A High-Performance Sharding Consortium Chain Autonomous Evolution Design Method

Abstract

A high-performance sharded consortium blockchain autonomous evolution design method is proposed. This method utilizes blockchain to provide a platform for cross-organizational co-construction and sharing of 6G networks, and introduces concepts such as reputation systems, node evaluation, and master-slave game theory to empower sharded blockchains, achieving a triangular balance of scalability, security, and decentralization that meets application requirements. First, this method coordinates the three dilemmas of blockchain based on the throughput and security requirements of actual applications, designing a sharded blockchain network that meets application needs. Then, since the blockchain network will be built on physical equipment used by network operators and equipment providers for co-constructing the 6G network, this method provides dynamic scoring for the massive amounts of physical equipment provided by different organizations through reputation systems and node evaluation. Finally, through master-slave game theory between shard leader nodes and replica nodes, nodes participating in consensus in the blockchain network are mapped to the optimal physical equipment based on the physical equipment scores.

Description

Technical Field

[0001] This invention relates to the field of mobile communication technology, specifically to a high-performance sharded consortium blockchain autonomous evolution design method. Background Technology

[0002] The Internet of Things (IoT) era is one where everything is interconnected, with a vast number of devices connecting to the network. These devices belong to different industrial sectors and have different characteristics and needs. 5G and IoT coexist, realizing a shift from communication-oriented to service-oriented architecture. 6G, building upon 5G, will fully support the digitization of the entire physical world and the virtualization of resources, achieving ubiquitous sharing of digital resources. Therefore, future 6G networks need to integrate a centralized mobile communication network architecture with an open IoT architecture, moving towards a new architecture that combines openness and distributed control.

[0003] Compared to 5G networks, which are jointly built and shared by operators, 6G networks require more base stations, which in turn requires huge investments. Higher communication frequency bands place higher demands on the number of base stations, while larger network capacity requires smaller cell radii. All of these mean that the construction of 6G networks requires an unprecedented amount of resources.

[0004] In traditional networks, network infrastructure, such as macro base stations (MBS), small base stations (SBS), and edge computing nodes (MEC), which are usually provided and maintained by network operators, may belong to different operators, enterprises, organizations, or individuals in 6G networks. The same equipment may also be jointly built and shared by multiple participants. Such cooperation methods lead to a large number of participants, complex costs, and a complex situation of multi-party and multi-level settlement. Specifically, the current multi-organizational cooperation in 6G co-construction and sharing will face the following difficulties: (1) Difficulty in sharing basic data: The base station energy consumption costs and leasing costs involved are all stored in the internal systems of each company, and the data affecting settlement costs are not interconnected; (2) Difficulty in coordinating cross-company settlements: It involves multi-party and multi-level settlements, making it difficult to reach a consensus on cost settlement, and the settlement process is complicated; (3) Difficulty in tracing business scenarios: Cross-company settlements, business lines are not connected, multiple entities maintain their own ledgers, and the settlement process is difficult to trace. Therefore, a new trust and security management solution is needed to ensure the safe and reliable sharing of resources and the secure flow of data and privacy protection.

[0005] Like most real-world networks, blockchain systems are open, meaning new nodes are constantly joining the network as it operates. In conventional consortium blockchains, consensus algorithms such as BFT or CFT are typically used. An excessively large number of nodes can significantly reduce transaction throughput. Furthermore, most existing work assumes that the nodes joining are identical; however, in real-world systems, nodes exhibit diverse differences in computing power, communication bandwidth, and historical behavior. Weaker nodes can easily become bottlenecks in the blockchain system's performance. Therefore, a method is needed to select an appropriate number of secure and powerful nodes to meet the functional requirements of the blockchain system, balancing scalability, security, and decentralization to form a suitable blockchain network. Summary of the Invention

[0006] To address the problems existing in the aforementioned background technologies, this invention proposes a high-performance sharded consortium blockchain autonomous evolution design method. It uses blockchain to provide a platform for cross-organizational co-construction and sharing of 6G networks, and introduces concepts such as reputation systems, node evaluation, and master-slave game theory to empower sharded blockchains, achieving a triangular balance of scalability, security, and decentralization that meets application requirements.

[0007] A high-performance sharded consortium blockchain autonomous evolution design method includes the following steps:

[0008] Step S1: Based on the application's security and scalability, construct a blockchain to meet these two requirements. Calculate the probability of an attacker's successful attack and obtain constraints on the number of nodes in each shard based on the attack cycle and the total number of nodes in the blockchain network. Calculate the total number of blockchain nodes, shards, and shard nodes that meet the actual application requirements.

[0009] Step S2: In order to map virtual elements in the blockchain virtual network to appropriate physical elements, resource value is used to score node resources; the evaluation of resource value is divided into two parts: scalability score and trustworthiness score.

[0010] Step S3: Conduct consensus node election, select trustworthy nodes from all nodes based on their trustworthiness scores to participate in the subsequent election game and consensus work;

[0011] Step S4, Leader Selection and Sharding: Nodes that meet the scalability score and intend to become leaders issue a declaration, and the remaining nodes will select suitable leaders to join their shards. The leader node comprehensively considers the scalability score and the credibility score to select the best replica node to form a shard.

[0012] Furthermore, step S1 specifically involves the following steps:

[0013] Step S1.1: Calculate the probability of a successful attack by the attacker; N is the total number of nodes in the system. If the attacker attacks r nodes simultaneously, and the attack speed is t... a The time interval between initiating a transaction request and the time interval between initiating verification are both t. d If it follows an exponential distribution with parameter λ, then during attack period i, we have: A number of nodes are compromised, where Y is the number of nodes recovered in each attack cycle, and the probability of a successful attack by the attacker is [missing information].

