised or hostile participants.
[0156] Signature generation is performed by coordinating participant p (c) By default, any key slice can act as a coordinating participant, and the requirements come down to individual implementations of the protocol. (c) Upon generation of a valid raw transaction by p, the transaction and its message hash are encrypted and shared among all participants p (i≠c) will be broadcast to
[0157] A. Ephemeral Key Share D k(i) Generate.
[0158] Participants use Algorithm 1 to calculate Z by a polynomial of degree (k-1). n * A temporary key D uniformly distributed in k Generate and share
number
[0159] D k The shares of each participant are kept private and confidential.
[0160] B. Mask Share α i Generate.
[0161] Each participant will share
number
[0162] α i The shares are confidential and are kept by the corresponding participants.
[0163] C. Mask Share β i ,c i Generate.
[0164] Run Algorithm 2 twice using a polynomial of degree 2(k-1).
[0165] The shares generated by these protocols
number
[0166] The shares of β and c should be kept secret by the participants.
[0167] D. Calculate the digest of message m:e = H(m).
[0168] This value is p (c) The received hash of the transaction is checked against the received hash of the transaction obtained from
[0169] E.v i =D k(i) α i +β i modn and ω i = G × α i Broadcast.
[0170] Participant P i is ν i =D k(i) α i +β i modn and ω i = G × α i Broadcast.
[0171] P i If no response is received from , the value used is set to null.
[0172]
number
[0173] F.μ=Interpolate(ν1,...,ν j ) mod n.
[0174] Interpolate()[2]: Here, {ν1, ,ν such that most (k-1) are null and all the residuals are in the (k-1) degree polynomial F(.) n}(j ≥ (2k-1)) form a set, then μ = F(0).
[0175] The polynomials can be calculated using general polynomial interpolation. The function "Interpolate()" is the Berlekamp-Welch interpolation [2], defined as an error-correcting algorithm for BCH and Reed-Solomon codes. 2 This is described in more detail at http: / / mathworld.wolfram.com / LagrangeInterpolatingPolynomial.html and in Whittaker, E. T. and Robinson, G. "Lagrange's Formula of Interpolation." §17 in The Calculus of Observations: A Treatise on Numerical Mathematics, 4th ed. New York: Dover, pp.28-30, 1967, and at https: / / jeremykun.com / 2015 / 09 / 07 / welch-berlekamp / .
[0176] G.θ=Exp-Interpolate(ω1,...,ω j ) to calculate
[0177] Exp-Interpolate()
[10] : At most (k-1) values are null, and the remaining values are G×α i Let the form be {ω1, ,ω j }(j≧(2k-1)) is set, and each α i exists in some (k-1)th degree polynomial H(·), then θ=G×H(0).
[0178] This value is θ=Σ i∈V ω i ×λ i =Σ i∈V (G×H(i))×λ i β. In this case, V can be calculated by the correct ω i is a (k)-subset of values, and further, λ i represent the resulting Lagrangian interpolation coefficients. The polynomials can be computed by using a Berlekamp-Welch decoder.
[0179] H.(R x ,R y )=θ×μ -1 Calculate.
[0180] Ir=R x Assign modq. If r=0, go to step A.
[0181] Each participant p (i) is step J i Calculate their slices of the adjuster p (c) can use those values to reconstruct s if it receives a threshold number of responses.
[0182] Js i =D k(i) (e+D A(i) r)+c i Broadcast modn.
[0183] The response is P i If not solicited / received from, the value used is set to null.
[0184]
number
[0185] Ks = Interpolate(s1, ,s n ) mod n.
[0186] If s=0, go to step I.
[0187] where the function Interpolate() is as defined above.
[0188] Each participant p (i) is step J i Calculate their slices of the adjuster p (c) is the threshold s i When responses are received, their values can be used to reconstruct s.
[0189] Return L.(r,s).
[0190] M. Replace the signature section of the raw transaction and broadcast it to the network.
[0191] <Dealer's share of slices> The above embodiment can be made much more flexible through the introduction of group shares: in this way, share allocation can be divided at any level of hierarchical depth among a dealer, multiple dealers, a group that does not include a dealer, or any possible combination thereof.
[0192] Value d A and its corresponding key slice d A(i) By replacing with values derived using the same algorithm, a voting hierarchy can be generated. For example, the scheme is: 1) Dealer-based allocation 2) Multiple dealers 3) No dealer The shares derived from the above can be simultaneously constructed.
[0193] The scheme is therefore extensible and can be made to incorporate any business structure or organisational system.
[0194] The allocation of slices is also extensible. Deploying a non-uniform allocation process allows weighting of shares to be added. In the scheme shown in Figure 3, it is possible to build a virtual organization with 5 top members. Note that this does not require setting an equally weighted share value of n=5. In this virtual organization, it is possible to set the voting structure for the top schema as follows: ·Threshold(0) 61 shares D L02 15 Shares D L1 15 Shares D L2 15 Shares ·D145 Share ·D210 Share Here, n=100 is set. As stated, this is an arbitrary value that can reflect any organizational structure. The organization in Figure 3 allows for veto scenarios (D1) and, through the introduction of tiered allocation, allows for any imaginable voting structure.
