A Reputation-Based Blockchain Secure Sharding Method and System

Abstract

This invention relates to a blockchain secure sharding method and system based on a reputation mechanism. The method includes dividing the network into multiple shards, establishing a quantifiable dynamic reputation model for each node and assigning an initial reputation value to each node; electing coordinator nodes within each shard to propose new blocks; introducing reputation-weighted voting in the BFT consensus to allow high-reputation nodes to quickly confirm blocks; evaluating and updating the reputation value of coordinator nodes based on their block production performance after each consensus round; and periodically distributing nodes with similar reputation values ​​evenly across different shards to prevent malicious node aggregation. This invention structurally prevents low-reputation nodes from aggregating within a single shard, thereby fundamentally improving robustness and security against single-shard takeover attacks and achieving a balance between high scalability, high security, and low overhead.

Description

Technical Field

[0001] This invention relates to the field of blockchain technology, and in particular to a blockchain secure sharding method and system based on a reputation mechanism. Background Technology

[0002] As a distributed ledger technology, blockchain has been widely used in fields such as digital currencies like Bitcoin and Ethereum, and the Internet of Things due to its decentralized, immutable, and traceable characteristics. However, the performance and scalability of blockchain are the core bottlenecks to its development. Under the decentralized architecture, traditional blockchain systems require all nodes in the network to verify, store, and broadcast all transactions, which severely limits the system throughput. For example, the Bitcoin network can only process about 7 transactions per second. Compared with Visa's centralized payment system, which can process thousands of transactions per second, the performance gap makes it difficult for blockchain to meet the needs of high-frequency transaction scenarios.

[0003] To address this scalability issue, the industry has proposed numerous technical solutions, including off-chain channels, sidechains, DAGs, and sharding. Among these, sharding is considered a highly promising core technology due to its ability to achieve "horizontal scaling" at the Layer 1 level. Sharding logically divides the entire blockchain network into multiple smaller network clusters (i.e., "shards"), allowing transactions to be processed in parallel within each shard. Each shard stores only a subset of its own state. Theoretically, this parallel processing enables the system's total transaction throughput to increase linearly with the number of nodes.

[0004] Based on differences in consensus mechanisms, existing sharding technologies can be categorized as follows:

[0005] Proof-of-Work (PoW) based sharding: such as Elastico and Monoxide, which use a PoW election committee, offer high security but consume a lot of energy;

[0006] Local Byzantine Fault Tolerant (BFT) Sharding: Such as OmniLedger, RapidChain, and Zilliqa, which use PBFT or its variants, offer fast transaction confirmation and low energy consumption, but have an inherent security weakness, namely low fault tolerance: Under the BFT framework, if more than one-third of the nodes in a single shard are malicious nodes, these malicious nodes can completely control the consensus result of that shard, and then launch a so-called "single shard takeover attack". This lack of fault tolerance limits the security and throughput of the system.

[0007] To improve the error tolerance of BFT consensus, an effective supplement or alternative is proposed—the reputation mechanism (PoR). The reputation mechanism establishes trust by evaluating the past behavior of participants. Its application in consensus includes: (1) only a subset of high-reputation nodes participate in the consensus process; (2) conducting reputation-based weighted voting, i.e., the voting weight of a node is proportional to its reputation value.

[0008] Some studies have attempted to combine reputation mechanisms with sharding systems. For example, Repchain uses independent BFT consensus and Raft consensus to determine the reputation chain and the transaction chain. However, in this approach, the evaluation of reputation and the processing of transactions are separated, requiring two independent consensus processes. This not only causes excessive system overhead and unsatisfactory throughput, but also most solutions only update reputation within a fixed period, failing to reflect changes in node behavior in real time, resulting in a lag in reputation updates. More importantly, this approach does not directly apply the node's reputation value to the consensus decision-making process within the shard, failing to fully leverage the potential of the reputation mechanism to improve the security within the shard.

