KR20200083145A - Fault tolerance consensus method for eliminating interference factors in blockchain networks - Google Patents

Abstract

The present invention can provide a fault tolerance consensus method of a blockchain network, which may return to normal consensus by adjusting voting rights of normal nodes and abnormal nodes by changing a consensus boundary by applying a penalty for the abnormal nodes based on voting information of a blockchain verification node.

Description

Fault tolerance consensus method for removing consensus obstacles of blockchain network {FAULT TOLERANCE CONSENSUS METHOD FOR ELIMINATING INTERFERENCE FACTORS IN BLOCKCHAIN NETWORKS}

The present invention relates to a method for consensus to allow an obstacle in a blockchain network, and to a method for consensus to allow an obstacle to remove an obstacle for consensus in a blockchain network.

Blockchain technology is a high-reliability decentralized information storage (DB: Database) technology that supports transparency, security, and irreversibility.As a simple structure, it is easy to secure irreversibility, so financial/logistics/public sector where integrity is very important Many projects are in progress as it has been in the limelight.

Consensus Algorithm is a technology that determines which information to select to maintain one DB between nodes when information mismatch occurs in a P2P network in which multiple nodes exist. Blockchain decides the authority for block creation and branched blockchain selection through consensus algorithm. Representative consensus algorithms include Proof of Work (PoW), Proof of Stake (PoS), and Practical Byzantine Fault Tolerance (PBFT).

If a blockchain branching phenomenon occurs in which blocks with different data occur at the same timestamp, PoW and PoS select the longest extended blockchain. However, this lowers the data reliability of the blockchain in a way that allows for temporary information discrepancies.

On the other hand, PBFT prevents the inconsistency in the blockchain data by adding blocks to the blockchain only after the normal consensus between the consensus nodes for the data to be added to the blockchain is completed, thereby improving the reliability of the data. The PBFT consensus algorithm allows normal consensus of the Byzantine Fault node under a certain ratio even in an asynchronous network where Byzantine Faulty nodes that do not perform the promised action can exist. When the total number of verify nodes is n and the number of fault nodes is f, consensus may be successful in the PBFT if the number of fault nodes f is less than (n-1)/3. That is, if the total number of verification nodes (n) exceeds three times the number of faulty nodes (f) (N ≥ 3f+1), normal consensus is possible even if a faulty node is included.

1 shows an example of a consensus procedure through a PBFT consensus algorithm.

Referring to FIG. 1, the PBFT consensus algorithm includes a request (S11), a pre-prepare (S12), a prepare (S13), a commit (S14), and a response (Reply) may be configured in step (S15).

Validation nodes (Validating Nodes) 110 are responsible for verifying and agreeing on a transaction request transmitted from a client 120 among a plurality of nodes in the blockchain network and adding it to the block. In FIG. 1, for example, four verification nodes 111 to 114 exist in the blockchain network, and the first verification node 111 among the four verification nodes 111 to 114 receives a transaction request from the client 120. It is a leader node, and the remaining second to fourth verification nodes 112 to 114 are backup nodes that consensusly verify the requested transaction with the leader node 111.

In the request step (S11) of the PBFT consensus algorithm, the leader node 111 receives the status change of data from at least one client 120, that is, the data requesting the transaction, and synthesizes, verifies, and sorts it. And in the preliminary preparation step (S12), the leader node 111 propagates the applied block to the remaining verification nodes 112 to 114, that is, backup nodes. In the preparation step (S13), each of the backup nodes 112 to 114 verifies the block in the leader node 111, and if the verification result is true, propagates the preparation message to the other verification nodes 111 to 114. In the determination step S14, each of the verification nodes 111 to 114 receives the same value from 2f or more nodes and propagates the determination message to other nodes. In the response step (S15), each of the verification nodes 111 to 114 receives a determination message of the same value from 2f+1 or more other nodes, adds the corresponding block to the chain, and delivers the result to the client 120 . Accordingly, when the client receives the same result from other nodes of f+1 or higher, the client confirms that the result is processed.

In this way, since the block data can be changed only by obtaining more than 2/3 of the total verification node (n) in the PBFT, securing more than 51% of the nodes is more resistant to data tampering attacks than PoW and PoS capable of data tampering. However, there is a problem in that a data falsification attack can cause a normal agreement to fail even if only more than one third of the total verification nodes are secured, and a deadlock may not be returned to the normal agreement.