[0195] What is often overlooked in multi-level sub-tier structures is that even if secret slices are allocated, they do not have to be equally distributed, and furthermore, subgroup ownership does not have to reflect ownership at other levels. In Figure 3, there is a seemingly powerful block that controls 45% of the total number of shares at a threshold of 75%. If sub-level allocation of shares is considered in the case, the scenario becomes much more complicated. It is possible to create cross-ownership with individuals holding voting shares at multiple levels and positions on the table.
[0196] The distribution in Table 3 is defined as (retained shares, threshold, allocation {n}).
[0197] From Table 3, participants P1 and P2 each dominate the vote, but their coalition with participant P4 provides a sufficient voting block for either P1 or P2, unless P1 or P2 vetoes the vote.
[0198] Since there are no limitations to the implementation and structure of the voting format in the present invention, it can be used to build any organizational hierarchy imaginable while ensuring secure backup and recovery methods.
[0199] [Table 2] As a result, it is possible to have veto and voting rights assigned to higher level signature shares. In this example, the ownership of a share in S is L02 , D 1A and D. 2A can be held at
[0200] <Secure Multi-Party Computation> n participants p (1) ,···,p (i) ,···,p (n) Secure multiparty function computation by x (i) A function F(x1, ,x i ,···,x n This is a question based on the need to evaluate (i) p (i) A secret value provided by a participant p (j≠i) Or an outside party x (i) must be kept secret so that no one can know the value of the sigma-based algorithm. Thus, the goal is to preserve the secrecy of each participant's value while being able to guarantee the accuracy of the computation.
[0201] In this scenario, a trusted third party T can communicate with various participants p (i:1···n) From all values of x (i:1···n) and returns the computation. This design only works in an ideal world where we can implicitly trust T. If there is any possibility that T is malicious, rogue, or otherwise compromised, then the use of a trusted third party becomes infeasible. This scenario mirrors existing elections where participants are voters and the trusted third party is conducted by the government.
[0202] Any value that can be computed in a secure manner using a trusted third party can be used to calculate the individual secret x (i) It has been shown that a secure protocol can be computed without a trusted party while still retaining the security of [Bar-Ilan, 1989]. The protocol presented here is secure for individual computations and provides secure computation even when non-threshold groups of compromised participants are able to cooperate.
[0203] <Simple multiplication> n participants p (i:1···n) If there are two secret values x and y that are distributed among participants p(i), then we can maintain the secrecy of both input variables x and y, and at the same time, each participant p(i) can maintain the secrecy of their individual secret x (i:1···n) and y (i:1···n) It is possible to compute the product xy while ensuring that holds.
[0204] In this scheme, x and y are each shared among a threshold group of participants using a polynomial of degree (k-1). (i) is the degree (k-1) polynomial of x (i:1···n) and y in the polynomial of degree (k-1) for y (i:1···n) Their shares of can be multiplied.
[0205] Introducing Algorithm 2 means that participant p (i) In this case, we use the polynomial z of degree (2k-1). (i) This value returns the share of each participant p (i) is the value x (i) y (i) +z (i) Calculate.
[0206] x (i) y (i) +z (i) The returned values of represent the effective shares of the computation of x ⋅ y for a polynomial of degree (2k-1). Any participant acting for a threshold number of shares or a coordinator can use the returned values held by each participant to compute the true value of x ⋅ y without gaining any knowledge of the individual shares.
[0207] <Simple addition> n participants p (i:1···n) If there are two secret values x and y that are distributed among participants p(i), then we can maintain the secrecy of both input variables x and y, and at the same time, each participant p(i) can maintain the secrecy of their individual secret x (i:1···n) and y (i:1···n) It is possible to compute the sum x+y while ensuring that holds.
[0208] As in the process for simple multiplication, each participant p (i) is the value x (i) +y (i) +z (i) Calculate z (i) is not necessary, but adds an extra level of randomness and secrecy to the process.
[0209] x (i) +y (i) +z (i) The returned values of represent the effective shares of the calculation of x+y for a polynomial of degree (2k-1). Any participant acting for a threshold number of shares or a coordinator can use the returned values held by each participant to calculate the true value of x+y without gaining any knowledge of the individual shares.
[0210] If the participants are not too hostile, this can be done without any additional steps. (i) +y (i) It can be simplified as addition.
[0211] <inverse or reciprocal> The distributed secret value, i.e., x among j participants (i:1···j) For x mod n, which is secretly distributed as (i) , x or x -1 without disclosing any information disclosing -1 It is possible to generate a polynomial share related to the value of modn [Gennaro, 1996]. As before, for each participant p (i) is a polynomial of degree (k-1) x (i) holds the share of the value x represented by
[0212] Using Algorithm 1, each participant calculates a share x of the unknown secret x y in a polynomial of degree (k-1). (i) Then, each participant executes Algorithm 2 to compute the zero-secret (k-1) with a polynomial of degree (2-1). Each participant (2-1) generates the value x (i) y (i) +z (i) Do the calculations to calculate.
[0213] Using the Interpolate() routine above, each participant calculates μ=x (i) y (i) +z (i) Calculate the value of μ i Then, each participant can return a value μ from the collected values of μ -1 The value of modn can be calculated.
[0214] These values are given by p (i) is a polynomial of degree (2k-1) and ζ i =γ i μ -1 Using x i -1 It is sufficient to be able to compute the relevant share of . The Berlekamp-Welch decoding scheme [Berlekamp, 1968] provides one of several methods that may be used to complete this process.