[0014] Step S1.2, if required and When obtaining the number of nodes n in each shard and the total number of nodes |N| in the blockchain, the constraints are as follows: For any attack period M, the following holds true:

[0015]

[0016]

[0017] Step S1.3: Calculate the total number of blockchain nodes, the number of shards, and the number of shard nodes to meet the scalability requirements of the actual application. The application provides scalability requirements: transaction throughput ≥ TPS0, transaction latency ≤ T0. Based on the transaction throughput requirements, calculate the corresponding total number of blockchain nodes N≥N0, the optimal number of shards S, and the number of nodes n in each shard. This is the set of arrays containing the total number of blockchain nodes, the number of shards, and the number of shard nodes that meet the conditions.

[0018] Where j is the number of sets that satisfy the condition.

[0019] Each array element in set S corresponds to an average transaction latency T. Elements with T ≤ T0 are selected to form a new array, and the set of the new array is...

[0020] S k ={[N1,S1,n1,T1],[N2,S2,n2,T2],……,[N k ,S k ,n k ,T k ]|T k ≤T0}

[0021] Where k is the number of sets that satisfy the time delay condition.

[0022] Furthermore, step S2 specifically involves the following steps:

[0023] Step S2.1: For the leader node, the scalability score includes the node's local capabilities and node importance; for replica nodes, the scalability score includes the node's local capabilities and distance from the leader node; the node's local capabilities are as follows:

[0024]

[0025] Where CPU(i) and Stor(i) represent the computing power and storage power of node i, respectively, ij is the link directly connected to node i, and Band(ij) represents the bandwidth of the link. max Stor max and Band max These are the maximum values ​​of computing power, storage capacity, and bandwidth, respectively; if A(i)>σ1, where σ1 is the capacity threshold, which can be set according to the capacity requirements of the node, then it means that node i serves as the shard leader node;

[0026] Step S2.2: For nodes that meet the requirements to serve as the leader node of a shard, their importance is measured using complex network structure characteristic metrics, namely the degree, betweenness, and clustering coefficient of the node.

[0027] Step S2.3: Since the local capabilities required for sharded replica nodes are relatively low, nodes that have not been selected as leader nodes are chosen. In addition, the selection of sharded replica nodes focuses on the distance from the leader node, i.e., the number of hops.

[0028] Record S shard For the set of all partitions, a partition s j ∈S shard During the shard leader node's selection of shard replica nodes, it searches for nodes with a smaller hop distance h from itself. i:j Let i be the shortest number of hops required to get from node i to node j, and calculate its score accordingly.

[0029]

[0030] Step S2.4: Each node scores the credibility of the other nodes, using subjective logic to represent the quadruple of opinion: ω y:x ={b y:x ,d y:x ,u y:x ,a y:x Trust level b y:x distrust level d y:x These represent the proportions of historical benign and undesirable behaviors among all behaviors, respectively, with uncertainty u. y:x The basic rate a represents the confidence level of node y in the knowledge about node x. y:x This determines the extent to which uncertainty affects reputation;

[0031] In traditional subjective logic (TSL), local opinions are typically updated in the following way.

[0032] Where α represents the number of historical positive behaviors of node x, and β represents the number of historical negative behaviors of node x. Nodes select the best leader node and shard candidate group based on their reputation scores;

[0033] The result of the trustworthiness calculation of node y to node x

[0034] e y:x =b y:x +u y:x ×a y:x

[0035] If there exists a node y with respect to node x whose uncertainty is 1, i.e., ω y:x ={0,0,1,a y:x}, then the reliability calculation result of node y with respect to node x is e y:x =a y:x In other words, a node's default opinion on completely unknown nodes is the basic rate a. y:x ;

[0036] Considering the timeliness of events, the scale for measuring new and old events is t. recent If a node engages in malicious behavior, its credibility is weakened; this is combined with the weights of benign / malicious behavior.

[0037]

[0038] Where, α y:x and β y:x α' represents the weighted benign and malignant behaviors, respectively. y:x and β' y:x They represent t respectively <t recent Newer benign and disgusting behaviors, α″ y:x and β″ y:x They respectively represent t>t recent The evaluation criteria are: old benign behaviors and bad behaviors; newer events have a greater impact on reputation value, so we have ζ + σ = 1, ζ > σ, where ζ is the weight of newer events and σ is the weight of old events; in addition, bad behaviors should have a greater impact on local opinion than benign behaviors, so we have τ + θ = 1, τ > θ, where τ is the weight of bad behaviors and θ is the weight of benign behaviors.

[0039] Update local opinions using the following method.

[0040]

[0041] Step S2.5: The method of evaluating an unfamiliar node by sharing opinions is to set a constant β∈[0.0,1.0] as the threshold for unfamiliar nodes. If the uncertainty exceeds the threshold, a "recommendation" request is broadcast.

[0042] Furthermore, in step S2.2, the degree of the node is:

[0043]

[0044] Where, if there is a direct edge connecting node i and node j in the network, then δ ij If it is 1, otherwise δ ij =0;

[0045] The betweenness of a node is:

[0046]

[0047] Among them, l jk (i) represents the number of nodes i that are the shortest path between node j and node k. jk This represents the total number of shortest paths between node j and node k;

[0048] The clustering coefficient of a node is expressed as:

[0049]

[0050] Among them, e i Let d be the number of edges actually connecting the neighboring nodes of node i. i (d i -1) / 2 is the maximum number of possible edges between these neighboring nodes;

[0051] Normalize the node degree and betweenness respectively.

[0052]

[0053]

[0054] Where (N-1) is the maximum value of the node degree. It is the maximum betweenness centrality of a node; therefore, the parameter for measuring the importance of a node in terms of network topology is:

[0055]

[0056] By combining the node's local capabilities and importance parameters, a scalability resource score is obtained for the node as a sharding leader node.

