A dynamic delay evaluation method based on a permissioned blockchain database system

Abstract

The application relates to the technical field of blockchains, in particular to a dynamic delay evaluation method based on a permissioned blockchain database system, which comprises the following steps: S1: obtaining a node quantity and establishing a delay regression curve model according to the influence of consistency sigma, concurrency gamma and resource competitiveness beta on delay, and calculating a delay starting value under a single node and a delay growth rate under a current node quantity according to the node quantity and the delay regression curve model; S2: comparing the delay starting value and the delay growth rate of the delay regression curve model, if the results are the same, no processing is performed, if the results are different, a dynamic delay index QDL is calculated k , and the delay comparison result is determined according to the size of the dynamic delay index QDL k . The application proposes a universal delay law ULL regression curve model, superparameters are determined by using Newton interpolation and the least square method, and a quantitative dynamic delay QDL model is constructed, the model considers the size and growth rate of delay, and comprehensively evaluates the delay.

Description

Technical Field

[0001] This invention relates to the field of blockchain technology, and in particular to a dynamic delay evaluation method based on a permissioned blockchain database system. Background Technology

[0002] Blockchain technology, with its enhanced security and decentralized properties, has attracted the interest of researchers in the database field. However, these advantages are accompanied by increased latency. The scalability of a blockchain system depends on two properties: throughput and latency. However, most research focuses on improving system throughput, neglecting latency studies. While a few studies have investigated latency, they either focus on the latency of specific algorithms or only provide brief numerical simulations to predict latency. Our research aims to develop a more comprehensive model to qualitatively assess latency in blockchain systems.

[0003] Early database systems treated latency as part of availability, leading to the CAP theorem. This theorem states that in a partitioned distributed system, there is a trade-off between consistency and availability. Later, the CAP theorem was extended to the PACELC theorem: in the presence of partitions, the system faces a trade-off between consistency and availability; in the absence of partitions, a trade-off is needed between consistency and latency. In the context of blockchain systems, the DCS conjecture also emerged: it emphasizes the trade-off between decentralization, consistency, and scalability, where latency is considered part of scalability. Therefore, changes in decentralization and consistency within a blockchain system can affect latency. In practical applications, creating a block in a blockchain system takes approximately 10 minutes, which is clearly unacceptable to users. Research shows that users can only tolerate latency of less than 4 seconds. Thus, in addition to typical public blockchain systems, private blockchain systems have emerged. Private blockchains sacrifice some decentralization features to compensate for latency. For example, ResilientDB is a private blockchain system that reduces latency to 3.3 seconds. Therefore, current private blockchains generally meet user needs.

[0004] As mentioned earlier, scalability and consistency assessments are closely related to latency. However, current methods for assessing latency compare the average latency of different consensus protocols with the same number of nodes, which has limitations. It does not consider latency variations with different numbers of nodes and may fail to yield satisfactory results when the assessment contradicts the latency growth rate. This paper proposes a framework to address these issues and provide a comprehensive assessment of latency. Summary of the Invention

[0005] The purpose of this invention is to address the shortcomings of existing technologies by providing a dynamic latency evaluation method based on a permissioned blockchain database system, comprising the following steps:

[0006] S1: Obtain the number of nodes and establish a latency regression curve model based on the impact of consistency σ, concurrency γ, and resource competition β on latency. Calculate the initial latency value for a single node and the latency growth rate for the current number of nodes based on the number of nodes and the latency regression curve model.

[0007] S2: Compare the initial delay value and the delay growth rate of the delayed regression curve model. If the results are the same, no action is taken. If the results are different, calculate the dynamic delay index QDL. k Through the dynamic latency index QDL k The size of the delay determines the comparison result.

[0008] Further, in step S1, establishing the delayed regression curve model further includes:

[0009] An optimistic delay regression loss model is established when the resource competitiveness β does not affect the delay, and a pessimistic delay regression loss model is established when the resource competitiveness β begins to affect the delay.