[0215] Figure 9 shows the quantity ζ=x -1 9 illustrates, in flow chart form, an example of a method for performing secure multi-party computation of a secret x, y, z1, where the secure multi-party computation is performed by multiple first nodes and the computed secret is recovered by a second node. First, in step 901, shares x, y, and z1 of a secret are i , y i , x 1i is distributed among the first nodes by the second nodes, where x, y, and z1 are the free terms of each polynomial function, and z1 is zero.
[0216] In step 902, each first node i =x i y i +z i By polynomial interpolation, the product μ=xy is calculated in step 903 as the threshold share μ i In step 904, the inverse μ -1 modn=x -1 y -1 modn is then determined. In step 905, the secret w, the share w of z2 i , z 2i is then distributed among the first nodes by the second nodes, where w and z2 are the free terms of each of the polynomial functions and z2 is zero. In step 906, each first node distributes its share of the common secret, ζ i Here, ζ i =y i μ -1 +w i +z 2i =y i x -1 y -1 +w i +z 2i The second node then determines in step 907 the threshold share ζ i from the first node and performs polynomial interpolation in step 908 to obtain the calculated secret ζ such that the individual secrets x, y, or w are not available to any of the first nodes. i =x -1 Determine +wmodn.
[0217] <Assignment> The ability to sign transactions in a verifiable and provable manner provides the opportunity to privately prove ownership and even revoke or exchange ownership of Bitcoin private keys and associated Bitcoin addresses without publicly moving anything on the blockchain. In this way, Bitcoin addresses can be funded and the contents of the addresses can be transferred or sold without leaving a public record. Because the process is a threshold system, the allocation of key slices can be accomplished securely without recording any additional payments on the blockchain.
[0218] In this way, it is possible to separate the ownership of notes that have already been settled on the blockchain from the process of transacting those notes.
[0219] <cltv> Bitcoin messages, or in more general terms, transactions, can be created with the inclusion of a CLTV [BIP 65] entry. This addition makes a transaction recoverable even in the event of catastrophic loss of all key slices, or when multiple slices from an entity are lost or deemed untrusted in a way that does not allow for the secure reconstruction of a signature with a minimum threshold.
[0220] This is further possible if an entity uses a third-party service and wants to ensure that the service cannot retain or deny access to the key. In structuring Bitcoin transactions with a time-based failsafe, users know that a malicious third party or a compromised exchange site or bank cannot force them to access their key. In a worst case scenario, a catastrophic level of compromise would result in a time-based reversion of transactions to a predefined address based on the CLTV condition. This predefined address can be generated using the protocol disclosed in this application. As such, it is possible to construct a set of transactions and keys that cannot be easily compromised.
[0221] <Safety concerns> Benger et. al. (2014) propose an example of ECDSA private key recovery by the flash and reload method. This event is just one example of an attack on system RAM and caches. These methods leave out the use of procedures such as Shamir's SSS
[1979] , since they reconstruct the private key. Furthermore, in any scenario where the private key is reconstructed at any time, a requirement for reliability is introduced. This scenario requires reliance on the systems and processes of the entity that holds the private key.
[0222] Even if the trusted parties are not malicious, we must rely on their process, and as a number of recent compromises have shown, this reliance on reconstructing the private key leaves an avenue for attack.
[0223] Being both a drop-in replacement for existing ECDSA implementations and completely transparent and compatible with the current Bitcoin protocol, the implementation does not require hard or soft forks and is indistinguishable from any current transactions. The invention can treat individuals as separate participants allowing group signing of keys with resilience. As an example, a two-of-two scheme can be implemented with four key slices, where the online wallet provider or exchange holds two key slices and the end user holds two slices. The exchange and user each have a two-of-two process for their key slices, which can be used in conjunction with each other for secure signing of messages as needed.
[0224] <Threshold Vault> The above processes and techniques may be used to generate public-private key pairs such that no node can independently generate a private key. The private key shares held by each node individually are never revealed in interactions and may be refreshed as needed for additional security or to address key share loss. This functionality encourages the implementation of "threshold vaults" in blockchain networks. A threshold vault is an address on the blockchain with which a digital asset may be recorded. A threshold number of nodes holding private key shares are then required to cooperate to access, transfer, or otherwise handle the digital asset in the threshold vault. A threshold vault may be established such that all or a portion of the key shares are required to digitally sign a transaction that has the threshold vault as an input. The threshold may be set depending on the purpose and goal of the vault. For example, in some cases where the primary purpose is to protect against disclosure of the private key, the threshold may be set at 100% of the key shares. In some cases where the primary purpose is a voting scheme, 51% or more of the key shares may be set as the threshold. In one example, both security and protection against key share loss may be the objective, in which case a 3 / 5 threshold (for example) may be set by the primary vault owner controlling 3 / 5 of the key shares and by one or more third party repositories holding the remaining 2 / 5 of the key shares, thereby allowing the primary vault owner to cooperate with the third party repositories to allow them to use their key shares in situations in which the primary vault owner loses access to one or more (but not all) of their key shares. In this way, the third party repositories can become the guardians of the "backup" key shares in the event that one of the primary vault owner's key shares is lost.