[0057] V L (i) = A(i) × S(i).

[0058] Furthermore, in step S2.5, v i This indicates the degree of influence of node i's opinion on the recommendation; v i Based on the node's familiarity with the recommended node, c i:x =b i:x +d i:x The decision is as follows:

[0059]

[0060] The recommendations were integrated by considering the credibility of the recommenders and their familiarity with node x.

[0061]

[0062] in This represents the level of trust that node y gains about node x from other recommending nodes. This represents the level of trust that node i has in node x. The final opinion is obtained by combining local opinions and recommendations.

[0063]

[0064] Basic rate a A:B This determines the degree to which uncertainty affects reputation and can be set as a default value. The final result of the credibility calculation for requesting node A towards target node B is:

[0065] Furthermore, the specific steps of step S3 are as follows:

[0066] Step S3.1: Nodes i∈N that intend to participate in the consensus all N all It is a set of all nodes in the system that sends requests to participate in the consensus node election.

[0067] Step S3.2: The neighbor node j∈N that received the request. all If j≠i, a judgment is made based on the credibility score based on familiarity. Then node j sends a confirmation message with a timestamp and digital signature to node i, indicating that node j agrees that node i should become a consensus node. Here, e0 is the trustworthiness score threshold. If it is greater than the threshold, it means that based on the historical behavior of node i and node j, node j considers node i to be a trustworthy node.

[0068] Step S3.3: Node i participating in the consensus node election collects confirmation information from neighboring nodes. If the number of confirmation messages... Then broadcast consensus node selection information SELECT c This information includes the number of confirmation messages (CON). iAll confirmation information is provided for verification. A consensus node group will be formed after the consensus node election concludes. |N c | represents the number of nodes in the consensus node group.

[0069] Furthermore, the specific steps of step S4 are as follows:

[0070] Step S4.1: Nodes that have achieved the required scalability score and intend to become leaders issue a declaration, referring to the nodes participating in the leadership election as leaders, and the nodes that choose to follow the leader and join its shard as followers. The master-slave game between the nodes participating in the leadership election and the nodes that choose to follow the leader and join its shard is represented as follows:

[0071]

[0072] It mainly includes the following basic elements:

[0073] C L ={l1,l2,…,l i ,…,l o |A(i)>σ1} represents the set of candidate leader nodes for the sharding. Based on the leader node scalability score V L (i) Nodes that meet the conditions and issue a declaration to participate in the leader election are ranked according to V. L Divide into C from high to low L ; N represents the set of followers; c The remaining consensus nodes, excluding the candidate shard leader nodes, form the follower set;

[0074] X j For set C F middle followers f j The set of leadership strategies to choose from;

[0075] Represents |N c The utility function for |-o followers;

[0076] P i Represents set C L Candidate Sharding Leader Node i The sharding strategy, i.e., the candidate set of sharded replica nodes;

[0077] {I1,…,I o Let} be the utility function for the o candidate shard leader nodes;

[0078] Step S4.2, follower group node f j Based on the candidate group C of the segmented leader node LThe scalability score of the replica nodes is calculated sequentially, and this score represents the number of followers f. j Join candidate shard leader node l i The scalability of the managed fragments is expressed as a utility function of the followers, as follows:

[0079] U j (l i ) = V R (f j :l i )

[0080] Based on the principle of maximizing returns, the strategy for selecting follower nodes is to recommend the node l that meets the credibility score requirement and has the highest utility function. i As a candidate for leadership, that is

[0081] max V R (f j :l i )

[0082]

[0083] Where e0 is the credibility score threshold;

[0084] Step S4.3, Leader Node l i Then, from among its follower nodes, it selects r-1 replica nodes to form a shard, i.e.

[0085] max V R (f j :l i )

[0086] str i <r-1

[0087] Where r i For the leader node l i The number of nodes in the shard it leads. Therefore, the candidate shard leader node l i The utility function is

[0088]

[0089] Step S4.4: Design a leader exit mechanism. After the followers have selected a leader node, there will be leader candidate nodes. i Number of followers r i If the number of sharding nodes is less than the required number r-1, then for leader candidate nodes whose follower count meets the requirement, the selection of sharding nodes proceeds normally; however, for nodes with too few followers, fewer followers mean that node l iThe less suitable a node is to become a shard leader, the better; sort the nodes by the number of followers from highest to lowest, and select those that exceed half of the optimal number of shards S. Leader candidate nodes with fewer followers are selected from shard leader candidate group C. L Move to follower group C F In the process, the leader and its original followers will re-select a leader node.

[0090] The beneficial effects achieved by this invention are as follows:

[0091] (1) First, based on the throughput, latency, and security requirements of the actual application, the three dilemmas of blockchain will be coordinated according to the specific circumstances, and a sharded blockchain network that meets the application requirements will be designed.

[0092] (2) Provide dynamic scoring for the massive physical devices provided by different organizations through reputation systems and node evaluation methods.

[0093] (3) Through master-slave game between sharded leader nodes and replica nodes, nodes participating in consensus in the blockchain network will be mapped to the best physical devices based on the physical device scores. Attached Figure Description

[0094] Figure 1 This is the EvoChain self-evolved consortium blockchain system model in this embodiment of the invention.

[0095] Figure 2 This is a schematic diagram of the opinion sharing and combination process in an embodiment of the present invention. Detailed Implementation

[0096] The technical solution of the present invention will be further described in detail below with reference to the accompanying drawings.