[0009] In addition, to prevent nodes within a shard from being corrupted over a long period of time, the sharding system needs to periodically refactor the shards, that is, to redistribute the nodes within each shard. Most of the current mainstream refactoring algorithms use randomization methods, such as Verifiable Random Functions (VRF). However, such random algorithms not only bring additional computational overhead to the system, but also have unstable refactoring effects. They cannot guarantee from a mechanism perspective that they can effectively disperse malicious nodes, and under random circumstances, malicious nodes may still gather in a certain shard. Summary of the Invention

[0010] In view of this, the present invention provides a blockchain secure sharding method and system based on reputation mechanism, which deeply integrates reputation mechanism into the consensus protocol within the shard to effectively improve the error tolerance and consensus efficiency of a single shard, and enhance the security and robustness of the entire system.

[0011] To achieve the above objectives, the present invention provides a blockchain secure sharding method based on a reputation mechanism, comprising the following steps:

[0012] S1, Network Sharding: The blockchain network is divided into S shards, and each shard contains a group of nodes;

[0013] S2, Reputation Initialization: Assign an initial reputation value to each newly added node in the shard. ;

[0014] S3. Perform reputation consensus within each shard;

[0015] S301. Elect a coordinator node within each shard to propose new blocks;

[0016] S302, The coordinator node broadcasts the proposed new block to other nodes within the shard;

[0017] S303, When the node receives at least A response message or received A number exceeding the preset reputation threshold A high-reputation node responds to the message, and the node submits a block;

[0018] S4. Evaluate and update the reputation value of the coordinator node based on its block production performance.

[0019] S5. Perform shard reconstruction according to the preset cycle, distribute nodes with similar reputation values ​​to different shards, and optimize the node composition within the shards.

[0020] Preferably, the coordinator node is elected using a random method within the shard. node The probability of being elected as the coordinator is :

[0021] ;

[0022] in, Indicates within a fragment node .

[0023] Preferably, when the system enters a stable period, the coordinator node is elected in rotation among trusted nodes with a reputation value that is not zero.

[0024] Preferably, evaluating and updating the reputation value of the coordinator node includes the following steps:

[0025] S401, Configure the monitor node Monitor each coordinator node The behavior is evaluated using performance indicators. Represents the coordinator node The regularity of each block production is expressed as:

[0026] ;

[0027] in, Represents the coordinator node The correct number of blocks produced. Represents the coordinator node The number of times a node has been elected as the coordinator;

[0028] S402. The reputation value of the coordinator node is evaluated using the sigmoid function, wherein the expression of the sigmoid function is:

[0029] ;

[0030] in, , The parameters of the sigmoid function control the slope and the growth point, respectively. The sigmoid function, representing the input value, limits the coordinator node's reputation value to between 0 and 1; input value The expression is:

[0031] ;

[0032] in, This represents the growth rate parameter.

[0033] Preferably, the evaluation and updating of the reputation value of the coordinator node also includes incentive and penalty mechanisms;

[0034] The incentive mechanism is to grant block rewards to coordinator nodes that successfully produce blocks and nodes with high reputation values ​​that participate in consensus in a manner proportional to their own reputation values.

[0035] The penalty mechanism reduces the reputation value of a node that has been elected as a coordinator three times in a row but has failed to produce a block to 0.

[0036] Preferably, the penalty mechanism also includes limiting the upper limit of a node's reputation value for intermittently publishing erroneous blocks.

[0037] Preferably, the fragment reconstruction includes the following steps:

[0038] Integrate nodes with the same reputation value in the shards and divide them into different reputation groups;

[0039] Nodes are uniformly extracted from each of the aforementioned reputation groups and redistributed to different new shards.

[0040] This invention also provides a blockchain security sharding system based on a reputation mechanism, comprising the following modules:

[0041] Reputation assessment module: Assigns and updates reputation value to each node, sets the reputation value of nodes that continuously engage in malicious behavior to zero, and limits the reputation value of nodes that indirectly engage in malicious behavior. The reputation value is determined based on the block production regularity and number of blocks when the node acts as a coordinator.