[0225] By way of example, reference is now made to FIG. 4, which diagrammatically illustrates an example of share distribution associated with a threshold vault implementation. In this example, a partial blockchain 100 is depicted that includes a block 102. The block 102 is comprised in part of a transaction, and includes a transaction 104 that features an output 106 associated with a vault address P. The vault address P is publicly available, whereby an unspent transaction output (UXTO) associated with the vault address is available to be an input to other transactions.
[0226] In order for the UXTO to be used as an input to other transactions, the private key V corresponding to the vault address needs to be used in generating the digital signature. In this case, the private key V has not been generated and will never be generated. Instead, the private key is distributed across multiple nodes 110 as private key shares (V1, V2, ... V n These key shares may be stored in a full blockchain node (e.g., a server, a personal computer, etc.), a wallet node (e.g., implemented within a personal computer, tablet, laptop, mobile device, or any other networked computing device capable of running blockchain wallet software), offline memory / storage (e.g., a USB stick, memory card, or other persistent digital storage medium), an online repository (e.g., a blockchain exchange account, a banking network device, or other remote third-party operating service that stores blockchain data), or any other such computing device.
[0227] In some example implementations, each node may be operated or controlled by a separate entity such that cooperation among a threshold number of owners is required to generate a digital signature for vault address P. In some example implementations, a threshold number of nodes may be controlled by a single entity even if the key shares are physically or logically in separate electronic devices for key protection. Additional key shares, rather than a threshold number, are stored in electronic devices operated by other entities that may serve to act as key recovery services in the event that the primary owner loses a key share. In yet other example implementations, two (or more) entities may each control enough key shares to collaborate in digital signature generation, e.g., in the case of two (or more) partners, i.e., a principal and a cosignatory (e.g., a bank, a parent company, etc.). Those skilled in the art will appreciate the range and types of logical and legal relationships that may be reflected in the distribution of key shares between parties and the setting of associated thresholds for signing.
[0228] Reference is now made to Figure 5, which illustrates, in flow chart form, an example method 200 for receiving and transferring digital assets using a threshold vault. In this example, multiple nodes should cooperate in generating threshold vault addresses and in privately and securely generating and distributing corresponding key shares without any central dealer or authority, thereby ensuring that no node has access to any other node's key shares during the generation process or digital signing process.
[0229] At operation 202, the nodes collaboratively generate threshold vault addresses (e.g., public keys). To do so, each node independently generates and stores a respective private key based on a respective polynomial selected secretly by each individual node. The polynomial selected by each node has an order based on a threshold agreed upon between the nodes. For example, among j nodes, in initiating the threshold vault generation process, the nodes have agreed to set a threshold number of k cooperating nodes required to use (the term "use" as used herein means that the private key is never actually reconstructed by any node) a private key, where k≦j. The threshold vault address may be determined by any or each of the nodes based on a particular polynomial value shared by the nodes and using polynomial interpolation.
[0230] At operation 204, a transaction is added (and validated and approved) to a block on the blockchain that moves the digital asset to the threshold vault address.
[0231] At some point in the future, in operation 206, a transaction is generated (which may be more generally referred to as a "message"). The transaction proposes to use a digital asset as an input to the transaction. This proposed transaction may be generated by one node or by other unrelated nodes. Upon receiving the message featuring the threshold vault address as input, the node determines whether to approve the transaction by digitally signing it.
[0232] If the nodes are instructed to approve the transaction, at least a threshold number k of nodes cooperate to generate a digital signature for the message, in operation 208. This cooperation relies on the use of joint zero secret sharing to generate randomized mask shares between the nodes that are used in the process to mask any key shares used, thereby ensuring that no node receives any other node's key share. Any mask shares or key shares used in the digital signature process are generated and distributed using a secret dealerless distribution. The digital signature process includes the determination of a first signature component r and a second signature component s. In the determination process, any temporary or private key shares are masked before being distributed, and any shares have a value that includes the masked key share embedded within another formula, as detailed above and further described below. Thus, the digital signature for the message is generated without even generating a perfect private key and without revealing any private key shares held by any of the nodes.
[0233] Reference is now made to FIG. 6, which illustrates, in flow chart form, an example method 300 for generating a threshold vault address. The threshold k is set in advance. Each of the nodes establishes secure communications with each of the other nodes. This may include determining a common secret for the secure exchange of information, as described in [Wright, 2016]. Note that any other encryption key generation process may be used, provided that each node is capable of sending encrypted communications to each of the other nodes and receiving and decrypting encrypted communications from each of the other nodes.
[0234] The method 300 includes: i Each of (i=0~j) has its own first secret a0 (i) The first secret may be obtained or determined in any suitable manner.
[0235] Each node also has its own respective polynomial f of degree k-1. i Select (x) a free term of a polynomial, e.g., f i (0) is the node's first secret, i.e., coefficient a0. Each polynomial has a different coefficient a i each having the formula:
number
[0236] Each node determines the value of its own polynomial for each of the other nodes. In this regard, x may be set to the index h associated with the other nodes h=1,...,(i-1),(i+1),...,j. That is, each node determines the value of f for all h, h ≠ i. i Determine the value of (h). Operations of method 300 are described below with respect to a "first node" i, although the same or similar operations are performed by each of the other nodes.