[0097] An autonomous evolution design method for high-performance sharded consortium blockchains is proposed in this invention. The EvoChain high-performance sharded consortium blockchain is designed to coordinate and manage heterogeneous network infrastructure resources with varying performance and security from different participants in a 6G network. It is an autonomous evolution high-performance sharded consortium blockchain that selects nodes from the 6G network. Its system model is as follows: Figure 1As shown, EvoChain's self-evolving consortium blockchain system model is mainly divided into four layers from bottom to top: the resource layer, the network layer, the blockchain layer, and the application layer. The resource layer consists of network infrastructures with varying capabilities and security levels provided by different network operators, enterprises, and even individuals. Various organizations participate in the co-construction and sharing of the resource layer as resource providers. The network layer consists of network functional nodes obtained after virtualizing the network infrastructure of the resource layer, indicating the network capabilities and connectivity of the nodes provided by each organization. The blockchain layer, based on the network layer, selects a certain number of nodes to participate in blockchain consensus according to the requirements of scalability, security, and decentralization of actual applications. Consensus nodes are then divided into leader nodes and replica nodes for sharding consensus based on factors such as capability, importance, and trustworthiness. The application layer encapsulates various application scenarios and cases in 6G mobile networks, responsible for tightly integrating the blockchain platform with actual 6G network applications. It can be implemented through various scripts, algorithms, and smart contract programming to complete the customized transformation of the blockchain network.

[0098] The high-performance sharded consortium blockchain autonomous evolution design method is characterized by selecting secure, reliable, and high-performance nodes as leader nodes for the sharded blockchain from nodes provided by different participants in the 6G network. This selection is based on differences in computing power, communication bandwidth, and historical behavior among the nodes, the characteristics of the network topology, and the actual performance requirements of the blockchain in coordinating and managing resources in the 6G network. Each leader node then selects nodes to build its own sharded blockchain. Finally, the leader nodes aggregate transaction information from the sharded blockchains and achieve consensus on the main chain, flexibly enabling various applications. The method mainly includes the following steps:

[0099] Step S1: Based on the application's security and scalability requirements, construct a blockchain to meet these two needs, and calculate the total number of blockchain nodes, the number of shards, and the number of shard nodes to meet the actual application requirements.

[0100] Step S2: To map virtual elements in the blockchain virtual network to appropriate physical elements, the Resource Valuation Method (RVM) is used to score node resources. The resource value evaluation consists of two parts: scalability score and trustworthiness score.

[0101] Step S3: Conduct consensus node election, select trustworthy nodes from all nodes based on their trustworthiness scores to participate in the subsequent election game and consensus work;

[0102] Step S4: Leader Selection and Shard Division. Nodes that meet the scalability score and intend to become leaders issue a declaration, and the remaining nodes will select suitable leaders to join their shards. The leader node comprehensively considers the scalability score and the credibility score to select the best replica node to form a shard.

[0103] First, in step S1, a blockchain is constructed based on the application's security and scalability requirements. The total number of nodes, the number of shards, and the number of nodes in each shard are calculated to meet the actual application's security requirements. The specific steps are as follows:

[0104] Step S1.1: To meet the security requirements of the blockchain, it is necessary to calculate the probability of an attacker successfully launching an attack. Let n be the total number of nodes in the system. If an attacker attacks r nodes simultaneously, and the attack speed is t... a The time interval between initiating a transaction request and the time interval between initiating verification are both t. d If it follows an exponential distribution with parameter λ, then during attack period i, we have: A number of nodes are compromised, where Y is the number of nodes recovered in each attack cycle, representing the probability of a successful attack.

[0105] Step S1.2, if required and At that time, the constraints on the number of nodes n in each shard and the total number of nodes N in the blockchain network can be obtained as follows: For any attack period M, the following holds true:

[0106]

[0107]

[0108] Step S1.3: Calculate the total number of blockchain nodes, the number of shards, and the number of shard nodes to meet the scalability requirements of the actual application. If the specific application specifies scalability requirements, such as transaction throughput ≥ TPS0 and transaction latency ≤ T0, then based on the transaction throughput requirements, the corresponding total number of blockchain nodes N≥N0, the corresponding optimal number of shards S, and the number of nodes n in each shard can be calculated. This is the set of arrays containing the total number of blockchain nodes, the number of shards, and the number of shard nodes that meet the conditions.

[0109]

[0110] Where j is the number of sets that satisfy the condition.

[0111] Each array element in set S corresponds to an average transaction latency T. Elements with T ≤ T0 are selected to form a new array, and the set of the new array is...

[0112] S k ={[N1,S1,n1,T1],[N2,S2,n2,T2],……,[N k ,S k ,n k ,T k ]|Tk ≤T0}

[0113] Where k is the number of sets that satisfy the time delay condition.

[0114] Next, in step S2, the Resource Valuation Method (RVM) is used to score the node resources. The specific steps are as follows:

[0115] Step S2.1: For the leader node, the scalability score includes the node's local capabilities and node importance; for replica nodes, the scalability score includes the node's local capabilities and distance from the leader node. The node's local capabilities are as follows:

[0116]

[0117] Where CPU(i) and Stor(i) represent the computing power and storage power of node i, respectively, ij is the link directly connected to node i, and Band(ij) represents the bandwidth of the link. max Stor max and Band max These represent the maximum values ​​for computation, storage, and bandwidth, respectively. If A(i) > σ1, then node i can serve as the shard leader node.

[0118] Step S2.2: For nodes that meet the requirements to serve as shard leader nodes, their importance is measured using complex network structure characteristic metrics, namely, degree, betweenness, and clustering coefficient. The degree of a node refers to the number of nodes directly connected to it in the network, reflecting the magnitude of a node's direct influence within the network. The degree of a node is:

[0119]

[0120] Where, if there is a direct edge connecting node i and node j in the network, then δ ij If it is 1, otherwise δ ij It is 0.

[0121] The betweenness of a node reflects its control over information within the entire network; it is a global characteristic. The betweenness of a node is:

[0122]

[0123] Among them, l jk (i) represents the number of nodes i that are the shortest path between node j and node k. jk This represents the total number of shortest paths between node j and node k.

