Method for evaluating cloud disaster recovery service quality based on expectation weighting and fuzzy mathematics

Abstract

The invention relates to a cloud disaster recovery service quality evaluation method based on expected empowerment and fuzzy mathematics, and the method comprises the steps: obtaining service quality score data of a business system based on a preset evaluation index in response to a cloud disaster recovery service quality evaluation request initiated by a user or a preset disaster recovery service evaluation trigger signal detected by the business system; using a parabolic membership function to convert the service quality score data into a fuzzy relation matrix; and carrying out hierarchical weighting calculation by combining a preset evaluation index weight and the fuzzy relation matrix, and generating a quantitative cloud disaster recovery service comprehensive evaluation result. According to the method, the functionality, reliability and user experience of the cloud disaster recovery service are comprehensively covered by constructing a multi-dimensional hierarchical index system; an index weight is objectively determined by using a reverse cloud model, and subjective deviation is reduced; a K-order parabolic membership function is adopted to effectively model a user fuzzy score, and the evaluation accuracy is improved; and generating a traceable comprehensive score through hierarchical weighting calculation, identifying a service bottleneck, and optimizing disaster recovery performance.

Description

Technical Field

[0001] This invention belongs to the field of cloud computing service quality assessment technology, and in particular relates to a cloud disaster recovery service quality assessment method based on expectation weighting and fuzzy mathematics. Background Technology

[0002] With the rapid development of cloud computing, virtualization, and data center technologies, cloud platforms, with their high elasticity, high availability, and resource scalability, have become core infrastructures in government, healthcare, finance, and other fields. Cloud disaster recovery services provide capabilities such as data protection, fault takeover, and RPO / RTO management through cloud backup and disaster recovery functions, ensuring business continuity and data security. However, existing cloud disaster recovery service quality evaluation methods have significant shortcomings: First, the evaluation system is incomplete, focusing primarily on performance indicators such as RPO / RTO while neglecting dimensions such as functional compatibility, information security, and usability, leading to biased assessments; second, traditional weighting methods such as the Analytic Hierarchy Process (AHP) rely on subjective comparisons, which are prone to consistency bias, while the entropy weighting method ignores expert knowledge and has poor weight stability; furthermore, user ratings are often expressed in vague language, making it difficult for existing models to effectively model their uncertainty; finally, the lack of a traceable mathematical modeling mechanism results in poor repeatability of evaluation results, making it difficult to guide service optimization.

[0003] This invention proposes a cloud disaster recovery service quality evaluation method that combines reverse cloud modeling and fuzzy mathematics. It solves the problems of incomplete evaluation system, strong subjectivity of weights, and insufficient fuzzy information modeling in existing technologies. Through a multi-dimensional indicator system, objective weighting, and hierarchical evaluation, a traceable comprehensive score result is generated. Summary of the Invention

[0004] The purpose of this invention is to provide a cloud disaster recovery service quality evaluation method based on expected weighting and fuzzy mathematics, in order to solve the technical problems of incomplete existing cloud disaster recovery service quality evaluation systems, strong subjectivity in weight allocation, insufficient fuzzy modeling of user scores, and lack of traceability of results.

[0005] To achieve one of the above-mentioned objectives, one embodiment of the present invention provides a cloud disaster recovery service quality evaluation method based on expected weighting and fuzzy mathematics, the method comprising:

[0006] In response to a user-initiated cloud disaster recovery service quality evaluation request or a business system detecting a preset disaster recovery service evaluation trigger signal, the service quality score data of the business system is obtained based on preset evaluation indicators.

[0007] A parabolic membership function is used to convert service quality score data into a fuzzy relation matrix.

[0008] By combining the preset evaluation index weights with the fuzzy relation matrix, a hierarchical weighted calculation is performed to generate a quantitative comprehensive evaluation result of cloud disaster recovery services.