[0042] Consensus module: Sets consensus parameters for each shard, including reputation threshold. and high-value response threshold The consensus process is executed within each shard, and whether to submit a block is determined based on the total number of response messages received and the number of response messages from high-reputation nodes.

[0043] Shard Reconstruction Module: According to a preset period, all nodes are redistributed, distributing nodes with similar reputation values ​​to different shards.

[0044] Compared with the prior art, the beneficial effects of the present invention are:

[0045] This invention organically integrates dynamic reputation assessment, reputation-weighted consensus, and deterministic sharding reconstruction, breaking through the triangular contradiction between performance, security, and overhead in traditional sharding systems; it introduces reputation values ​​and sets up a reasonable incentive mechanism to ensure that honest nodes stably serve the system.

[0046] This invention applies a reputation-based consensus mechanism in sharding, relying on a small number of honest nodes to quickly confirm blocks, in order to make up for the problems of low error tolerance and insecurity in traditional sharding systems, while improving consensus speed.

[0047] The fast fragmentation and reassembly method based on reputation value adopted in this invention aims to improve the security within a single fragment and prevent the aggregation of malicious nodes.

[0048] Compared to traditional BFT sharding systems, this invention improves throughput and scales linearly with the number of nodes and shards. In terms of communication and computational overhead, the sharding reconstruction complexity provided by this invention is significantly reduced compared to existing random algorithms. Reduce to The reconstruction latency increases only linearly and slowly with scale, which is significantly better than VRF-type solutions. At the same time, by implementing a reward and punishment mechanism to maintain high-reputation nodes and quickly isolate malicious nodes, the system maintains high robustness and low maintenance costs throughout its life cycle, truly achieving a unity of high scalability, high security, and low overhead. Attached Figure Description

[0049] Figure 1 This is a schematic diagram of the process of a blockchain secure sharding method based on a reputation mechanism according to the present invention;

[0050] Figure 2 This is a schematic diagram of the overall architecture of a reputation-based blockchain security sharding method according to the present invention.

[0051] Figure 3 This is a schematic diagram of the consensus process of the present invention;

[0052] Figure 4 This is a schematic diagram of the fragment reconstruction process in an embodiment of the present invention;

[0053] Figure 5 This is a schematic diagram illustrating the scalability evaluation in an embodiment of the present invention;

[0054] Figure 6 This is a schematic diagram comparing the time consumption of fragmented reconstruction in an embodiment of the present invention;

[0055] Figure 7 This is a schematic diagram comparing throughput in an embodiment of the present invention;

[0056] Figure 8 This is a schematic diagram of delay comparison in an embodiment of the present invention;

[0057] Figure 9 This is a schematic diagram illustrating the reputation value changes of different behavioral nodes in an embodiment of the present invention;

[0058] Figure 10 This is a schematic diagram comparing attack prevention capabilities in embodiments of the present invention. Detailed Implementation

[0059] To further illustrate the technical means and effects of the present invention in achieving its intended purpose, the following detailed description of the specific implementation methods, structures, features, and effects of the present invention, in conjunction with the accompanying drawings and preferred embodiments, is provided below.

[0060] This invention aims to resolve the inherent contradiction between performance, security, and overhead in existing sharding technologies. Its core lies in constructing a trust framework integrating dynamic evaluation, weighted consensus, and deterministic reconstruction. First, a quantifiable dynamic reputation model is established for network nodes using a non-linear growth function based on block production regularity and block count, accurately reflecting their long-term behavior. Second, a reputation-weighted consensus mechanism coupled with this model is proposed. This mechanism introduces the concept of "high-value responses" on top of BFT consensus, allowing consensus to be reached by confirmation from a specified number of high-reputation nodes, thereby significantly reducing communication overhead while ensuring security. The key innovation of this invention is a reputation-based deterministic sharding reconstruction protocol. This protocol periodically groups nodes with similar reputation values ​​and systematically distributes them across different shards. This aims to structurally prevent low-reputation (potentially malicious) nodes from aggregating within a single shard, thus fundamentally improving robustness against single-shard takeover attacks.