[0124] The clustering coefficient of a node characterizes the probability that any two of a node's neighbors are also neighbors. It reflects the strength of a node's connections with its surrounding neighbors and is a local feature. The larger the clustering coefficient, the stronger the connections between the node's neighbors. The clustering coefficient of a node can be expressed as:

[0125]

[0126] Among them, e i Let d be the number of edges actually connecting the neighboring nodes of node i. i (d i -1) / 2 represents the maximum number of edges that these neighboring nodes may have.

[0127] Normalize the node degree and betweenness respectively.

[0128]

[0129]

[0130] Where (N-1) is the maximum value of the node degree. It is the maximum betweenness centrality of a node. Therefore, the parameter for measuring the importance of a node in terms of network topology is:

[0131]

[0132] By combining the node's local capabilities and importance parameters, we can obtain a scalability resource score for the node as a sharding leader node.

[0133] V L (i) = A(i) × S(i)

[0134] Step S2.3: Since sharded replica nodes have lower local capabilities, nodes that were not selected as leader nodes can be chosen. Furthermore, compared to the importance of a node in the network topology, the selection of sharded replica nodes focuses more on their distance from the leader node, i.e., hop count. A lower average hop count within a shard can reduce message passing latency within the shard, effectively improving consensus speed and security.

[0135] Record S shard For the set of all partitions, a partition s j ∈S shard During the shard leader node's selection of shard replica nodes, it searches for nodes with a smaller hop distance h from itself and calculates their scores accordingly.

[0136]

[0137] Step S2.4: Each node scores the credibility of the other nodes, using subjective logic to represent the quadruple of opinion: ω y:x ={b y:x ,d y:x ,u y:x ,a y:x Trust level b y:x distrust level d y:x These represent the proportions of historical benign and undesirable behaviors among all behaviors, respectively, with uncertainty u. y:x The basic rate a represents the confidence level of node y in the knowledge about node x. y:x This determines the extent to which uncertainty affects reputation;

[0138] In traditional subjective logic (TSL), local opinions are typically updated in the following way.

[0139] Where α represents the number of historical positive behaviors of node x, and β represents the number of historical negative behaviors of node x. Nodes select the best leader node and shard candidate group based on their reputation scores;

[0140] The result of the trustworthiness calculation of node y to node x

[0141] e y:x =b y:x +u y:x ×a y:x

[0142] If there exists a node y with respect to node x whose uncertainty is 1, i.e., ω y:x ={0,0,1,a y:x}, then the reliability calculation result of node y with respect to node x is e y:x =a y:x In other words, a node's default opinion on completely unknown nodes is the basic rate a. y:x ;

[0143] Considering the timeliness of events, the scale for measuring new and old events is t. recent If a node engages in malicious behavior, its credibility is weakened; this is combined with the weights of benign / malicious behavior.

[0144]

[0145] Where, α y:x and β y:x α' represents the weighted benign and malignant behaviors, respectively. y:x and β' y:x They represent t respectively <t recent Newer benign and disgusting behaviors, α″ y:x and β″y:x They respectively represent t>t recent The evaluation criteria are: old benign behaviors and bad behaviors; newer events have a greater impact on reputation value, so we have ζ + σ = 1, ζ > σ, where ζ is the weight of newer events and σ is the weight of old events; in addition, bad behaviors should have a greater impact on local opinion than benign behaviors, so we have τ + θ = 1, τ > θ, where τ is the weight of bad behaviors and θ is the weight of benign behaviors.

[0146] Update local opinions using the following method.

[0147]

[0148] Step S2.5: The method for evaluating an unfamiliar node is to share opinions. A constant β∈[0.0,1.0] is set as the threshold for unfamiliar nodes. If the uncertainty exceeds the threshold, a "recommendation" request is broadcast. Figure 2 This illustrates the process of trustworthiness combination based on a familiarity-based model employed in this invention. Requesting node A attempts to interact with target node B, but its familiarity level is below a certain threshold, making it impossible to determine the trustworthiness of target node B based on local information. Therefore, in the subjective logic model, requesting node A requests its neighbors C, D, and E for their subjective logic regarding target node B. Node E lacks knowledge of B and therefore does not respond to the request. After issuing a recommendation request for target node B, requesting node A receives the subjective logic, i.e., recommendations, from recommending nodes C and D. The more interactions between nodes, the higher the familiarity level, reducing the randomness in the recommendations provided by the recommending nodes. Requesting nodes are also more inclined to use more definitive recommendations. Therefore, the opinions of the nodes issuing recommendations need to be weighted according to their familiarity with the target node. (v) i This indicates the degree of influence of node i's opinion on the recommendation. i Based on the node's familiarity with the recommended node, c i:x =b i:x +d i:x The decision is as follows:

[0149]

[0150] Taking into account the recommenders' credibility and their familiarity with x helps to better utilize existing information and reduce the uncertainty of the recommendation information. This facilitates the integration of recommendations.

[0151]

[0152] in This represents the level of trust that node y gains about node x from other recommending nodes. This represents the level of trust that node i has in node x. The final opinion is obtained by combining local opinions and recommendations.

[0153]

[0154] Basic rate a A:B This determines the degree to which uncertainty affects reputation and can be set as a default value. The final result of the credibility calculation for requesting node A towards target node B is:

[0155] Then, in step S3, trusted nodes are selected from all nodes to participate in the subsequent election game and consensus process. The specific method is as follows:

[0156] Step S3.1: Nodes i∈N that intend to participate in the consensus all N all It is a set of all nodes in the system that sends requests to participate in the consensus node election.