[0009] As a further improvement to one embodiment of the present invention, the method further includes constructing a multi-dimensional, hierarchical cloud disaster recovery service quality evaluation index system during system startup, specifically including:

[0010] Based on the preset cloud disaster recovery service quality evaluation standard, the cloud disaster recovery service quality is decomposed into multiple primary evaluation dimensions, which cover the functionality, non-functionality and user experience characteristics of the cloud disaster recovery system.

[0011] At least one secondary evaluation indicator is configured for each primary evaluation dimension. The secondary evaluation indicators obtain quantitative data through standardized testing tools or technical detection interfaces to generate an indicator data set that can be used for expert importance scoring and user service quality scoring.

[0012] The secondary evaluation indicators are mapped to specific components or operation and maintenance processes of the cloud disaster recovery system to support the localization of service quality issues;

[0013] The weights of the primary evaluation dimensions are calculated using the analytic hierarchy process (AHP) to determine their priority. The correlation between the primary evaluation dimensions and the secondary evaluation indicators is also analyzed to ensure that the indicator dataset supports the accuracy of subsequent expert and user ratings.

[0014] As a further improvement to one embodiment of the present invention, the method further includes obtaining the importance score data of the expert group for the secondary evaluation indicators, including,

[0015] Based on the set of indicator data for secondary evaluation indicators, multiple experts rank the secondary evaluation indicators under the same primary evaluation dimension by importance through a standardized online scoring system or a scoring interface provided by a cloud computing platform, and assign scores on a percentage scale based on the importance ranking to generate scoring data; wherein the scores of low-importance secondary evaluation indicators are not higher than those of high-importance secondary evaluation indicators to ensure scoring consistency.

[0016] Construct a rating matrix based on the rating data:

[0017]

[0018] in, This represents the rating matrix for the k-th primary evaluation dimension. This represents the scoring data for the m-th secondary evaluation indicator under the k-th primary evaluation dimension. This represents the score data of the nth expert for the mth secondary evaluation indicator, where m represents the number of secondary evaluation indicators and n represents the number of experts.

[0019] By calculating each row vector of the scoring matrix coefficient of variation To verify the consistency of the scoring data, the formula is:

[0020]

[0021] in, for The mean, This indicates the relative dispersion of the rating data;

[0022] like If the preset threshold is exceeded, the scoring data collection will be repeated or the scoring data will be adjusted.

[0023] As a further improvement to one embodiment of the present invention, the method further includes calculating the statistical feature parameters of the scoring data using a reverse cloud model algorithm to generate the weight of each secondary evaluation indicator, including...

[0024] The rating matrix Each row vector As input, the inverse cloud model algorithm is used to calculate statistical feature parameters, including,

[0025] Expected value, the formula is:

[0026]

[0027] in, , where is the expectation, represents the rating data. The average value reflects the core importance value of the i-th secondary evaluation indicator under the k-th primary evaluation dimension. The score given by the j-th expert to the i-th secondary evaluation indicator;

[0028] Variance, the formula is:

[0029]

[0030] in, Variance represents the volatility of the score data of the i-th secondary evaluation indicator under the k-th primary evaluation dimension, reflecting the degree of dispersion of the score data.

[0031] Entropy, the formula is:

[0032]

[0033] in, Entropy represents the uncertainty of the scoring data of the i-th secondary evaluation indicator under the k-th primary evaluation dimension, reflecting the randomness of the data;

[0034] Regarding the expected After normalization, the weight of the i-th secondary evaluation indicator under the k-th primary evaluation dimension is given by the formula:

[0035]

[0036] in, Let be the weight of the i-th secondary evaluation indicator under the k-th primary evaluation dimension. Let m be the expected value of the i-th secondary evaluation indicator under the k-th primary evaluation dimension, and m be the total number of secondary evaluation indicators under the k-th primary evaluation dimension.

[0037] Forming a weight vector ;in, Let be the weight vector of the k-th primary evaluation dimension.

[0038] As a further improvement to one embodiment of the present invention, the method further includes, in that the calculation of statistical characteristic parameters further includes,

[0039] Based on the variance Entropy The formula for calculating hyperentropy is:

[0040]

[0041] in, Hyperentropy represents the degree of consensus and stability of the rating data.