[0010] Furthermore, the establishment of the optimistic delayed regression loss model further includes:

[0011] When the resource contention β does not affect latency, the consistency σ and concurrency γ affect latency. Based on the characteristics of permissioned blockchain, the optimistic latency regression loss model OLRL is established, and the calculation formula of the optimistic latency regression loss model OLRL is as follows:

[0012]

[0013] Where L(N) is the delay value with N nodes, and L(1) is the initial delay value.

[0014] Furthermore, the method of establishing the pessimistic delayed regression loss model further includes:

[0015] When the resource contention β begins to affect latency, the pessimistic latency regression loss model (PLRL) is established based on the characteristics of the permissioned blockchain. The calculation formula for the pessimistic latency regression loss model (PLRL) is as follows:

[0016]

[0017] Wherein, N0 represents the maximum capacity of the communication channel, and N0 is a fixed value related to the current network bandwidth and information load.

[0018] Further, in step S1, calculating the initial delay value for a single node further includes:

[0019] Each node is obtained based on the delayed regression curve, and the initial delay value L(1) is calculated using Newton's interpolation method:

[0020] L(1)=a0+a1(1-x0)+a2(1-x0)(1-x1)+…+a m (1-x0)(1-x1)…(1-x m-1 ),

[0021] Among them, a i Let x represent the i-th order difference quotient, and let x represent the number of nodes.

[0022] Further, in step S1, calculating the latency growth rate under the current number of nodes further includes:

[0023] The delay growth rate is calculated using the delay regression curve and the target parameter calculated by the least squares method. Specifically, when the resource competitiveness β does not affect the delay, the delay growth rate is obtained by applying the optimistic delay regression loss model OLRL and the first target parameter calculated by the least squares method to the delay growth rate model. When the resource competitiveness β affects the delay, the delay growth rate is obtained by applying the pessimistic delay regression loss model PLRL and the second target parameter calculated by the least squares method to the delay growth rate model.

[0024] More preferably, the theoretical value of the impact of resource contention β on latency can be illustrated by the PBFT consensus algorithm. When the number of nodes n = B / (pm+3sm), resource contention β begins to affect latency, where B represents bandwidth, pm represents the scale of load information, and sm represents the scale of state information.

[0025] Furthermore, the first target parameter further includes:

[0026] The first target parameter is calculated using a first objective function, the formula for which is as follows:

[0027]

[0028] Wherein, parameter α0 = 1 / γ and parameter α1 = σ / γ, Q represents the error between parameter α0 and parameter α1, and L i N represents the current node. i Let represent the dataset of the i-th single node. The first hyperparameter of the optimistic delayed regression loss model OLRL, applied to the calculation formula of the least squares method, is as follows:

[0029]

[0030] Furthermore, the second target parameter further includes:

[0031] The second objective parameter is calculated using a second objective function, the formula for which is as follows:

[0032]

[0033] Wherein, the parameters α0 = (1-βx0), α1 = (β-σ) / γ, and α2 = σ / γ are obtained by applying the second hyperparameter calculated by the pessimistic delayed regression loss model PLRL to the least squares method. and

[0034] Furthermore, the delay growth rate model further includes:

[0035] The formula for calculating the delay growth rate (LGR) model is as follows:

[0036]

[0037] Where, k sn Let α be the slope and α be the objective parameter. Depending on the chosen consensus, specifically the optimistic delayed regression loss model OLRL and the pessimistic delayed regression loss model PLRL, the slope k... sn There are two calculation formulas:

[0038]

[0039]

[0040] Further, in step S2, the dynamic delay index QDL is calculated. k Further including:

[0041] The relative latency value RVAP is calculated using either the first hyperparameter or the second hyperparameter, given the current number of nodes. k The calculation formula is as follows:

[0042]

[0043] Among them, let L i (N) represents the delay value of the i-th data in ascending order with N nodes, k represents the k-th minimum value of the actual points of the current node in ascending order, and s represents the total number of actual points;

[0044] The relative delay value RVAP k The delay growth rate is input into the dynamic delay evaluation QDL model to obtain the dynamic delay index QDL. k The dynamic latency index QDL kThe formula is shown below:

[0045]

[0046] According to the dynamic latency index QDL k The comparison results determine that the dynamic delay index QDL k A smaller result indicates less latency, which means better performance, and vice versa, a larger result indicates greater latency, which means worse performance.