[0061] Example 1

[0062] Existing sharding technologies inherently conflict between performance, security, and overhead. This embodiment constructs a trust framework integrating dynamic evaluation, weighted consensus, and deterministic reconstruction to improve security within a single shard and prevent malicious node aggregation. This embodiment also provides a reputation-based blockchain secure sharding method, such as... Figure 1 As shown, the specific steps include:

[0063] S1, Network Sharding: The blockchain network is divided into S shards, and each shard contains a group of nodes;

[0064] S2, Reputation Initialization: Assign an initial reputation value to each newly added node in the shard. Initial reputation value As indicated by the community where the node is located, the system enters a stable period, and the coordinator node is elected through a rotation among trusted nodes with a reputation value that is not zero.

[0065] S3. Perform reputation consensus within each shard;

[0066] S301. Within each shard, a coordinator node is elected to propose new blocks. The coordinator node is elected using a random method within the shard. node The probability of being elected as the coordinator is :

[0067] ;

[0068] in, Indicates within a fragment node ;

[0069] S302, The coordinator node broadcasts the proposed new block to other nodes within the shard;

[0070] S303, When the node receives at least A response message or received A number exceeding the preset reputation threshold A high-reputation node responds to the message, and the node submits a block;

[0071] S4. Evaluate and update the coordinator node's reputation value based on its block production performance, including the following steps:

[0072] S401, Configure the monitor node Monitor each coordinator node The behavior is evaluated using performance indicators. Represents the coordinator node The regularity of each block production is expressed as:

[0073] ;

[0074] in, Represents the coordinator node The correct number of blocks produced. Represents the coordinator node The number of times a node has been elected as the coordinator;

[0075] S402. The sigmoid function is used as a model to evaluate the reputation value of the coordinator node, controlling its reputation growth. The expression for the sigmoid function is:

[0076] ;

[0077] in, , The parameters of the sigmoid function control the slope and the growth point, respectively. The sigmoid function, representing the input value, limits the coordinator node's reputation value to between 0 and 1; the reputation growth function takes the input value as its input value. The expression is:

[0078] ;

[0079] in, The growth rate parameter, in addition to the internal reputation value evaluated by the sigmoid function, also needs to consider the coordinator node's external reputation value. Therefore, the coordinator node's final reputation update is:

[0080] ;

[0081] The evaluation and updating of the reputation value of the coordinator node also includes incentive and penalty mechanisms;

[0082] The incentive mechanism rewards coordinator nodes that successfully mine blocks and nodes with high reputation values ​​that participate in consensus with block rewards proportional to their own reputation values. This is to encourage all honest coordinator nodes to continue mining blocks correctly in pursuit of higher reputation values. Each time a coordinator node mines a block that is recognized by nodes within the sharded network, it receives a portion of the fixed reward contained in the block, which is proportional to the coordinator node's reputation value. Unlike Bitcoin's incentive mechanism, in this embodiment, nodes with high reputation values ​​in the network also receive another portion of the block reward after consensus is reached. The distribution of this reward is related to the high reputation value of the node itself. A higher reputation value means that the node will receive a higher reward after consensus is reached. This ensures that high reputation nodes remain honest throughout the long-term operation of the system, making attackers more susceptible to corruption.

[0083] The penalty mechanism reduces the reputation value of a node that has been elected as a coordinator three times in a row but has failed to produce a block to 0, and limits the upper limit of the reputation value of a node that intermittently publishes erroneous blocks. In this embodiment, it is set to the middle value of the reputation value range, which is 0.5.