[0042] when When this occurs, hyperentropy diagnostic information is generated, indicating that the domain of the secondary evaluation index exceeds a preset threshold, triggering a refinement or redefinition operation of the secondary evaluation index.

[0043] The refinement includes decomposing secondary evaluation indicators whose scope exceeds a preset threshold into more specific sub-indicators.

[0044] As a further improvement to one embodiment of the present invention, the method further includes, in that the step of converting the service quality score data into a fuzzy relation matrix using a parabolic membership function, includes,

[0045] Based on the set of indicator data of secondary evaluation indicators, quantitative data of cloud disaster recovery services are obtained through standardized testing tools or technical testing interfaces, and service quality scores for secondary evaluation indicators are obtained through user terminals. User score data is generated using preset evaluation levels, the highest and lowest scores are removed, and the average score is calculated as the service quality score of secondary evaluation indicators.

[0046] The scoring data is mapped into a fuzzy relation matrix using a K-order parabolic membership function:

[0047]

[0048] in, This indicates the membership value between the score and the preset evaluation level;

[0049] The parabolic membership function adapts to different user rating styles by adjusting parameters, including the function exponent K, the center value, and the range width.

[0050] As a further improvement to one embodiment of the present invention, the method further includes, in the step of performing hierarchical weighted calculation by combining preset evaluation index weights and a fuzzy relation matrix to generate a quantitative comprehensive evaluation result of cloud disaster recovery services, the following steps are included:

[0051] Weight vector With fuzzy relation matrix Perform matrix multiplication to calculate the fuzzy evaluation vector, using the following formula:

[0052]

[0053] in, This indicates the evaluation results for each evaluation dimension;

[0054] Using a preset set of evaluation level scores E, the quantitative scores for each evaluation dimension are calculated using the following formula:

[0055]

[0056] in, This represents the overall score across the evaluation dimensions. The transpose of the evaluation grade score set E is used to convert the row vector E into a column vector;

[0057] Based on each primary evaluation dimension Calculate the comprehensive score of cloud disaster recovery services layer by layer. ;

[0058] according to and Identify weaknesses in service quality, including:

[0059] Compare the quantitative scores of each primary evaluation dimension with the preset thresholds, when When the score is below the threshold, the k-th primary evaluation dimension is marked as a weak point, and the overall score is taken into account. Assess the impact of weaknesses on overall service quality;

[0060] Based on the hyperentropy diagnostic information, secondary evaluation indicators whose domain range exceeds a preset threshold are identified, triggering refinement or redefinition operations to optimize the indicator system structure.

[0061] To achieve one of the above-mentioned objectives, an embodiment of the present invention also provides a cloud disaster recovery service quality evaluation system based on expected weighting and fuzzy mathematics. The system includes an indicator system construction module, a weight generation module, and a fuzzy evaluation and comprehensive calculation module.

[0062] The indicator system construction module is used to respond to user-initiated cloud disaster recovery service quality evaluation requests or business systems detecting preset disaster recovery service evaluation trigger signals, and to obtain service quality score data of the business system based on preset evaluation indicators.

[0063] The weight generation module is used to convert service quality score data into a fuzzy relation matrix using a parabolic membership function.

[0064] The comprehensive calculation module is used to perform hierarchical weighted calculation by combining preset evaluation index weights and fuzzy relation matrices to generate a quantitative comprehensive evaluation result of cloud disaster recovery services.

[0065] To achieve one of the above-mentioned objectives, an embodiment of the present invention also provides an electronic device, including a memory and a processor, characterized in that the memory stores a computer program that can run on the processor, and when the program is executed on the processor, it implements the steps in the cloud disaster recovery service quality evaluation method of expected weighting and fuzzy mathematics as described above.

[0066] To achieve one of the above-mentioned objectives, an embodiment of the present invention also provides a storage medium storing a computer program, characterized in that, when the computer program is executed by a processor, it implements the steps in the cloud disaster recovery service quality evaluation method based on expected weighting and fuzzy mathematics as described above.