Korean Open Patent No. 10-2018-0113140 (published 2018.10.15)

The object of the present invention is to incorporate a trust evaluation model based on the PBFT consensus algorithm, improve the success rate of consensus through imposing penalties for abnormal behaviors, and allow a consensus method of a blockchain network to allow a consensus deadlock to return to normal consensus. To provide.

In order to achieve the above object, a method of concluding a fault tolerance for a blockchain network according to an embodiment of the present invention, when a given leader node among n verification nodes receives a block including a transaction request transmitted from at least one client , Providing a leader reliability, which is a local reliability obtained from a previous voting result, to the leader node in the block, and propagating to the remaining verification nodes, backup nodes; When each of the backup nodes compares the backup reliability, which is the local reliability of the self obtained from the previous voting results, with the reader reliability, and if they are different from each other, propagates a preparation message that changes the reader reliability to the backup reliability to other verification nodes. step; Analyzing the preparation message received by each of the verification nodes, and if the number of local trusts having the same value is equal to or greater than a predetermined reference number, propagating the decision message by applying the local trust to the global trusts to other verification nodes ; And each of the verification nodes voting for a transaction request based on the sum of global reliability included in the decision message, adding a block to the blockchain; It includes.

)의 합이 기지정된 글로벌 기준 비율 이상인지 판별하는 단계; 상기 글로벌 기준 비율 이상이면, 상기 블록에 대한 글로벌 신뢰도 합의가 성공한 것으로 판별하여 블록을 블록체인에 추가하는 단계; 추가된 결과를 상기 클라이언트로 전달하는 단계; 및 합의 성공 여부에 따라 검증 노드 각각에 대한 로컬 신뢰도를 재계산하는 단계; 를 포함할 수 있다.The step of adding to the blockchain is the global reliability of each of the n verification nodes (

Determining whether the sum of) is greater than or equal to a predetermined global reference ratio; If it is equal to or greater than the global reference ratio, determining that the global reliability agreement for the block is successful and adding the block to the blockchain; Passing the added result to the client; And recalculating local reliability for each verification node according to whether the agreement is successful. It may include.

In the step of recalculating the local reliability, if it is determined that the global reliability agreement is successful, the local reliability of the t-th round for the k-th verification node calculated by the i-th verification node among the n verification nodes (C ^t _i,k ) Equation

는 k번째 검증 노드에 대한 t번째 라운드의 글로벌 신뢰도이고, V_k는 n개의 검증 노드 중 k번째 검증 노드를 의미하고, R은 정상적인 리더 노드의 제안에 동의하는 투표를 수행한 로얄 노드의 집합을 나타내고, F는 장애 노드 집합을 나타내며, α는 패널티 가중치이다.)에 따라 계산하는 단계; 및 글로벌 신뢰도 합의가 실패한 것으로 판별되면, 상기 검증 노드들 각각에 대한 로컬 신뢰도를 수학식(here

Is the global reliability of the t-th round for the k-th verification node, V _k is the k-th verification node among n verification nodes, and R is the set of royal nodes that have voted to agree to the proposal of the normal leader node. F is a set of faulty nodes, and α is a penalty weight. And when it is determined that the global reliability agreement has failed, the local reliability for each of the verification nodes is expressed by the equation.

Calculating according to; It may include.

The step of recalculating the local reliability is based on whether the global reliability agreement is successful, the consensus boundary is based on the global reliability.

It can be changed according to.

Propagating the determination message may include analyzing a preparation message received by each of the verification nodes to determine whether the number of local reliability values having the same value is greater than or equal to a predetermined reference number; If the number is greater than or equal to the reference number, reflecting the local reliability of the same value in the global reliability of the current round; If it is less than the reference number, reflecting the backup local reliability to the global reliability of the current round; And determining whether the sum of global reliability of verification nodes having the same local reliability value is equal to or greater than a predetermined global reference value, and if it is equal to or greater than the global reference value, propagating a decision message to another verification node. It may include.

Therefore, the method of allowing a fault to remove the obstacles to the consensus of the blockchain network according to the embodiment of the present invention reduces the influence on the consensus of the Byzantine obstacle nodes among the consensus nodes of the blockchain so that it can return to the normal consensus in a deadlock. By doing so, it is possible to increase the success rate of a normal agreement.