[0047] Compared with the prior art, the beneficial effects of the present invention are:

[0048] (1) This invention proposes a Universal Delay Law (ULL) regression curve model. By using Newton interpolation and least squares method to determine hyperparameters, a quantitative dynamic delay (QDL) model is constructed. This model considers the magnitude and growth rate of delay to comprehensively evaluate the delay. The QDL model can evaluate the delay of different consensus protocols under different numbers of nodes.

[0049] (2) The present invention evaluated the relative error of the ULL model. The average relative error was 6.60% after multiple experiments. The present invention studies the influence of hyperparameters on the model through numerical simulation. Attached Figure Description

[0050] Various other advantages and benefits will become apparent to those skilled in the art upon reading the following detailed description of preferred embodiments. The accompanying drawings are for illustrative purposes only and are not intended to limit the invention.

[0051] Figure 1 This is a flowchart of a dynamic delay evaluation method based on a permissioned blockchain database system according to the present invention;

[0052] Figure 2 This is a schematic diagram of the overall framework of a dynamic delay evaluation method based on a permissioned blockchain database system according to the present invention.

[0053] Figure 3 This diagram illustrates an application scenario of the dynamic delay evaluation method based on a permissioned blockchain database system according to the present invention.

[0054] Figure 4 This is a comparison of simulated regression curves for a dynamic delay evaluation method based on a permissioned blockchain database system according to the present invention. Detailed Implementation

[0055] To make the objectives, technical solutions, and advantages of this invention clearer, the invention will be further described in detail below with reference to the accompanying drawings and embodiments. It should be understood that the specific embodiments described herein are merely illustrative of the invention and are not intended to limit the invention. Obviously, the described embodiments are only some, not all, of the embodiments described in this application. All other embodiments obtained by those skilled in the art based on the embodiments in this application without creative effort are within the scope of protection of this application.

[0056] Those skilled in the art will understand that, unless otherwise stated, the singular forms “a” and “an” used herein, and “the”, may also include the plural forms. It should be further understood that the term “comprising” as used in this specification means the presence of the stated features, integers, steps, operations, elements, and / or components, but does not exclude the presence or addition of one or more other features, integers, steps, operations, elements, components, and / or groups thereof.

[0057] Example 1

[0058] Please see Figure 1 and Figure 2 As shown in the figure, this embodiment provides a dynamic latency evaluation method based on a permissioned blockchain database system, which includes the following steps:

[0059] S1: Based on the impact of consistency σ, concurrency γ, and resource contention β on latency, obtain the number of nodes and establish a latency regression curve model. Calculate the initial latency value for a single node and the latency growth rate for the current number of nodes based on the number of nodes and the latency regression curve model. Specifically, in this embodiment, the present invention uses the QDL model to simulate the latency evaluation process, comparing the latency evaluation of four blockchain consensus protocols with different numbers of nodes. When evaluating the latency of four points—single node a1, single node a2, single node a3, and single node a4—concurrency γ represents the parallelism level of the system. In the field of blockchain technology, public chains can promote parallel operations through sidechains and the Lightning Network, while private chains can transform multiple nodes into clusters or shards to achieve parallelization.

[0060] Preferably, consistency in a blockchain system stems from data replication or verification, which is a result of decentralization. As proposed by the DCS conjecture, in public chains, the consensus process essentially solves the consistency problem of the system. This leads to the inevitable trade-off between the consistency of public chains and the loss of latency. In private chains, blocks can only be created after consensus is reached, which also involves a certain degree of latency loss.

[0061] While blockchain systems are decentralized, resource contention still exists. Public blockchains, using broadcasting and block building methods, can theoretically accommodate an unlimited number of nodes without contention issues. Conversely, in private blockchains, the increased number of nodes leads to increased complexity in full-to-full communication. Contention issues may arise when bandwidth is fully utilized above a certain threshold.