[0157] Step S3.2: The neighbor node j∈N that received the request. all (j≠i) Make a judgment based on the credibility score based on familiarity. If the credibility score is... Then node j sends a confirmation message with a timestamp and digital signature to node i, indicating that node j agrees that node i should become a consensus node. Here, e0 is the trustworthiness score threshold. If it is greater than this threshold, it means that node j considers node i to be a trustworthy node based on the historical behavior of node i and node j. This value is set to 0.5 by default and can be adjusted up or down according to the system's scalability and security requirements to adjust the scale of consensus nodes.

[0158] Step S3.3: Node i participating in the consensus node election collects confirmation information from neighboring nodes. If the number of confirmation messages... Then broadcast consensus node selection information SELECT c This information includes the number of confirmation messages (CON). i All confirmation information is provided for verification. The pseudocode for the consensus node election process is as follows.

[0159]

[0160]

[0161] After the consensus node election is completed, a consensus node group will be formed. Based on the conclusions above, we can rely on N c Number of nodes | N c Calculate the optimal number of partitions S and the number of nodes r that should exist in each partition, where r × s = n c Next, we need to go to N. cNodes are selected to serve as leader nodes, and the number of leader nodes equals the number of shards. Each leader node then selects replica nodes to form a shard. In this invention, since the participating nodes are all rational entities pursuing the maximization of distributed personal utility, game theory is the most suitable tool for analyzing the leader selection and sharding problem.

[0162] Finally, in step S4, leadership selection and segmentation are carried out, and the specific methods are as follows:

[0163] Step S4.1: Nodes that have achieved the required scalability score and intend to become leaders issue a declaration, referring to the nodes participating in the leadership election as leaders, and the nodes that choose to follow the leader and join its shard as followers. The master-slave game between the nodes participating in the leadership election and the nodes that choose to follow the leader and join its shard is represented as follows:

[0164]

[0165] It mainly includes the following basic elements:

[0166] (1)C L ={l1,l2,…,l i ,…,l o |A(i)>σ1} represents the set of candidate leader nodes for the sharding. Based on the leader node scalability score V L (i) Nodes that meet the conditions and issue a declaration to participate in the leader election can be ranked from V. L Divide into C from high to low L . N represents the set of followers; c The remaining consensus nodes, excluding the candidate shard leader nodes, form the follower set.

[0167] (2)X j For set C F middle followers f j The set of leadership strategies to choose from.

[0168] (3) Represents n c The utility function for -o followers.

[0169] (4)P i Represents set C L Candidate Sharding Leader Node i The sharding strategy is the candidate set of sharded replica nodes.

[0170] (5){I1,…,I o Let} be the utility function for the o candidate shard leader nodes.

[0171] Step S4.2, follower group node f j Based on the candidate group C of the segmented leader node L The scalability score of the replica nodes is calculated sequentially, and this score represents the number of followers f. j Join candidate shard leader node l i The scalability performance of the managed shards is as follows: the higher the value, the better the overall system performance, and the more likely the shard leader node is to be selected into the shard. In a master-slave game, this means that the followers f j Select l i The strategy of being a candidate shard leader node yields higher returns; therefore, this invention uses it as the utility function of followers, as follows:

[0172] U j (l i ) = V R (f j :l i )

[0173] Follower nodes selecting shards with better scalability means they can process more transactions per unit of time, which helps with information acquisition and credibility scoring, increasing their chances of being selected in subsequent shards and creating a virtuous cycle. Therefore, based on the principle of maximizing returns, the follower node selection strategy is to recommend the node l that meets the credibility scoring requirements and has the highest utility function. i As a candidate for leadership, that is

[0174] max V R (f j :l i )

[0175]

[0176] Where e0 is the credibility score threshold.

[0177] Step S4.3, Leader Node l i Then, from among its follower nodes, it selects r-1 replica nodes to form a shard, i.e.

[0178] max V R (f j :l i )

[0179] str i <r-1

[0180] Where r i For the leader node l i The number of nodes in the shard it leads. Therefore, the candidate shard leader node l i The utility function is

[0181]

[0182] Step S4.4: Because shard leaders have higher decision-making power and access permissions in the system, they have a greater advantage in the competition. Therefore, nodes that meet the requirements will actively participate in the election of shard leader nodes. This will generally result in the number of shard leader node candidates, o, being greater than the optimal number of shards. However, an excessively large number of shards not only fails to improve the throughput performance of the blockchain system, but also leads to a decrease in shard security due to an insufficient number of nodes within a shard, which is undesirable. Therefore, to ensure a suitable number of shards, this invention also designs a leader exit mechanism. After the followers have selected a leader node, there will be leader candidate nodes l. i Number of followers r i The number of sharding nodes is less than the required number r-1. For leader candidate nodes with sufficient follower counts, the selection of sharding nodes proceeds normally; however, for nodes with too few followers, fewer followers mean that node l i The less suitable a node is to become a shard leader, the better. Therefore, the nodes are sorted by the number of followers from highest to lowest, and those exceeding half the optimal number of shards S are selected. Leader candidate nodes with fewer followers are selected from shard leader candidate group C. L Move to follower group C F In this process, the leader itself and its existing followers undergo a new round of leader selection. In summary, the pseudocode for leader selection and sharding is as follows.

[0183]

[0184]

[0185] In Algorithm 2, step 1 requires looping through log2(n) in the worst case. c -s) times, while each iteration from step 2 to step 15 requires traversing all participating nodes, therefore requiring a maximum of n iterations. c Therefore, the time complexity of Algorithm 2 is O(nlogn).

[0186] The above description is only a preferred embodiment of the present invention. The scope of protection of the present invention is not limited to the above embodiments. Any equivalent modifications or changes made by those skilled in the art based on the content disclosed in the present invention should be included within the scope of protection set forth in the claims.