[0067] Compared with existing technologies, this invention provides a cloud disaster recovery service quality evaluation method based on expected weighting and fuzzy mathematics. It comprehensively covers the functionality, reliability, and user experience of cloud disaster recovery services by constructing a multi-dimensional, hierarchical indicator system; objectively determines indicator weights using an inverse cloud model to reduce subjective bias; effectively models user fuzzy scores using a K-order parabolic membership function to improve evaluation accuracy; and generates a traceable comprehensive score through hierarchical weighted calculation to identify service bottlenecks and optimize disaster recovery performance. This method has high reproducibility and scalability in public cloud, private cloud, and hybrid cloud scenarios, providing a scientific basis for user selection and service optimization. Attached Figure Description

[0068] Figure 1 This is an overall flowchart of the cloud disaster recovery service quality evaluation method based on expected weighting and fuzzy mathematics as described in this invention.

[0069] Figure 2This is a schematic diagram of the cloud disaster recovery service quality evaluation system of the cloud disaster recovery service quality evaluation method based on expectation weighting and fuzzy mathematics described in this invention.

[0070] Figure 3 This is a schematic diagram of the parabolic membership function five-level scoring mapping of the cloud disaster recovery service quality evaluation method based on expected weighting and fuzzy mathematics described in this invention.

[0071] Figure 4 This is a schematic diagram of the architecture of the cloud disaster recovery service quality evaluation system based on expected weighting and fuzzy mathematics as described in this invention. Detailed Implementation

[0072] The present invention will now be described in detail with reference to the specific embodiments shown in the accompanying drawings. However, these embodiments do not limit the present invention, and any structural, methodological, or functional modifications made by those skilled in the art based on these embodiments are included within the scope of protection of the present invention.

[0073] Embodiments of the present invention are described in detail below. Examples of these embodiments are shown in the accompanying drawings, wherein the same or similar reference numerals denote the same or similar elements or elements having the same or similar functions throughout. The embodiments described below with reference to the accompanying drawings are exemplary and are only used to explain the present invention, and should not be construed as limiting the present invention.

[0074] In Embodiment 1 of the present invention, the present invention provides a cloud disaster recovery service quality evaluation method based on expected weighting and fuzzy mathematics, such as... Figure 1 As shown, the method includes,

[0075] S1: In response to a user-initiated cloud disaster recovery service quality evaluation request or the business system detecting a preset disaster recovery service evaluation trigger signal, obtain the service quality score data of the business system based on preset evaluation indicators.

[0076] S2: Use a parabolic membership function to convert the service quality score data into a fuzzy relation matrix;

[0077] S3: Combine the preset evaluation index weights with the fuzzy relation matrix to perform hierarchical weighted calculations and generate a quantitative comprehensive evaluation result of cloud disaster recovery services.

[0078] In one specific embodiment of the present invention, a multi-dimensional, hierarchical cloud disaster recovery service quality evaluation index system is constructed during system startup, specifically as follows:

[0079] Based on the preset cloud disaster recovery service quality evaluation standard, the cloud disaster recovery service quality is decomposed into multiple primary evaluation dimensions, which cover the functionality, non-functionality and user experience characteristics of the cloud disaster recovery system.

[0080] At least one secondary evaluation indicator is configured for each primary evaluation dimension. The secondary evaluation indicators obtain quantitative data through standardized testing tools or technical detection interfaces to generate an indicator data set that can be used for expert importance scoring and user service quality scoring.

[0081] The secondary evaluation indicators are mapped to specific components or operation and maintenance processes of the cloud disaster recovery system to support the localization of service quality issues;

[0082] The weights of the primary evaluation dimensions are calculated using the analytic hierarchy process (AHP) to determine their priority. The correlation between the primary evaluation dimensions and the secondary evaluation indicators is also analyzed to ensure that the indicator dataset supports the accuracy of subsequent expert and user ratings.