[0084] S5. Perform shard reconstruction according to a preset cycle, distributing nodes with similar reputation values ​​into different shards to optimize the node composition within each shard; shard reconstruction includes the following steps:

[0085] Integrate nodes with the same reputation value in the shards and divide them into different reputation groups;

[0086] Nodes are uniformly extracted from each of the aforementioned reputation groups and redistributed to different new shards;

[0087] The period from the end of each round of shard reconstruction to the start of the next round is called a period. After each period, malicious nodes may aggregate in a single shard. Nodes with the same reputation value in each shard are selected and these node sets are assigned to different shards during resharding. Considering that the proportion of malicious nodes in other shards will increase, even honest nodes with the same reputation value will undergo the same process. This is because the security of the sharding system must take into account the security of all shards and ensure the robustness of each shard.

[0088] Example 2

[0089] This embodiment provides a blockchain security sharding system based on a reputation mechanism, which divides the entire blockchain network into segments. Each segment: Each shard contains a set of nodes, a coordinator, and a monitor; the system includes a reputation assessment module, a consensus module, and a shard reconstruction module.

[0090] Reputation assessment module: Assigns and updates reputation value to each node, sets the reputation value of nodes that continuously engage in malicious behavior to zero, and limits the reputation value of nodes that indirectly engage in malicious behavior. The reputation value is determined based on the block production regularity and number of blocks when the node acts as a coordinator.

[0091] Consensus module: Sets consensus parameters for each shard, including reputation threshold. and high-value response threshold The consensus process is executed within each shard, and whether to submit a block is determined based on the total number of response messages received and the number of response messages from high-reputation nodes.

[0092] Shard Reconstruction Module: According to a preset period, all nodes are redistributed, distributing nodes with similar reputation values ​​to different shards;

[0093] System initialization and initial shard configuration: In a permissioned chain network environment, the initialization in this embodiment includes the following steps:

[0094] Node and Shard Initialization: Assuming there are n nodes in the network and the plan is to divide them into s s shards, the system initializes all nodes by shuffling the node order using a random function and then randomly assigning the nodes to the s s s shards in sequence to ensure the fairness of the initial allocation.

[0095] Reputation and parameter settings: When each node is assigned to a shard, it is given an initial reputation value set by the community or system to which the node belongs. At the same time for each fragment Assign a trusted monitor node And set the consensus parameters within that shard, including the reputation threshold. High-value response threshold In this embodiment , ;

[0096] Initial coordinator node election: In the early stages of system operation, since the reputation values ​​of nodes have not yet formed an effective distinction, a random method is used to elect coordinator nodes. The system uses a random number generator to randomly elect a node for each shard as the coordinator for this round of consensus.

[0097] Reputation-based consensus within shards, such as Figure 2 As shown, when the number of broadcast transactions within a shard reaches a certain level, the monitor node... The consensus process is initiated by combining PBFT voting and reputation-based weighted voting, and specifically includes the following steps:

[0098] REQUEST: Sharding monitor node Initiate a consensus request to all nodes within the shard;

[0099] PROPOSAL: After receiving the request, the elected coordinator node packages the transaction into a block and encapsulates it in the PROPOSAL message, broadcasting it to all other nodes in the shard.

[0100] RESPONSE response: After receiving the PROPOSAL message, other nodes verify the legality of the block. If the verification is successful, they forward the RESPONSE response message, indicating their agreement with the proposal, to other nodes.

[0101] Submitting a COMMIT: When a node receives at least A response message or received A number exceeding the preset reputation threshold High-reputation nodes respond to messages to monitor nodes. Send a COMMIT message;

[0102] Block on-chain: Monitor node Upon receiving A response message or After receiving a high-value response message, the legality of the block is confirmed and all nodes in the shard are notified to put the block on the chain.

[0103] Reputation assessment and updates occur after each round of consensus (or after each view transformation), via monitor nodes. The system will conduct a reputation assessment on the coordinator nodes of this round of consensus. The reputation assessment module sets up monitor nodes to monitor the coordinator nodes, assigns and updates reputation values ​​for each node, sets the reputation value of nodes that continuously engage in malicious behavior to zero, and restricts the reputation value of nodes that indirectly engage in malicious behavior. The reputation value is determined based on the block production regularity and number of blocks produced when the node acts as a coordinator.