1 shows an example of a consensus procedure through a PBFT consensus algorithm.
2 shows a block chain network configuration according to an embodiment of the present invention.
Fig. 3 shows a method of concluding a disability allowance for removing a blockage of consensus in a blockchain network.
4 shows a modified block structure according to an embodiment of the present invention.
5 shows a schematic structure of a disability-tolerant blockchain terminal according to an embodiment of the present invention.

In order to fully understand the present invention, the operational advantages of the present invention, and the objects achieved by the practice of the present invention, reference should be made to the accompanying drawings and the contents described in the accompanying drawings, which illustrate preferred embodiments of the present invention.

Hereinafter, the present invention will be described in detail by explaining preferred embodiments of the present invention with reference to the accompanying drawings. However, the present invention may be implemented in various different forms, and is not limited to the described embodiments. In addition, in order to clearly describe the present invention, parts irrelevant to the description are omitted, and the same reference numerals in the drawings indicate the same members.

Throughout the specification, when a part “includes” a certain component, this means that other components may be further included, rather than excluding other components, unless otherwise specified. In addition, terms such as "... unit", "... group", "module", and "block" described in the specification mean a unit that processes at least one function or operation, which is hardware or software or hardware. And software.

2 shows a blockchain network configuration according to an embodiment of the present invention.

Referring to FIG. 2, the blockchain network 200 according to the present embodiment includes n verification nodes 211 to 21n (where n is a natural number) capable of mutual communication in a mesh-type among a plurality of nodes. Including, transaction requests sent from multiple clients can be verified and agreed to be added to the block. In FIG. 2, only four verification nodes 211 to 21n are illustrated for convenience of description, but the number of verification nodes may be variously adjusted. In addition, the first verification node 211 among the plurality of verification nodes 211 to 21n is a leader node receiving a transaction request from at least one of the plurality of clients 221 to 22m, and the remaining second to fourth verification nodes It is assumed that 212 to 21n are backup nodes.

As described above, among the plurality of verification nodes 211 to 21n, the leader node 211 verifies and sorts the transaction requests authorized from at least one client 221 to 22m to backup nodes 212 to 21n, which are other verification nodes. ), and confirm the block through consensus among multiple verification nodes 211 to 21n, and extend and maintain the block by adding it to the blockchain.

All n verification nodes (211 to 21n) participating in the blockchain participate in the consensus for each round of consensus, and the malicious node operating from a specific point in time (t ₀ ) performs block tampering attempts or non-consensus. It is effective to send a message to more verification nodes in order to overcome the malicious intention, because the malicious node normally participates in the consensus, so it is not considered to send the normal message and the tamper message to other verification nodes at the same time. . At this time, even if the normal node fails to participate in the agreement due to a time delay, network congestion, or the like, depending on the network environment, a penalty is given because it has a negative effect on the agreement.

When a predetermined leader node 211 among the n verification nodes 211 to 21n of the blockchain network generates a block, and the number of faulty nodes among the n verification nodes 211 to 21n is f, The verification node set (V), the failure node set (F), and the set of royal nodes (R) that voted to agree to the proposal of the normal leader node 211 are V={V _k |k=1, respectively. ..., n}, F={F _k |k=1, ..., f}, R={R _k |k=1, ..., nf}.

Each of the n verification nodes 211 to 21n of the blockchain network calculates local credibility (C ^t _i,k ) by performing a reliability evaluation based on the voting information collected in the commit phase. Here, the local reliability set C ^t _i for the entire verification node calculated by the i-th verification node 21i is expressed by Equation 1.

Each of the verification nodes 211 to 21n maintains the local reliability (C ^t _i,k ) evaluated by voting in round t-1 until round t. The leader node 211 proposes its local reliability (C ^t ₁ ) together with the block at t round. The reliability determined by the agreement between the verification nodes 211 to 21n is recognized as global reliability and can be expressed by Equation (2).

In the Genesis block, the first block of the blockchain, the reliability of the previous node is 1, and the reliability is calculated according to the voting results of the previous round from the second block. First, if a vote is agreed to the proposal of the normal leader node 211, the node is judged as a royal node (R) and maintains the reliability of the previous round.