[0083] It should be noted that this invention, based on national standards (such as GB / T 29765—2021, GB / T 25000.10-2016, GB / T 37046-2018, GB / T 20988-2007) and the principles of completeness, hierarchy, and consistency in cloud disaster recovery service quality evaluation, decomposes cloud disaster recovery service quality into multiple primary evaluation dimensions, covering the functional, non-functional, and user experience characteristics of cloud disaster recovery systems.

[0084] like Figure 2 As shown, the primary evaluation dimensions include:

[0085] Basic functional requirements ( ): Covers functions such as backup object support, operating platform support, and backup mode support, ensuring the completeness of the core functions of the disaster recovery system.

[0086] Reliability requirements ( ): Focus on the stability of the system during long-term operation, such as the reliability of product functions and lifespan.

[0087] Performance efficiency requirements ( ): Evaluate the efficiency of data backup and recovery, such as recovery point objective (RPO) and recovery time objective (RTO).

[0088] Information security requirements ( ): Ensure the security of data and configurations, such as user data protection and identity authentication.

[0089] Usability requirements ( ): Measure ease of use, such as the availability of product manuals and online help.

[0090] Furthermore, at least one secondary evaluation indicator is configured for each primary evaluation dimension, ensuring that quantitative data can be obtained through standardized testing tools or technical detection interfaces, forming an indicator dataset that can be used for expert importance scoring and user service quality scoring. The generated indicator dataset is stored in a standardized data format, providing unified input data for subsequent expert scoring (through an online scoring system) and user scoring (through standardized questionnaires), ensuring the objectivity and consistency of the scoring process.

[0091] The secondary evaluation indicators under the basic functional requirements include backup object support ( ), Operating platform support ( Backup mode support () ), backup media support ( ), system management functions ( ) and additional features ( ).like Quantitative data is obtained by testing the consistency of backup and recovery data (using checksums or hash comparisons).

[0092] Secondary evaluation indicators under performance efficiency requirements include simultaneous access backup capability ( ), and simultaneously access recovery capabilities ( ), RPO ( ) and RTO ( ).like Quantitative data is obtained by calculating data gaps through simulated interruption scenarios.

[0093] These secondary evaluation metrics obtain specific data through performance testing tools, log analysis, or security testing tools to ensure the objectivity of the evaluation results, which will be used for subsequent user evaluation.

[0094] Furthermore, secondary evaluation indicators are mapped to specific components or operational processes of the cloud disaster recovery system to support the localization of service quality issues. For example:

[0095] (Backup object support) Map to the backup agent module to verify its ability to back up and restore objects such as files, databases, and virtual machines.

[0096] (RTO) is mapped to the fault takeover process, and performance bottlenecks are located by measuring the total time from disaster triggering to business recovery.

[0097] (User data protection) is mapped to the data encryption module to check the validity of static encryption algorithms (such as AES) and transport encryption protocols (such as TLS).

[0098] This mapping relationship facilitates the rapid identification of the root cause of problems during the evaluation process and guides service optimization.

[0099] Furthermore, to ensure the scientific validity and logical consistency of the cloud disaster recovery service quality evaluation index system, this invention calculates the weights of the primary evaluation dimensions using the Analytic Hierarchy Process (AHP) and analyzes the correlation between the primary evaluation dimensions and the secondary evaluation indicators to verify the rationality of the index system. The specific implementation process is as follows:

[0100] Construct a judgment matrix for primary evaluation dimensions: Invite a group of experts (usually 5 to 10 people) with professional backgrounds in cloud computing and disaster recovery to conduct pairwise comparisons of primary evaluation dimensions (including basic functional requirements, reliability requirements, performance efficiency requirements, information security requirements, and usability requirements), and construct a judgment matrix based on the relative importance scores.

[0101] Calculate the weights of the first-level evaluation dimensions: Calculate the eigenvalues ​​and eigenvectors of the judgment matrix using the eigenvector method to obtain the weights of each first-level evaluation dimension, and perform a consistency check.