[0237] As indicated by operation 302, a first node i receives a second secret f for node h. i (h) with its node h securely (i.e., over an encrypted channel). That is, for each other node h, the first node i shares a value f i Then, the first node i also broadcasts all the second secret elliptic curve point multiplications (ECPMs) that it has determined to all other nodes h. That is, the first node i sends f i The first node i also sends the ECPMa of its own polynomial coefficients for K = 0, , (k-1). K (i) ×G to each of the other nodes. Thus, the first node i does not disclose its polynomial coefficients, but rather a “public key” corresponding to each coefficient, and does not disclose the determined second secret in general, but rather a “public key” corresponding to each determined second secret. The first node i shares one second secret only privately with each of the other nodes.
[0238] In operation 304, the first node i receives from each of the other nodes (or at least k-1 of them) a "second secret" determined by the other nodes for index x. That is, the first node i receives a value f for each h≠i (or at least k-1 of the j-1 nodes). h The first node i also receives the ECPM values broadcast by the other nodes, i.e. the second secret ECPM and the ECPM of each of the polynomial coefficients.
[0239] Although not explicitly shown in Figure 6, nodes may independently verify and validate second secrets received from other nodes based on the broadcasted values, i.e., each node may determine that the second secret obtained from node h is valid by calculating its ECPM value and comparing it to the broadcasted ECPM values from the other nodes.
[0240] In operation 306, the first node i receives each second secret f from each of the other (at least k−1) nodes. h The first private key share d is the sum of (i) A(i) The sum of each of these second secrets determined by node i is:
number
[0241] In operation 308, one or more of the nodes (which may be delegated to a node other than the first node, or may be performed by more than one node or all nodes) obtains the threshold vault address, i.e., the address at which the nodes each receive their private key share d. A(i) has a private key d A The public key Q corresponding to A , the public key is obtained from the sum of the ECPMs of a threshold number of second secrets using polynomial interpolation. As mentioned above, this is
number
[0242] As can be seen, the method 300 uses a public key Q A A private key share d generated independently at a node corresponding to a threshold vault address given by A(i) results.
[0243] Reference is now made to FIG. 7, which illustrates, by way of example, in flow chart form, a method 400 for digitally signing a message relative to a threshold vault. The method 400 is used to generate a digital signature for a message m, which may be represented by a message hash e=H(m). The message m may be considered to be related to a proposed blockchain transaction in which the threshold vault is an input. Thus, for the transaction to be validated, it must be signed using the threshold vault public key Q. A Of course, none of the nodes possesses a private key and does not have the ability to generate a private key; they only possess a private key share d A(i) Nonetheless, method 400 provides a mechanism for a threshold number of nodes to cooperate in generating valid digital signatures without even disclosing their private key shares.
[0244] As indicated by operation 402, method 400 uses dealerless secret distribution to generate a temporary key share D k(i) and multiplication mask share a i Node i may generate a secret key share d generated using method 300. A(i) In operation 404, the joint zero secret sharing (JZSS) algorithm calculates the first and second sum mask shares β i and c i (eg, as described above in connection with Algorithm 2) to generate the first sum mask share and the second sum mask share are not related.
[0245] Then, in operation 406, each node publishes a double masked key share formed based on the temporary key share masked by the multiplication mask share and the first additive mask share. In one example, the double masked key share is θ i =D k(i) α i +β i modn This is expressed as:
[0246] Each node also publishes a public multiplication mask share that corresponds to the ECPM of the multiplication mask share. ω i = G × α i It can be expressed as:
[0247] No node ever reveals its public key shares, ephemeral key shares, or mask shares.
[0248] At operation 408, a first signature component r is determined from the published data. The determination may be made by each of the nodes independently, or by one of the nodes and shared and / or verified by the other nodes.
[0249] The Elliptic Curve Digital Signature Algorithm (ECDSA) involves determining the inverse of a random value k, which in this example is an ephemeral key D k That is, the inverse is D k -1 To determine the first signature component r, a masked ephemeral key μ is determined using polynomial interpolation on the doubly masked key shares: μ = Interpolate(θ1, ,θ n )modn Since the sum mask is based on joint zero shares, it falls off the interpolation, resulting in μ = D k Similarly, the exponential polynomial interpolation for the ECPM of the multiplicative mask share is given by θ=Exp-Interpolate(ω1,...,ω n ) θ=G×α Then, the first signature component r is given by r x,y =(R x ,R y )=θ×μ -1 This is obtained by calculating r x,y =(G×α)×[D k αmodn] -1 =G×D k -1 The first signature component r is R x modn.
[0250] At operation 410, each node (or at least k of them) receives a second signature component share s i The second signing component includes the ephemeral key, the private key, the hash of the message, and the first signing component. No node holds the ephemeral key or the private key; they only have key shares. Thus, each node determines the second signing component share s i of s i =D k(i) (e+d A(i) r)+c i modn This value is then broadcast or shared with other nodes. The second additive mask share is used to improve the security of revealing the second signature component share.
[0251] Then, in operation 412, a second ECDSA signature component s is obtained from the second signature component shares through polynomial interpolation: s = Interpolate(s1, ,s j )modn Here, as before, the second additive mask share c i falls out of the interpolation since it originates from joint zero shares. Nevertheless, it serves to provide a random mask for the published second signature component share. Operation 412 may be performed by one or more of the nodes, or by a proxy node. Having generated the two components (r, s), a digital signature for message m is thus generated without revealing any of the private key shares or any of the ephemeral key shares.