)와 같다. 이를 통해 로컬 신뢰도(C^t _i,k)는 수학식 3과 같이 계산된다.However, if it is judged that the failure node (F) who votes incorrectly or does not vote, the reliability of the total failure node in the previous round divided by the reliability of the total verification node is reduced by the value multiplied by the penalty weight α. At this time, if global reliability is normally determined through credibility consensus procedures, the t-1 round local reliability of the kth node (C ^t-1 _i,k ) is the t round global reliability of the kth verification node of the global reliability. (

). Through this, the local reliability (C ^t _i,k ) is calculated as in Equation 3.

On the other hand, if the global reliability agreement fails, the reliability of the round is recalculated using Equation 4 using the local reliability of the previous round.

In equations (3) and (4), as the number (f) of fault nodes increases, the probability of failure of the normal agreement increases. Therefore, by increasing the penalty by the reliability agreement ratio of the failure nodes and the entire verification nodes 211 to 21n in proportion to the increase in the reliability agreement of the failure node, the influence of the failure node agreement can be reduced to increase the agreement success rate.

This adaptively changes the consensus boundary by a number of verification nodes, so that despite the existence of a disability node that is a consensus obstacle, the success rate can be improved and the consensus deadlock can be returned to the normal consensus, and the obstacle according to this embodiment The allowable consensus algorithm is called an adaptive consensus bound PBTF (hereafter referred to as ACB-PBFT) algorithm.

The reliability, which means the trust value of each verification node, is calculated for each node. Using the calculated reliability, consensus is performed in a 3-phase protocol method in order for all the verification nodes to share a single trust value.

FIG. 3 shows a method of allowing a consensus for disabling an obstacle to consensus in a blockchain network, and FIG. 4 shows a modified block structure according to an embodiment of the present invention.

Prior to explaining the method of consensing to allow the failure of FIG. 3, first, the modified block structure according to the present embodiment will be described with reference to FIG. 4, and the block includes a block header 410 and a block body. 420. In addition, the block header 410 includes a version field (Version) 411, a previous block hash field (Hash of Prev. Block) 404, a hash field (Hash of Merkle Root) 405, and a time stamp (Time Stamp). ) 406. Meanwhile, the block body 420 may be composed of a transaction counter 421, a reliability transaction 422, and transactions 423.

When the block structure according to the present embodiment shown in FIG. 4 is compared with the block structure of the existing representative blockchain system Bitcoin, the block chain network of the present embodiment selects the block generator leader node as non-competitive, so the block constructor The Difficulty and Nonce fields used in Bitcoin have been removed to determine. In addition, the ACB-PBFT according to the present embodiment uses credibility, which means voting power of each verification node, so each verification node agreed through the Credibility Consensus Procedures described below. The global reliability, which is a full set of reliability, is recorded in the reliability transaction field 422 in the block body 420 so that all verification nodes maintaining the blockchain can share the same global reliability.

Hereinafter, with reference to FIGS. 2 and 4, a method for concluding a fault tolerance of FIG. 3 will be described. The disability tolerant consensus method according to the present embodiment may also basically be performed by including a request step, a preliminary preparation step, a preparation step, a decision step, and a response step, like the PBFT algorithm shown in FIG. 1.

First, the reliability consensus procedure is performed. When a transaction request is transmitted from at least one client to the leader node 211 in the requesting step (not shown), in the preliminary preparation step, the leader node 211 is the leader local reliability calculated as a result of voting in the previous round (t-1). (C _L ^t-1 ) is recorded in the reliability transaction field 422 of the block data of FIG. 4 to propagate the block to the backup nodes 212 to 21n, which are the remaining verification nodes, and propose (S31).

In addition, each of the backup nodes 212 to 21n that has received the block in the preparation phase has its own backup local reliability (C _i ^t-1 ) calculated in the previous round ( ^t-1 ) and the block transmitted from the leader node 211. It is determined whether the leader local reliability (C _L ^t-1 ) recorded in the reliability transaction field 422 is the same (S32).