[0104] Trusted sharding process as follows Figure 3 As shown, where, This represents the reputation value of a node. There are a total of 4 nodes participating in the consensus. Let's assume a reputation threshold for the consensus. A value of 0.6 represents the number of valuable responses required to satisfy the consensus requirements. The value is 2, which means that in this consensus, there are two suspected malicious nodes, node 1 and node 4. Even if node 1 and node 4 do not forward messages, the consensus can still be completed correctly, and the number of communications within the sharded network is reduced, which has a good effect on reducing the communication burden in PBFT. In addition, even when there are few malicious nodes in the shard, TrustShard can achieve consensus quickly and submit blocks, thereby improving the overall throughput of the system.

[0105] Fragmented reconstruction, such as Figure 4 As shown, when the system enters a stable period and the reputation value of a node is sufficiently representative, the system will periodically (in this embodiment, every cycle e) execute the sharding reconstruction protocol, specifically including the following process:

[0106] Stable period coordinator election: During the stable period, the coordinator election method is changed from random to round-robin election among all nodes with non-zero reputation values, to ensure that each node has the opportunity to participate in consensus and be evaluated;

[0107] Reputation Group Integration: At the start of the reconstruction, the system integrates all shard nodes and their latest reputation values, classifying nodes with the same reputation value across the entire network into the same "Reputation Group" (GRP). Attackers' attack behavior often leads to them having similar low reputation distributions, so this step can effectively identify potential attacker groups.

[0108] Reconfiguration and allocation: The system creates s new empty shards, iterates through each reputation group, takes out nodes one by one from the group, and allocates these nodes to different new shards in a round-robin manner; this process ensures that nodes with the same reputation value (especially low reputation value) are deliberately distributed to different shards, which greatly reduces the risk that the proportion of malicious nodes in a single shard exceeds the fault tolerance threshold. This process is also applied to honest nodes to ensure the security of all shards.

[0109] State migration: After a node is assigned to a new shard, its historical block production data and the number of times it has been elected coordinator, as well as other state information, are also migrated to the new shard to provide data for subsequent reputation assessment.

[0110] The cyclic execution in this embodiment can improve consensus and security in a sharded blockchain system by leveraging a reputation mechanism, and effectively resist single-shard takeover attacks through a reputation-based reconstruction algorithm, ultimately achieving a high-throughput, highly scalable, and robust blockchain system.

[0111] Example 3

[0112] This embodiment presents a series of simulation experiments on a blockchain secure sharding method and system based on a reputation mechanism. The simulation experiments are conducted in a network environment with 100 node instances, a default sharding size of 10, and each block containing 100 transactions. The core parameters of the reputation growth algorithm are set as follows: , , The core parameter of the consensus algorithm is set to the reputation threshold by default. High-value response threshold Inter-node communication is simulated using Python's multi-threaded module and UDP-based socket programming, with the sharding reconstruction cycle set to 10 rounds of consensus.

[0113] like Figure 5 As shown, this implementation evaluates the scalability of the TrustShard security sharding system by observing the changes in the total throughput (Transactions per Second, tps) of the system by changing the number of shards (3, 5, 7, 9, 11). The experimental results show that as the number of shards (i.e. the total number of nodes) increases, the system throughput exhibits an approximately linear growth trend. This embodiment has good scalability through the mechanism of parallel transaction processing by sharding.

[0114] like Figure 6 The fragmentation reconstruction latency analysis shown in this embodiment evaluates the time consumption of a reputation-based fragmentation reconstruction algorithm and compares it with the OmniLedger scheme using a verifiable random function (VRF-based). As the number of fragments increases from 5 to 9, the latency of the secure fragmentation method provided in this embodiment increases linearly and slowly, while the latency of OmniLedger increases sharply. At 9 fragments, the latency difference between the two reaches 7 seconds. This is because the VRF and inter-node gossip communication used by OmniLedger introduce latency issues. The complexity of [previous method] is much higher, while the secure fragmentation method provided in this embodiment has a complexity of only [other complexity]. The secure fragmentation method provided in this embodiment has a significant advantage in efficiency.