[0062] S2: Compare the initial delay value and the delay growth rate of the delayed regression curve models. If the results are the same, no action is taken. If the results are different, calculate the dynamic delay index QDL. k Through the dynamic delay metric QDL k The size of the delay determines the comparison result.

[0063] Preferably, this embodiment analyzes the main factors affecting the latency of a blockchain system and identifies concurrency and consistency as the primary determinants. However, as the number of nodes increases, latency is also affected by resource contention. Based on this, simulated regression curves of distributed databases are analyzed, deriving a regression model called the Universal Delay Law (ULL) model, i.e., a latency regression curve model. This model describes the relationship between latency and the number of nodes. Depending on whether the impact of competition on latency is considered, we further divide the ULL model into the Optimistic Delay Regression Loss Model (OLRL) and the Super Loss Model (PLRL).

[0064] Further, in step S1, establishing a delayed regression curve model further includes:

[0065] An optimistic delay regression loss model is established when resource competitiveness β does not affect delay, and a pessimistic delay regression loss model is established when resource competitiveness β begins to affect delay.

[0066] refer to Figure 3 As shown, an optimistic delayed regression loss model is established, which further includes:

[0067] When resource contention β has no effect on latency, consistency σ and concurrency γ affect latency. Based on the characteristics of permissioned blockchains, an optimistic latency regression loss model (OLRL) is established. The calculation formula for the optimistic latency regression loss model OLRL is as follows:

[0068]

[0069] Where L(N) is the delay value with N nodes, and L(1) is the initial delay value. Specifically, in this embodiment, most private blockchain systems use all-to-all communication and voting, requiring O(N) latency. 2To reduce communication complexity and achieve higher security, OLRL can obtain a loss function: f(N) = N(N-1). Specifically, in this embodiment, the optimistic delayed regression loss model OLRL, which uses a simulated regression curve model, is applied to reconstruct curve A.

[0070] Furthermore, the feature is that establishing a pessimistic delayed regression loss model further includes:

[0071] When resource competition β begins to affect latency, a pessimistic latency regression loss model (PLRL) is established based on the characteristics of permissioned blockchains. PLRL yields a loss function f(N) = N(N-1) + λ(N-N0), where λ is an independent parameter. The calculation formula for the pessimistic latency regression loss model PLRL is as follows:

[0072]

[0073] Wherein, N0 represents the maximum capacity of the communication channel. N0 is a fixed value that is related to the bandwidth and information load of the current network. Specifically, in this embodiment, the super loss model PLRL of the simulated regression curve model is used to restore curve B.

[0074] Preferably, the initial value of the delays of a1, a2, a3, and a4 is L. a1 (1) = L a3 (1)>L a2 (1) = L a4 (1).

[0075] Further, in step S1, calculating the initial delay value for a single node further includes:

[0076] Each node is obtained from the delayed regression curve, and the initial delay value L(1) is calculated using Newton's interpolation method:

[0077] L(1)=a0+a1(1-x0)+a2(1-x0)(1-x1)+…+a m (1-x0)(1-x1)…(1-x m—1 ),

[0078] Among them, a i Let represent the i-th order difference quotient, and x represent the number of nodes. Specifically, in this embodiment, the equation of the Newton interpolation polynomial N(x) is as follows:

[0079]

[0080] Among them, a i Let x represent the number of nodes, m represent the number of test datasets, i represent the i-th dataset, j correspond to the j-th polynomial, and n represent the number of nodes. i(x) can be calculated using the following formula:

[0081]

[0082] Preferably, we use Newton's interpolation method to address the latency of a system with only one node. It is worth noting that this single-node latency includes a consistency loss, which distinguishes it from traditional single-node systems because it is a theoretical value rather than an empirical measurement.