As a result of determination, if the backup local reliability (C _i ^t-1 ) and the leader local reliability (C _L ^t-1 ) are the same, the leader local reliability (C _L ^t-1 ) is selected, and the selected leader local reliability (C ^{i t-1} ) is selected. _L ^t-1 ) is propagated to the other backup nodes except the leader node 211 as a preparation message (S33). However, if the backup local reliability (C _i ^t-1 ) and the reader local reliability (C _L ^t-1 ) are different from each other, the backup local reliability (C _i ^t-1 ) is recorded in the block's reliability transaction field 422 to read the reader. It propagates to other verification nodes except the node 211 (S34).

In the determining step, all the verification nodes 211 to 21n analyze the local reliability (C _i ^t-1 ) of the prepared message propagated from other verification nodes and the local reliability selected by them (S35 ). Then, it is determined whether the number of local reliability having the same value in the analyzed local reliability is greater than or equal to a predetermined reference number with respect to the number n of all the verification nodes 211 to 21n (S36). Here, the reference number may be set to 2(n-1)/3 as an example.

)에 반영하여, 즉 로컬 신뢰도를 블록의 신뢰도 트랜잭션 필드(422)에 기록하여 결정 메시지로서 다른 노드로 전파한다(S37). 그러나 동일한 값을 갖는 로컬 신뢰도의 개수가 기준 개수 미만이면, 백업 로컬 신뢰도(C_i ^t-1)를 현재 라운드의 글로벌 신뢰도(

)에 반영하여 결정 메시지를 다른 검증 노드로 전파한다(S38).If the number of local reliability (C _k ) having the same value is greater than or equal to the reference number, the corresponding verification node sets the local reliability of the same value to the global reliability of the current round (

), that is, the local reliability is recorded in the block's reliability transaction field 422 and propagated to other nodes as a decision message (S37). However, if the number of local reliability values having the same value is less than the reference number, the backup local reliability (C _i ^t-1 ) is set to the global reliability of the current round (

) To propagate the decision message to another verification node (S38).

) 이상인지 판별하고, 글로벌 기준 값 이상이면, 결정 메시지를 다른 검증 노드로 전파하도록 구성될 수 있다.In this case, each of the verification nodes 211 to 21n is a global reference value (for example, a sum of global reliability of verification nodes that transmit a preparation message, that is, verification nodes having the same local reliability value), for example.

), and may be configured to propagate the decision message to another verification node if it is greater than or equal to the global reference value.

Through this process, the leader reliability proposed by the leader node 211 or the same local reliability shared by 2f+1 or more verification nodes among all verification nodes is recognized as global reliability, and the recognized global reliability value is the block's reliability transaction field (422 ).

)는 각 검증 노드(211 ~ 21n)의 투표 가중치로 이용될 수 있다. 다수의 검증 노드(211 ~ 21n) 각각은 자신을 포함한 다수의 검증 노드 중 동일한 결과 메시지를 전파한 검증 노드들의 글로벌 신뢰도의 합(

)이 기지정된 글로벌 기준 비율 이상인지 판별하고, 다수의 검증 노드(211 ~ 21n) 각각은 글로벌 신뢰도의 합(

)이 글로벌 기준 비율 이상이면, 합의가 성공된 것으로 판별하여 블록을 블록체인에 추가하고, 추가된 결과를 클라이언트로 전달한다(S40). 여기서 기준 비율은 일예로

로 설정될 수 있다.Global reliability in this embodiment (

) May be used as a voting weight for each verification node 211 to 21n. Each of the plurality of verification nodes 211 to 21n is the sum of the global reliability of the verification nodes that propagated the same result message among the plurality of verification nodes including itself (

) Is determined to be equal to or greater than a predetermined global reference ratio, and each of the plurality of verification nodes 211 to 21n is the sum of global reliability (

) Is greater than the global standard ratio, it is determined that the agreement is successful, the block is added to the blockchain, and the added result is transmitted to the client (S40). Here, the reference ratio is an example

Can be set to

As described above, in the ACB-PBFT consensus method according to the present embodiment, the consensus boundary for determining consensus success for a plurality of verification nodes 211 to 21n is adaptively changed, and each consensus boundary Can be expressed as Equation (5).

Then, the local reliability C ^t _i for all the verification nodes 211 to 21n is recalculated according to Equation 3 or 4 according to the voting result for the block (S41).

5 shows a schematic structure of a disability-tolerant blockchain terminal according to an embodiment of the present invention.

Referring to FIG. 5, the terminal may include an information input unit 501, a memory 502, a CPU 503, a message processing unit 504, and an information output unit 505 for information exchange and agreement.