[0115] like Figure 7 As shown, this embodiment compares the throughput of TrustShard with non-sharding BFT systems and traditional BFT-sharding systems. The comparisons were conducted with 50, 60, 70, and 80 nodes. The results show that the throughput of the non-sharding system decreases sharply with the number of nodes, reaching only 384 tps with 50 nodes. While the BFT-sharding system achieves linear growth, its throughput of approximately 15 ktps with 80 nodes is still lower than TrustShard's 17.5 ktps. This is because consensus completion is no longer mandatory in TrustShard after it enters a stable period. One reply, only The system can achieve consensus by quickly responding to a high-reputation node, thereby shortening the consensus time and increasing the system throughput.

[0116] like Figure 8 As shown, this embodiment compares the confirmation latency experienced by TrustShardyu and the BFT sharding system from block publication to commit. This embodiment is conducted with shard sizes of 5, 10, 15, and 20. The results show that although the latency of both systems increases linearly with the number of nodes in the shard, the confirmation latency of TrustShardyu is significantly lower than that of the BFT sharding system under all settings. This is also due to the fact that high-reputation nodes can quickly confirm blocks, thus bringing an advantage in confirmation interval.

[0117] like Figure 9 As shown, this embodiment verifies the effectiveness of the reputation calculation algorithm of the present invention. This embodiment simulates the reputation value changes of honest nodes, intermittent malicious nodes, and continuous malicious nodes during 20 elections as coordinators. The reputation value changes of honest nodes are as expected, that is, they grow slowly in the early stage, grow rapidly during the 5th to 12th evaluations, and then the growth rate slows down and approaches 1. The reputation value of intermittent malicious nodes decreases after the 3rd, 5th, and 7th malicious acts, and even if they stop malicious acts later, their reputation value cannot break through the upper limit of 0.5. After two consecutive malicious acts, the reputation value of continuous malicious nodes is directly reduced to 0 when they commit malicious acts for the third time, and they are excluded from the coordinator candidate range. The experimental results show that the reputation mechanism of this embodiment can effectively incentivize honest behavior and impose effective penalties on different modes of attack behavior, significantly increasing the attack cost for attackers.

[0118] like Figure 10As shown, this embodiment evaluates TrustShard's ability to resist single-shard takeover attacks under different attacker proportions and compares it with the BFT sharding system. With 100 nodes and 10 shards, the BFT allocation system shows a consensus flip probability as high as 50% when the attacker proportion reaches 15%. In contrast, TrustShard's security can be adjusted... Values ​​are controlled when When set to 4 or 3, the attack success rate can be controlled within 10%, which proves that the high-reputation nodes and adjustable [functions] used in this invention [are effective]. The consensus security of the parameters is significantly better than that of the traditional BFT sharding system.

[0119] Experimental results show that this scheme can improve throughput by nearly 27% compared with the traditional BFT sharding system, and can control the consensus reversal probability to within 30% with a 30% attacker ratio, providing an effective and low-overhead technical path for building a highly scalable and secure blockchain system.

[0120] The above description is merely a preferred embodiment of the present invention and is not intended to limit the present invention in any way. Although the present invention has been disclosed above with reference to preferred embodiments, it is not intended to limit the present invention. Any person skilled in the art can make some modifications or alterations to the above-disclosed technical content to create equivalent embodiments without departing from the scope of the present invention. Any simple modifications, equivalent changes and alterations made to the above embodiments based on the technical essence of the present invention without departing from the scope of the present invention shall still fall within the scope of the present invention.