[0083] Further, in step S1, calculating the latency growth rate under the current number of nodes further includes:

[0084] The delay growth rate is calculated by using the delay regression curve and the target parameter calculated by the least squares method. Specifically, when resource competitiveness β does not affect the delay, the delay growth rate is obtained by applying the first target parameter calculated by the optimistic delay regression loss model OLRL and the least squares method to the delay growth rate model. When resource competitiveness β affects the delay, the delay growth rate is obtained by applying the second target parameter calculated by the pessimistic delay regression loss model PLRL and the least squares method to the delay growth rate model.

[0085] More preferably, the theoretical value of the impact of resource contention β on latency can be determined according to the PBFT consensus algorithm. When the number of nodes n = B / (pm+3sm), resource contention β begins to affect latency, where B represents bandwidth, pm represents the scale of load information, and sm represents the scale of state information.

[0086] Furthermore, the first target parameter further includes:

[0087] The first objective parameter is calculated using the first objective function, which is calculated as follows:

[0088]

[0089] Where, parameter α0 = 1 / γ and the number of parameters α1 = σ / γ, Q represents the error between the test data and the simulated data, and L i N represents the test latency data. i For the dataset representing the number of the i-th node, since Q is a squared value, it is always greater than or equal to zero. Therefore, the necessary condition for the existence of an extremum in the above equation is as follows:

[0090]

[0091] The first hyperparameter of the optimistic delayed regression loss model OLRL, when applied to the least squares calculation formula, is as follows:

[0092]

[0093] Furthermore, the second objective parameter further includes:

[0094] The second objective parameter is calculated using the second objective function, which is calculated as follows:

[0095]

[0096] Wherein, the parameters α0=(1-βx0), α1=(β-σ) / γ, and α2=σ / γ are obtained by applying the second hyperparameter calculated by the pessimistic delayed regression loss model PLRL to the least squares method. and The formula for applying the second hyperparameter to the least squares method is as follows:

[0097]

[0098] Using the normal equations mentioned above, we can solve for the relationship between the unknown variables as follows: σ=α2 / [α0+x0(α1+α2)], β=(α1+α2) / [α0+x0(α1+α2)] and γ=1 / [α0+x0(α1+α2)].

[0099] Furthermore, the delayed growth rate model further includes:

[0100] The formula for calculating the delay growth rate (LGR) model is shown below:

[0101]

[0102] Where, k sn Let α be the slope and α be the objective parameter. Depending on the chosen consensus, specifically in the cases of the optimistic delayed regression loss model (OLRL) and the pessimistic delayed regression loss model (PLRL), the slope k... sn There are two calculation formulas:

[0103]

[0104] Specifically, in this embodiment, α is a parameter for fine-tuning the slope value, while k sn This represents the current system slope at a given number of nodes. To determine the slope in the OLRL equation with the current number of nodes, we express it as the difference between N+1 nodes and N nodes. Considering that the number of nodes must be a positive integer, the calculation result is as follows:

[0105]

[0106] Similar to the process of calculating the slope under the optimistic delayed regression loss model OLRL, we obtain the following results for calculating the slope under the pessimistic delayed regression loss model PLRL:

[0107]

[0108] refer to Figure 3 As shown, in step S2, the dynamic delay index QDL is calculated. k Further including:

[0109] The relative delay value RVAP is calculated using either the first or second hyperparameter, given the current number of nodes. k The calculation formula is as follows:

[0110]

[0111] Among them, let L i (N) represents the delay value of the i-th data point arranged in ascending order with N nodes, k represents the k-th minimum value of the actual points of the current node in ascending order, and s represents the total number of actual points. Specifically, in this embodiment, the relative delay value RVAP k It is a relative measure of actual data points and may vary when compared between different objects. The relative delay value (RVAP) is... k It also falls within the range of [0,1]. The closer the value is to 1, the higher the latency, which is more detrimental to the assessment of latency.