The information input unit 501 serves to receive vote information received from another node.

The memory 502 stores information processed by the CPU. Specifically, the memory stores information such as voting information of each node, calculated reliability information, and blockchain data.

The CPU 503 transmits the received voting information to the message processing unit 504, and transmits the voting information received by the message processing unit to the information output unit 505. In addition, operation of the information input unit 501, the information output unit 505, the memory 502, and the message processing unit 504 is controlled.

The message processing unit 504 analyzes the voting information of other nodes to calculate the reliability information and compares and validates the received messages at each stage of the agreement.

The method according to the present invention can be implemented as a computer program stored in a medium for execution on a computer. Computer readable media herein can be any available media that can be accessed by a computer, and can also include any computer storage media. Computer storage media includes both volatile and nonvolatile, removable and non-removable media implemented in any method or technology for storage of information such as computer readable instructions, data structures, program modules or other data, and ROM (readable) Dedicated memory), RAM (random access memory), CD (compact disk)-ROM, DVD (digital video disk)-ROM, magnetic tape, floppy disk, optical data storage, and the like.

Although the present invention has been described with reference to the embodiments shown in the drawings, these are merely exemplary, and those skilled in the art will understand that various modifications and other equivalent embodiments are possible therefrom.

Therefore, the true technical protection scope of the present invention should be defined by the technical spirit of the appended claims.

Claims (5)

When a block containing a transaction request transmitted from at least one client is received by a designated leader node among n verification nodes, the remaining portion of the block includes the leader reliability, which is the local reliability obtained from the previous voting result, for the leader node. Propagating to backup nodes that are verification nodes;
When each of the backup nodes compares the backup reliability, which is the local reliability of itself obtained from the previous voting results, with the reader reliability, and if they are different, propagates a preparation message that changes the reader reliability to the backup reliability to other verification nodes. step;
Analyzing the preparation message received by each of the verification nodes, and if the number of local trusts having the same value is equal to or greater than a predetermined reference number, propagating the decision message by applying the local trust to the global trusts to other verification nodes ; And
Each of the verification nodes voting for a transaction request based on the sum of global reliability included in the decision message, and adding a block to the blockchain; A method of consensus tolerant of a block chain network comprising a. The method of claim 1, wherein the step of adding to the blockchain
Global reliability for each of the n verification nodes (

Determining whether the sum of) is greater than or equal to a predetermined global reference ratio;
If it is equal to or greater than the global reference ratio, determining that the global reliability agreement for the block is successful and adding the block to the blockchain;
Passing the added result to the client; And
Recalculating local reliability of each verification node according to whether the agreement is successful; A method of consensus tolerant of a block chain network comprising a. The method of claim 2, wherein the step of recalculating the local reliability is
If it is determined that the global reliability agreement is successful, the local reliability (C ^t _i,k ) of the t-th round for the k-th verification node calculated by the i-th verification node among the n verification nodes is expressed by Equation

(here

Is the global reliability of the t-th round for the k-th verification node, V _k is the k-th verification node out of n verification nodes, and R is the set of royal nodes that have voted to agree to the proposal of the normal leader node. F is a set of faulty nodes, and α is a penalty weight.)
Calculating according to; And
If it is determined that the global reliability agreement is unsuccessful, the local reliability for each of the verification nodes is expressed by the equation.

Calculating according to; A method of consensus tolerant of a block chain network comprising a. The method of claim 3, wherein the step of recalculating the local reliability is
Consensus Bound is based on the global reliability according to the success or failure of the global reliability agreement.

A method of consensus to allow for the failure of a blockchain network that varies depending on. The method of claim 1, wherein the propagating the decision message is
Analyzing a preparation message received by each of the verification nodes, and determining whether the number of local reliability values having the same value is greater than or equal to a predetermined reference number;
If more than the reference number, reflecting the local reliability of the same value in the global reliability of the current round;
If it is less than the reference number, reflecting the backup local reliability to the global reliability of the current round; And
Determining whether the sum of global reliability of verification nodes having the same local reliability value is equal to or greater than a predetermined global reference value, and if it is equal to or greater than a global reference value, propagating a decision message to another verification node; A method of consensus tolerant of a block chain network comprising a.

Images