Claims

1. A blockchain secure sharding method based on a reputation mechanism, characterized in that, Includes the following steps: S1, Network Sharding: The blockchain network is divided into S shards, and each shard contains a group of nodes; S2, Reputation Initialization: Assign an initial reputation value to each newly added node in the shard. ; S3. Perform reputation consensus within each shard; S301. Elect a coordinator node within each shard to propose new blocks; S302, The coordinator node broadcasts the proposed new block to other nodes within the shard; S303, When the node receives at least A response message or received A number exceeding the preset reputation threshold A high-reputation node responds to the message, and the node submits a block; S4. Evaluate and update the reputation value of the coordinator node based on its block production performance. S401, Configure the monitor node Monitor each coordinator node The behavior is evaluated using performance indicators. Represents the coordinator node The regularity of each block production is expressed as: in, Represents the coordinator node The correct number of blocks produced. Represents the coordinator node The number of times a node has been elected as the coordinator; S402. The reputation value of the coordinator node is evaluated using the sigmoid function, wherein the expression of the sigmoid function is: in, , The parameters of the sigmoid function control the slope and the growth point, respectively. The sigmoid function, representing the input value, limits the coordinator node's reputation value to between 0 and 1; input value The expression is: in, Indicates the growth rate parameter; S5. Perform shard reconstruction according to the preset cycle, distribute nodes with similar reputation values ​​to different shards, and optimize the node composition within the shards.

2. The blockchain secure sharding method based on a reputation mechanism according to claim 1, characterized in that, The coordinator node is elected using a random method within the shard. node The probability of being elected as the coordinator is : in, Indicates within a fragment node .

3. The blockchain secure sharding method based on a reputation mechanism according to claim 2, characterized in that, Once the system enters a stable period, the coordinator node will be elected through a rotation among trusted nodes with a reputation value that is not zero.

4. The blockchain secure sharding method based on a reputation mechanism according to claim 1, characterized in that, The evaluation and updating of the reputation value of the coordinator node also includes incentive and penalty mechanisms; The incentive mechanism is to grant block rewards to coordinator nodes that successfully produce blocks and nodes with high reputation values ​​that participate in consensus in a manner proportional to their own reputation values. The penalty mechanism reduces the reputation value of a node that has been elected as a coordinator three times in a row but has failed to produce a block to 0.

5. A blockchain secure sharding method based on a reputation mechanism according to claim 4, characterized in that, The penalty mechanism also includes limiting the upper limit of a node's reputation value for intermittently publishing erroneous blocks.

6. A blockchain secure sharding method based on a reputation mechanism according to claim 1, characterized in that, The fragment reconstruction includes the following steps: Integrate nodes with the same reputation value in the shards and divide them into different reputation groups; Nodes are uniformly extracted from each of the aforementioned reputation groups and redistributed to different new shards.

7. A blockchain secure sharding system based on a reputation mechanism, characterized in that, Includes the following modules: Reputation assessment module: Assigns and updates reputation value to each node, sets the reputation value of nodes that continuously engage in malicious behavior to zero, and limits the reputation value of nodes that indirectly engage in malicious behavior. The reputation value is determined based on the block production regularity and number of blocks when the node acts as a coordinator. Configure monitor node Monitor each coordinator node The behavior is evaluated using performance indicators. Represents the coordinator node The regularity of each block production is expressed as: in, Represents the coordinator node The correct number of blocks produced. Represents the coordinator node The number of times a node has been elected as the coordinator; The reputation value of the coordinator node is evaluated using the sigmoid function, the expression of which is: in, , The parameters of the sigmoid function control the slope and the growth point, respectively. The sigmoid function, representing the input value, limits the coordinator node's reputation value to between 0 and 1; input value The expression is: in, Indicates the growth rate parameter; Consensus module: Sets consensus parameters for each shard, including reputation threshold. and high-value response threshold The consensus process is executed within each shard, and whether to submit a block is determined based on the total number of response messages received and the number of response messages from high-reputation nodes. Shard Reconstruction Module: According to a preset period, all nodes are redistributed, distributing nodes with similar reputation values ​​to different shards.

Images