[0252] At operation 414, the digital signature is incorporated into a blockchain transaction that includes the now digitally signed message, which may then be submitted to the network for validation and final confirmation.
[0253] Reference is now made to Figure 8, which illustrates in block diagram form a simplified example of a node 800. The node 800 includes a processor 802, which may include one or more microprocessors, application specific integrated chips (ASICs), microcontrollers, or similar computer processing devices. The node 800 further includes a memory 806, which may include persistent and non-persistent memory for storing values, variables, and in some cases, processor-executable program instructions. The node 800 further includes a network interface 804 that provides network connectivity over a wired or wireless network.
[0254] The node 800 includes a processor-executable blockchain application 808 that includes processor-executable instructions that, when executed, cause the processor 802 to perform one or more of the functions or operations described herein.
[0255] It will be understood that any modules, routines, processes, threads, applications, or other software components implementing the devices and processes described herein, as well as the described methods / processes for configuring nodes, may be implemented using standard computer programming techniques and languages, and the application is not limited to any particular processor, computer language, computer programming specifications, data structures, or other such implementation details.
[0256] <Conclusion> This invention forms the basis of what Bitcoin tries to achieve with the introduction of a group signature process. The addition of a fault-tolerant signature system in combination with a distributed key generation system removes all centralization and trust requirements.
[0257] Furthermore, the introduction of an implicitly decentralized system allows the creation of more robust and resilient protocols. The compatibility between ECDSA [Johnson, 2001] and Shamir's SSS [Shamir, 1979] allows the present inventors to introduce a system that extends Bitcoin with a new verifiable secret sharing scheme that is much more effective than those derived by Feldman [Feldman, 1987] or Pedersen [Pedersen, 1992] without losing anything in security.
[0258] In this application, a system has been described that extends the functionality of Bitcoin without requiring modifications to the base protocol. According to the invention: 1. A trusted third party is no longer required for the selection or distribution of key secrets. 2. A decentralized banking system can be built that does not rely on third-party trust. 3. Each member or group of members can independently verify that the private key share they hold corresponds to a published Bitcoin address and public key. 4. A protocol exists for refreshing private key slices in a way that mitigates the effects of eavesdropping and related attacks. 5. A trusted third party is not required for group signing of transactions and messages.
[0259] The present invention completely solves many existing security risks by ensuring that sensitive data never appears in memory.
[0260] It should be noted that the above embodiments illustrate rather than limit the invention, and that those skilled in the art can design many alternative embodiments without departing from the scope of the invention as defined by the appended claims. In the claims, any reference signs in parentheses shall not be construed as limiting the scope of the claims. The words "comprising" and "comprises" and the like do not exclude the presence of elements or steps other than those listed in any claim or the specification as a whole. In this specification, "comprising" means "includes or consists of" and "comprising" means "including or consisting of". A single reference of an element does not exclude a plurality of references of such elements and vice versa. The invention may be implemented by means of hardware comprising several distinct elements, and by means of a suitably programmed computer. In a device claim enumerating several means, several of these means may be embodied by one and the same item of hardware. The mere fact that certain measures are recited in mutually different claims does not indicate that a combination of these measures cannot be used to advantage.
[0261] <References> 1)Bar-Ilan, J. Beaver, "Non-Cryptographic Fault-Tolerant Computing in a Constant Number of Rounds", Proc. of 8th PODC, pp.201-209, 1989. 2)Berlekamp, Elwyn R. (1968), Algebraic Coding Theory, McGraw‐Hill, New York, NY. 3)Benger, N., van de Pol, J., Smart, N.P., Yarom, Y.: "Ooh Aah... Just a Little Bit": A Small Amount of Side Channel Can Go a Long Way. In: Batina, L., Robshaw, M. (eds.) Cryptographic Hardware and Embedded Systems | CHES 2014, LNCS, vol.8731, pp.75-92. Springer (2014) 4)Ben-Or, M., Goldwasser, S., Wigderson, A.: "Completeness theorems for noncryptographic fault-tolerant distributed computation”. In: Proceedings of the Twentieth Annual ACM Symposium on Theory of Computing. pp.1‐10. STOC ’88, ACM, New York, NY, USA (1988) 5)BIP 65 OP_CHECKLOCKTIMEVERIFY https: / / github.com / bitcoin / bips / blob / master / bip-0065.mediawiki 6)Chaum, David (1983). "Blind signatures for untraceable payments" (PDF). Advances in Cryptology Proceedings of Crypto. 82 (3): 199‐203. 7)Dawson, E.; Donovan, D. (1994), "The breadth of Shamir's secret-sharing scheme", Computers & Security, 13: Pp.69‐78 8)Desmedt. You (1987). "Society and Group Oriented Cryptography: A New Concept”. In A Conference on the Theory and Applications of Cryptographic Techniques on Advances in Cryptology (CRYPTO '87), Carl Pomerance (Ed.). Springer-Verlag, London, UK, UK, 120-127. 