[0112] The above formula calculates the relative delay value RVAP. k This refers to the relative evaluation index. It's important to note that the evaluation index for the same actual point may differ across different sets of actual points (s), considering the delay growth rate (LGR) and RVAP models. k The calculation formulas all provide evaluation indicators within the same value range, and the following QDLs can be used. k The tradeoff equation easily balances these two results, with the delay relative value RVAP. k The delay growth rate is input into the dynamic delay evaluation QDL model to obtain the dynamic delay index QDL. k Dynamic latency metric QDL k The formula is shown below:

[0113]

[0114] According to the dynamic latency index QDL k The comparison results confirm that the dynamic delay index QDL k A smaller result indicates lower latency, i.e., better performance, while a larger result indicates higher latency, i.e., worse performance. Specifically, in this embodiment, the assessment of latency in the latency-node graph largely depends on the slope under the current number of nodes. The larger the slope, the faster the latency changes as the number of nodes increases, which is detrimental to the user. The larger the slope, the worse the system performance.

[0115] More ideally, for the three points where the initial delay magnitude and delay growth rate are contradictory, it is necessary to calculate the relative actual point values ​​for the remaining three points a1, a2, and a3 separately. First, based on the known information, calculate the delay magnitude using the current number of nodes at the three points a1, a2, and a3 respectively, and then import the results into the relative delay value RVAP. k The formula was used to calculate the scores for the three points, normalized to 0 to 1. The results showed that the RVAP score was RVAP. a3 >RVAP a1 >RVAP a2 .

[0116] Then, according to the dynamic delay evaluation formula QDL k By combining the RVAP score and the LGR score, a three-point QDL is obtained. k The result is QDL a3 QDL a2 QDL a2 .

[0117] Even better, LGR a2 >LGR a3 >LGR a1 ≈LGR a4 Based on the initial latency value and the latency growth rate, single node a4 has the smallest L(1) and LGR values ​​in the comparison results. Therefore, it can be directly determined that single node a4 has the smallest latency and thus better performance. In summary, it can be concluded that single node a3 has the largest latency, indicating the worst performance; while single node a4 has the smallest latency, indicating the best performance. The overall dynamic latency ranking is as follows: QDL a3 QDL a2 QDL a1 QDL a4 .

[0118] The latency growth rate model (LGR) proposed in this invention aims to normalize the slope and calculate a dynamic latency index (QDL) in the range of 0 to 1. k Using the dynamic delay metric QDL k The magnitude of the slope determines the final delay comparison result, where a smaller slope value corresponds to a delay growth rate closer to 0, and a larger slope value corresponds to a delay growth rate closer to 1.

[0119] refer to Figure 4As shown, this embodiment compares the regression performance of the Delayed Regression (OLRL) model with other models. Based on the data processing described above, we can see that using the reciprocal of the traditional USL model to simulate the regression curve analysis is completely inconsistent with the sample data. To better distinguish between the two models, we calculate the relative error by comparing the sample data with the regression curves of these two models. The results show that the average relative error of the OLRL model is 6.68%, and the average relative error of the PLRL model is 6.33%. The OLRL equation has a smaller correlation error in the middle range of the sample data, indicating better regression performance in this range. On the other hand, the PLRL equation shows a smaller correlation error at the maximum and minimum values ​​of the sample data, indicating better regression performance in these regions. Therefore, it is recommended to apply the OLRL equation to the middle values ​​of the sample data and use the PLRL equation at the maximum and minimum values.

[0120] The above description is merely a preferred embodiment of the present invention. The scope of protection of the present invention is not limited to the above embodiments. All technical solutions falling within the scope of the present invention's concept are within the scope of protection of the present invention. It should be noted that for those skilled in the art, any improvements and modifications made without departing from the principles of the present invention should also be considered within the scope of protection of the present invention.