Claims

1. A high-performance sharded consortium blockchain autonomous evolution design method, characterized in that: Includes the following steps: Step S1: Based on the application's security and scalability, construct a blockchain to meet these two requirements. Calculate the probability of an attacker's successful attack and obtain constraints on the number of nodes in each shard based on the attack cycle and the total number of nodes in the blockchain network. Calculate the total number of blockchain nodes, shards, and shard nodes that meet the actual application requirements. The specific steps of step S1 are as follows: Step S1.1: Calculate the probability of a successful attack by the attacker; N is the total number of nodes in the system. If the attacker attacks r nodes simultaneously, and the attack speed is t... a The time interval between initiating a transaction request and the time interval between initiating verification are both t. d If it follows an exponential distribution with parameter λ, then during attack period i, we have: A number of nodes are compromised, where Y is the number of nodes recovered in each attack cycle, and the probability of a successful attack by the attacker is [missing information]. Step S1.2: If the probability of the entire blockchain network being compromised is required to be ≤ The probability of each fragment being compromised is ≤ At that time, the constraints for obtaining the number of nodes n in each shard and the total number of nodes N in the blockchain network are as follows: For any attack period M, the following holds true: Step S2: In order to map virtual elements in the blockchain virtual network to appropriate physical elements, resource value is used to score node resources; the evaluation of resource value is divided into two parts: scalability score and trustworthiness score. Step S3: Conduct consensus node election, select trustworthy nodes from all nodes based on their trustworthiness scores to participate in the subsequent election game and consensus work; Step S4, Leader Selection and Sharding: Nodes that meet the scalability score and intend to become leaders issue a declaration, and the remaining nodes will select suitable leaders to join their shards. The leader node comprehensively considers the scalability score and the credibility score to select the best replica node to form a shard.

2. The high-performance sharded consortium blockchain autonomous evolution design method according to claim 1, characterized in that: Step S1 also includes the following steps: Step S1.3: Calculate the total number of blockchain nodes, the number of shards, and the number of shard nodes to meet the scalability requirements of the actual application. The application provides scalability requirements: transaction throughput ≥ TPS0, transaction latency ≤ T0. Based on the transaction throughput requirements, calculate the corresponding total number of blockchain nodes N≥N0, the optimal number of shards S, and the number of nodes n in each shard. This is the set of arrays containing the total number of blockchain nodes, the number of shards, and the number of shard nodes that meet the conditions. Where j is the number of sets that satisfy the condition; Each array element in set S corresponds to an average transaction latency T. Elements with T ≤ T0 are selected to form a new array, and the set of the new array is: S k ={[N1,S1,n1,T1],[N2,S2,n2,T2],……,[N k ,S k ,n k ,T k ]|T k ≤T0} Where k is the number of sets that satisfy the time delay condition.

3. The high-performance sharded consortium blockchain autonomous evolution design method according to claim 1, characterized in that: The specific steps of step S2 are as follows: Step S2.1: For the leader node, the scalability score includes the node's local capabilities and node importance; for replica nodes, the scalability score includes the node's local capabilities and distance from the leader node; the node's local capabilities are as follows: Where CPU(i) and Stor(i) represent the computing power and storage power of node i, respectively, ij is the link directly connected to node i, and Band(ij) represents the bandwidth of the link. max Stor max and Band max These are the maximum values ​​for computing power, storage capacity, and bandwidth, respectively; if A(i)>σ1, it means that node i is the shard leader node, and σ1 is the capability threshold, which is set according to the capability requirements of the node; Step S2.2: For nodes that meet the requirements to serve as the leader node of a shard, their importance is measured using complex network structure characteristic metrics, namely the degree, betweenness, and clustering coefficient of the node. Step S2.3: Since the local capabilities required for sharded replica nodes are relatively low, nodes that were not selected as leader nodes are chosen. In addition, the selection of sharded replica nodes focuses on the distance from the leader node, i.e., the number of hops; Record S shard For the set of all partitions, a partition s j ∈S shard During the shard leader node's selection of shard replica nodes, it searches for nodes with a smaller hop distance h from itself. i:j Let i be the shortest number of hops required to get from node i to node j, and calculate its score accordingly. Step S2.4: Each node scores the credibility of the other nodes, using subjective logic to represent the quadruple of opinion: ω y:x ={b y:x ,d y:x ,u y:x ,a y:x Trust level b y:x distrust level d y:x These represent the proportions of historical benign and undesirable behaviors among all behaviors, respectively, with uncertainty u. y:x The basic rate a represents the confidence level of node y in the knowledge about node x. y:x This determines the extent to which uncertainty affects reputation; In traditional subjective logic TSL, local opinions are updated in the following way: Where α represents the number of historical positive behaviors of node x, and β represents the number of historical negative behaviors of node x; nodes select the best leader node and shard candidate group based on their reputation value; The result of the credibility calculation of node y with respect to node x: e y:x =b y:x +u y:x ×a y:x If there exists a node y with respect to node x whose uncertainty is 1, i.e., ω y:x ={0,0,1,a y:x }, then the reliability calculation result of node y with respect to node x is e y:x =a y:x In other words, a node's default opinion on completely unknown nodes is the basic rate a. y:x ; Considering the timeliness of events, the scale for measuring new and old events is t. recent If a node engages in malicious behavior, its credibility is weakened; this is combined with the weights of benign / malicious behavior. Where, α y:x and β y:x α' represents the weighted benign and malignant behaviors, respectively. y:x and β' y:x They represent t respectively <t recent Newer benign and disgusting behaviors, α” y:x and β” y:x They respectively represent t>t recent The evaluation criteria are: old benign behavior and bad behavior; newer events have a greater impact on reputation value, so we have ζ + σ = 1, ζ > σ, where ζ is the weight of newer events and σ is the weight of old events; in addition, bad behavior should have a greater impact on local opinion than benign behavior, so we have τ + θ = 1, τ > θ, where τ is the weight of bad behavior and θ is the weight of benign behavior. The following methods should be used to update local comments: Step S2.5: The method of evaluating an unfamiliar node by sharing opinions is to set a constant β∈[0.0,1.0] as the threshold for unfamiliar nodes. If the uncertainty exceeds the threshold, a "recommendation" request is broadcast.