9)Feldman. P. "A practical scheme for non-interactive verifiable secret sharing”. In Proceedings of the 28th IEEE Annual Symposium on Foundations of Computer Science, pages 427‐437, 1987. 10)Gennaro, R., Jarecki, S., Krawczyk, H., Rabin, T.: "Robust threshold DSS signatures”. In: Proceedings of the 15th Annual International Conference on Theory and Application of Cryptographic Techniques. pp.354‐371. EUROCRYPT’96, SpringerVerlag, Berlin, Heidelberg (1996) 11)Ibrahim, M., Ali, I., Ibrahim, I., El-sawi, A.: "A robust threshold elliptic curve digital signature providing a new verifiable secret sharing scheme”. In: Circuits and Systems, 2003 IEEE 46th Midwest Symposium on. vol.1, pp.276‐280 (2003) 12)Johnson, D., Menezes, A., Vanstone, S.: "The elliptic curve digital signature algorithm (ecdsa)”. International Journal of Information Security 1(1), 36‐63 (2001) 13)Kapoor, Vivek, Vivek Sonny Abraham, and Ramesh Singh. "Elliptic Curve Cryptography." Ubiquity 2008, no. May (2008): 1-8. 14)Knuth, D. E. (1997), "The Art of Computer Programming, II: Seminumerical Algorithms” (3rd ed.), Addison-Wesley, p.505. 15)Koblitz, N. "An Elliptic Curve Implementation of the Finite Field Digital Signature Algorithm" in Advances in Cryptology ‐ Crypto '98. Lecture Notes in Computer Science, vol.1462, pp.327-337, 1998, Springer-Verlag. 16)Liu, C. L. (1968), "Introduction to Combinatorial Mathematics”, New York: McGraw-Hill. 17)National Institute of Standards and Technology: FIPS PUB 186-4: "Digital Signature Standard” (DSS) (2003) 18)Pedersen, T.: "Non-interactive and information-theoretic secure verifiable secret sharing”. In: Feigenbaum, J. (ed.) Advances in Cryptology ‐ CRYPTO ’91, LNCS, vol. 576, pp.129‐140. Springer (1992) 19)Rabin T. & Ben-Or. M. (1989) "Verifiable secret sharing and multiparty protocols with honest majority". In Proc. 21st ACM Symposium on Theory of Computing, pages 73--85, 1989. 20)Shamir, Adi (1979), "How to share a secret", Communications of the ACM, 22 (11): Pp.612‐613 21)Wright, C. & Savanah, S. (2016) "Determining a common secret for two Blockchain nodes for the secure exchange of information” International Patent Application Number Application Number: WO 2017 / 145016.< / cltv> < / ecdsa>
Claims
1. 1. A computer-implemented method for securely generating a digital signature for a message related to a threshold vault, comprising: the threshold vault having a public key address, a private key, and a threshold value, each node among the plurality of nodes having a respective private key share, a temporary key share, a multiplication mask share, a first additive mask share, and a second additive mask share, the method being performed by a first node among the plurality of nodes, comprising: generating and sharing a double masked key share from the temporary key share masked by the multiplication mask share and by the first addition mask share, and an elliptic curve point multiplication of the multiplication mask share; receiving from a plurality of other nodes respective double masked key shares corresponding to the plurality of other nodes and respective elliptic curve point multiplications of respective multiplication mask shares corresponding to the plurality of other nodes; determining a first signature component using polynomial interpolation on the double masked key shares and polynomial interpolation on the elliptic curve point multiplication of the multiplication mask shares; determining and publishing a second signature component share based on the message masked by the second additive mask share, the private key share, the first signature component, and the ephemeral key share; having a second signature component is obtainable from signature shares of at least a threshold number of nodes, the signature shares including a second signature component share of each of the plurality of nodes; The method.
2. generating a private key share, an ephemeral key share, and a multiplication mask share for the first node using a dealerless secret distribution. The method of claim 1.
3. generating a first sum mask share and a second sum mask share for the first node using the joint zero secret share. The method of claim 1.
4. generating the double masked key shares, [0010] Calculating θ i is the double masked key share, D k(i) is the ephemeral key share of the first node, and α i is the multiplication mask share of the first node, and β i is the first additive mask share of the first node, and n is the integer order of the elliptic curve. The method of claim 1.
5. the polynomial interpolation on the doubly masked key shares results in a multiplicatively masked ephemeral key; the polynomial interpolation to the elliptic curve point multiplication of the multiplication mask shares results in an elliptic curve point multiplication of a multiplication mask; determining the first signature component includes determining a product of an elliptic curve point multiplication of the multiplication mask and an inverse of the multiplicatively masked ephemeral key. The method of claim 1.
6. The second signature component share s i To determine [0025] determining D k(i) is the first node's ephemeral key share, e is the hash of the message, and d A(i) is the first node's private key share, r is the first signature component, and c i is the second additive mask share, and n is the integer order of the elliptic curve. The method of claim 1.
7. determining the second signature component by polynomial interpolation on the signature shares of at least the threshold number of nodes. The method of claim 1.
8. adding the first signature component and the second signature component to a transaction including the message and submitting the transaction to a blockchain network. The method according to claim 7.