Claims

1. A dynamic latency evaluation method based on a permissioned blockchain database system, characterized in that, Includes the following steps: S1: Based on consistency Concurrency and resource competition To assess the impact on latency, the number of nodes is determined and a latency regression curve model is established. Based on the number of nodes and the latency regression curve model, the initial latency value for a single node and the latency growth rate for the current number of nodes are calculated. S2: Compare the initial delay value and the delay growth rate of the delayed regression curve model. If the results are the same, no action is taken. If the results are different, calculate the dynamic delay index QDL. k Through the dynamic latency index QDL k The magnitude determines the comparison delay result; In step S1, establishing the delayed regression curve model includes: When the resource competition An optimistic delay regression loss model is established when the delay is not affected, and the resource competitiveness is... Establish a pessimistic delay regression loss model when delay begins to be affected; The establishment of the optimistic delayed regression loss model includes: When the resource competition When the delay is not affected, the consistency Concurrency To mitigate latency, an optimistic latency regression loss model (OLRL) is established based on the characteristics of permissioned blockchains. The calculation formula for the OLRL model is as follows: ， in, In order to be in Latency value for a given number of nodes The initial delay value; The establishment of the pessimistic delayed regression loss model includes: When the resource competition When latency begins to be affected, the pessimistic latency regression loss model (PLRL) is established based on the characteristics of permissioned blockchains. The calculation formula for the pessimistic latency regression loss model (PLRL) is as follows: in, Indicates the maximum capacity of the communication channel, the The value is a fixed value and is related to the current network bandwidth and information load; In step S2, the dynamic delay index QDL is calculated. k ,include: Calculate the relative latency (RVAP) for the current number of nodes. k The formula is shown below: ， in, express With a number of nodes, the nodes arranged in ascending order are... The size of the delay value of each data point. This indicates the number of the actual points of the current node sorted in ascending order. The minimum value, This represents the total number of actual points; The relative delay value RVAP k The delay growth rate is input into the dynamic delay evaluation QDL model to obtain the dynamic delay index QDL. k The dynamic latency index QDL k The formula is shown below: ， According to the dynamic latency index QDL k The comparison results determine the delay result, the dynamic delay index QDL. k A smaller result indicates less latency, which means better performance, and vice versa, a larger result indicates greater latency, which means worse performance.

2. The delay evaluation method based on a permissioned blockchain database system according to claim 1, characterized in that, In step S1, calculating the initial delay value for a single node further includes: Each node is obtained from the delayed regression curve, and the initial delay value is calculated using Newton's interpolation method. : ， in, Indicates the first Step difference quotient This indicates the number of nodes.

3. The delay evaluation method based on a permissioned blockchain database system according to claim 2, characterized in that, In step S1, calculating the latency growth rate under the current number of nodes further includes: The delay growth rate is calculated using the delayed regression curve and the target parameter calculated by the least squares method, i.e., in the context of resource competition. When delay is not affected, the delay growth rate is obtained by applying the optimistic delay regression loss model OLRL and the first objective parameter calculated by the least squares method to the delay growth rate model. When resource competition... When the delay is affected, the pessimistic delay regression loss model PLRL and the second objective parameter calculated by the least squares method are applied to the delay growth rate model to obtain the delay growth rate.

4. The delay evaluation method based on a permissioned blockchain database system according to claim 3, characterized in that, The first target parameter further includes: The first target parameter is calculated using a first objective function, the formula for which is as follows: ， Among them, parameters and parameters , Indicates the parameter With the parameters The error between them Indicates the current node, Indicates the first For the dataset of the single node, the delay value calculated by the optimistic delayed regression loss model OLRL is applied to the least squares calculation formula as follows: 。 5. The delay evaluation method based on a permissioned blockchain database system according to claim 4, characterized in that, The second target parameter further includes: The second objective parameter is calculated using a second objective function, the formula for which is as follows: Among them, parameters ,parameter and parameters The delay value calculated by the pessimistic delayed regression loss model PLRL is applied to the least squares method to obtain... .

6. The delay evaluation method based on a permissioned blockchain database system according to claim 5, characterized in that, The delay growth rate model further includes: The formula for calculating the delay growth rate (LGR) model is as follows: ， in, The slope The target parameter is the slope, which varies depending on the chosen consensus, specifically in the cases of the optimistic delayed regression loss model OLRL and the pessimistic delayed regression loss model PLRL. There are two calculation formulas: 。

Images