4. The high-performance sharded consortium blockchain autonomous evolution design method according to claim 3, characterized in that: In step S2.2, the degree of the node is: Where, if there is a direct edge connecting node i and node j in the network, then δ ij If it is 1, otherwise δ ij =0; The betweenness of a node is: Among them, l jk (i) represents the number of nodes i that are the shortest path between node j and node k. jk This represents the total number of shortest paths between node j and node k; The clustering coefficient of a node is expressed as: Among them, e i Let d be the number of edges actually connecting the neighboring nodes of node i. i (d i -1) / 2 is the maximum number of possible edges between these neighboring nodes; Normalize the node degree and betweenness respectively. Where (N-1) is the maximum value of the node degree. It is the maximum betweenness centrality of a node; therefore, the parameter for measuring the importance of a node in terms of network topology is: By combining the node's local capabilities and importance parameters, a scalability resource score is obtained for the node as a sharding leader node. V L (i)=A(i)×S(i)。 5. The high-performance sharded consortium blockchain autonomous evolution design method according to claim 3, characterized in that: In step S2.5, v i This indicates the degree of influence of node i's opinion on the recommendation; v i Based on the node's familiarity with the recommended node, c i:x =b i:x +d i:x The decision is as follows: The recommendations were integrated by considering the credibility of the recommenders and their familiarity with node x. in This represents the level of trust that node y gains about node x from other recommending nodes. This represents the level of trust that node i has in node x; the final opinion is obtained by combining local opinions and recommendations. Basic rate a A:B The degree to which uncertainty affects reputation is determined; the final result of the credibility calculation for requesting node A to target node B is:

6. The high-performance sharded consortium blockchain autonomous evolution design method according to claim 1, characterized in that: The specific steps of step S3 are as follows: Step S3.1: Nodes i∈N that intend to participate in the consensus all N all It is a set of all nodes in the system that sends requests to participate in the consensus node election; Step S3.2: The neighbor node j∈N that received the request. all If j≠i, a judgment is made based on the credibility score based on familiarity. Then node j sends a confirmation message with a timestamp and digital signature to node i, indicating that node j agrees that node i should become a consensus node. Here, e0 is the trustworthiness score threshold. If it is greater than the threshold, it means that based on the historical behavior of node i and node j, node j considers node i to be a trustworthy node. Step S3.3: Node i participating in the consensus node election collects confirmation information from neighboring nodes. If the number of confirmation messages... Then broadcast consensus node selection information SELECT c This information includes the number of confirmation messages (CON). i All confirmation information is available for verification; a consensus node group will be formed after the consensus node election concludes. |N c | represents the number of nodes in the consensus node group.

7. The high-performance sharded consortium blockchain autonomous evolution design method according to claim 1, characterized in that: The specific steps of step S4 are as follows: Step S4.1: Nodes that have achieved the required scalability score and intend to become leaders issue a declaration, referring to the nodes participating in the leadership election as leaders, and the nodes that choose to follow the leader and join its shard as followers. The master-slave game between the nodes participating in the leadership election and the nodes that choose to follow the leader and join its shard is represented as follows: It includes the following basic elements: C L ={l1,l2,…,l i ,…,l o |A(i)>σ1} represents the set of candidate leader nodes for the sharding. Based on the leader node scalability score V L (i) Nodes that meet the conditions and issue a declaration to participate in the leader election are ranked according to V. L Divide into C from high to low L ; N represents the set of followers; c The remaining consensus nodes, excluding the candidate shard leader nodes, form the follower set; X j For set C F middle followers f j The set of leadership strategies to choose from; Represents |N c The utility function for |-o followers; P i Represents set C L Candidate Sharding Leader Node i The sharding strategy, i.e., the candidate set of sharded replica nodes; {I1,…,I o Let} be the utility function for the o candidate shard leader nodes; Step S4.2, follower group node f j Based on the candidate group C of the segmented leader node L The scalability score of the replica nodes is calculated sequentially, and this score represents the number of followers f. j Join candidate shard leader node l i The scalability of the managed fragments is expressed as a utility function of the followers, as follows: U j (l i )=V R (f j :l i ) Based on the principle of maximizing returns, the strategy for selecting follower nodes is to recommend the node l that meets the credibility score requirement and has the highest utility function. i As a candidate for leadership, that is max V R (f j :l i ) Where e0 is the credibility score threshold; Step S4.3, Leader Node l i Then, from among its follower nodes, it selects r-1 replica nodes to form a shard, i.e. max V R (f j :l i ) s.t. r i <r-1 Where r i For the leader node l i The number of nodes in the shard it leads; therefore, the candidate shard leader node l i The utility function is: Step S4.4: Design a leader exit mechanism. After the followers have selected a leader node, there will be leader candidate nodes. i Number of followers r i If the number of sharding nodes is less than the required number r-1, then for leader candidate nodes with sufficient follower counts, the selection of sharding nodes proceeds normally; however, for nodes with too few followers, fewer followers mean that node l i The less suitable a node is to become a shard leader, the better; sort the nodes by the number of followers from highest to lowest, and select those that exceed half of the optimal number of shards S. Leader candidate nodes with fewer followers are selected from shard leader candidate group C. L Move to follower group C F In this process, the leader and its original followers will re-select a leader node.

Images