9. 1. A computer-implemented method of multi-party computation, comprising: distributing a plurality of shares of a secret among a plurality of first nodes, each said secret being a constant term of a respective polynomial function, each said share being a respective value of said polynomial function associated with said first node, said secret being accessible by polynomial interpolation of a threshold number of said shares and being inaccessible in the absence of said threshold number of shares; receiving, at a second node, respective shares of the computed secret from a plurality of said first nodes, each said first node computing its share of the computed secret by performing at least one predetermined computation on each of the plurality of shares of the secret assigned to said first node; determining, at the second node, the computed secret by polynomial interpolation of at least a threshold number of shares including the share of the computed secret; The method comprising the steps of:
10. For each polynomial function, the following formula [0030] is represented by f i is the polynomial function of the i-th node, n is the integer order of the elliptic curve, x is a variable in the polynomial function, and a t is the set of coefficients for the polynomial function, and t is the index.
10. The method of claim 9.
11. and sending, to a plurality of nodes, elliptic curve point multiplications of each coefficient of a plurality of the polynomial functions.
10. The method of claim 9.
12. sending, to a plurality of nodes, an elliptic curve point multiplication of at least one of the polynomial function values associated with each of the plurality of nodes.
10. The method of claim 9.
13. and verifying consistency between an elliptic curve point multiplication of at least one of the polynomial function values associated with at least one of the nodes and a sum of elliptic curve point multiplications of the polynomial function coefficients. The method of claim 11.
14. Determining the computed secret share comprises: μ i = x i ・y i mod determining x i is the share of the first secret, and y i is the second secret share, and n is the integer order of the elliptic curve.
10. The method of claim 9.
15. Determining the computed secret share comprises: x i +y i mod determining x i is the share of the first secret, and y i is the second secret share, and n is the integer order of the elliptic curve.
10. The method of claim 9.
16. Determining the computed secret share comprises: x i -1 mod determining x i -1 is the inverse of the first secret share modulo n, where n is the integer order of the elliptic curve.
10. The method of claim 9.
17. providing modified shares of the calculated secret by adding zero secret shares, which are the constant terms of a polynomial function equal to zero, such that the calculated secret is accessible by polynomial interpolation of a threshold number of the modified shares and is inaccessible in the absence of the threshold number of shares.
10. The method of claim 9.
18. determining the calculated secret comprises determining a polynomial function corresponding to the calculated secret share by determining coefficients of the polynomial function from a plurality of known values of the calculated secret share.
10. The method of claim 9.
19. determining the polynomial function comprises executing an error correction algorithm.
20. The method of claim 18.
20. determining the polynomial function comprises performing a Berlekamp-Welch decoding algorithm; 20. The method of claim 18.
21. Determining the polynomial function comprises: defining an error locator polynomial function and a second polynomial function which is the product of said polynomial function and said error locator polynomial function related to said calculated secret shares; determining coefficients of the second polynomial function and the error locator polynomial function from a plurality of known values of the calculated secret share; determining the polynomial function associated with the calculated secret share from the second polynomial function and the error locator polynomial function; having 20. The method of claim 18.
22. 1. A computer-implemented method for distributing shares of a private key of a public-private key cryptography system among a plurality of nodes, comprising: Each of the plurality of nodes has a respective first secret, a respective first polynomial function with the respective first secret of the node set as a free term, and a respective first share of a secret key, the secret key being a sum of the first secrets and a free term of a second polynomial function, the secret key being accessible by polynomial interpolation of at least a threshold number of the first shares and being inaccessible for less than the threshold number of the first shares, the method being performed by a first node included in the plurality of nodes, the method comprising: selecting each third polynomial function having zero set as its respective free term; for each other node of the plurality of nodes, setting a variable in the third polynomial function to a value associated with that other node, determining a result of each of the third polynomial functions for that value, and securely transmitting each of the results to that other node; receiving from each other node a respective share of a respective second secret, the share being a result of a third polynomial function of each of the other nodes on a first value associated with the first node; forming and storing a second share of each of said secret keys, said second share being the sum of said respective first shares and the sum of said received share of said second secret; The method comprising the steps of:
23. sending to the other nodes each of the elliptic curve point multiplications of the results; sending to the other nodes elliptic curve point multiplications of coefficients of the third polynomial function associated with the first node; Further comprising 23. The method of claim 22.
24. receiving from each other node the elliptic curve point multiplication of the result associated with that node; receiving from each other node an elliptic curve point multiplication of coefficients of said third polynomial function associated with that node; Further comprising 23. The method of claim 22.
25. and verifying the consistency of the result with respect to at least one other node with coefficients of the third polynomial function with respect to that other node.
25. The method of claim 24.
26. generating a share of a digital signature based on the message and the second share of the private key.
23. The method of claim 22.
27. receiving, from at least a threshold number of nodes, respective shares of a message and digital signatures based on a second share of each of the private keys; generating a digital signature by polynomial interpolation of at least a threshold number of shares of said digital signature; Further comprising 23. The method of claim 22.
28. A computing device that participates in a blockchain transaction, comprising: the computing device is a first node among a plurality of nodes; The computing device includes: A processor; Memory, a network interface providing a network connection to other nodes in the plurality of nodes; a blockchain application comprising computer executable instructions which, when executed by the processor, cause the processor to perform the method of any one of claims 1 to 27; having The computing device.
Citation Information
Patent Citations
Semi-deterministic digital signature generation
JP2017517770A