[0102] Verify the correlation: Analyze the dependencies between primary evaluation dimensions and secondary evaluation indicators to verify the logic and rationality of the indicator system. For example, check the basic functional requirements ( Secondary indicators (such as) Does the support for backup objects have performance efficiency requirements? Secondary indicators (such as) The RPO (Recovery Point Objective) has functional dependencies, ensuring that the indicator system covers the core characteristics of cloud disaster recovery services. Correlation analysis is completed through expert discussion or data verification, for example, by confirming it through test data. Backup consistency support Optimize RPO. If a dependency is found to be missing (e.g., a secondary metric is unrelated to the primary dimension), adjust the metric architecture, such as removing redundant metrics or adding missing metrics.

[0103] In one specific embodiment of the present invention, the importance score data of the expert group for the secondary evaluation indicators is obtained, specifically as follows:

[0104] Based on the set of indicator data for secondary evaluation indicators, multiple experts rank the secondary evaluation indicators under the same primary evaluation dimension by importance through a standardized online scoring system or a scoring interface provided by a cloud computing platform, and assign scores on a percentage scale based on the importance ranking to generate scoring data; wherein the scores of low-importance secondary evaluation indicators are not higher than those of high-importance secondary evaluation indicators to ensure scoring consistency.

[0105] Construct a rating matrix based on the rating data:

[0106]

[0107] in, This represents the rating matrix for the k-th primary evaluation dimension. This represents the scoring data for the m-th secondary evaluation indicator under the k-th primary evaluation dimension. This represents the score data of the nth expert for the mth secondary evaluation indicator, where m represents the number of secondary evaluation indicators and n represents the number of experts.

[0108] By calculating each row vector of the scoring matrix coefficient of variation To verify the consistency of the scoring data, the formula is:

[0109]

[0110] in, for The mean, This indicates the relative dispersion of the rating data;

[0111] like If the preset threshold is exceeded, the scoring data collection will be repeated or the scoring data will be adjusted.

[0112] It should be noted that, in order to ensure the objectivity and reliability of the weights of the secondary evaluation indicators in the quality evaluation of cloud disaster recovery services, this invention collects expert scoring data through standardized questionnaires or online scoring systems, constructs a scoring matrix, and verifies consistency through the coefficient of variation, providing high-quality input for subsequent reverse cloud model algorithms.

[0113] Furthermore, expert scoring is based on a dataset of secondary evaluation indicators and is performed through a standardized online scoring system or a scoring interface provided by a cloud computing platform. An expert panel (typically consisting of 7-10 professionals with experience in cloud computing or disaster recovery) ranks the secondary evaluation indicators defined above according to their importance. The standardized online scoring system uses a unified question template and scoring criteria (e.g., a percentage system, ranging from [0, 100]), clearly defining the secondary indicators and their importance evaluation requirements. For example, regarding basic functional requirements (… Secondary indicators The question regarding "(Backup Object Support)" is "How important is the completeness of backup object support to cloud disaster recovery services?" Experts first ranked the importance of secondary indicators within the same primary evaluation dimension, for example... (Backup object support) > (Backup frequency) >…> (Additional functionality). Subsequently, based on this ranking, experts assign scores on a percentage basis, ensuring that the scores of lower importance indicators do not exceed those of higher importance indicators. For example, It could be assigned a value of [95, 90, 95, 95, 90, 98, 90], while The scores are [50, 55, 50, 60, 55, 50, 50]. The online scoring system supports real-time data entry, validity verification (checking the scoring range), and scoring constraint verification to ensure a standardized scoring process. Scoring data is stored on a cloud computing platform, supporting concurrent operations by multiple users.

[0114] Furthermore, based on the rating data, a rating matrix for the k-th primary evaluation dimension is constructed. Furthermore, to ensure the reliability of the scoring data, the vector of each row of the scoring matrix is ​​calculated. The coefficient of variation (CV) is used to verify the consistency of expert ratings. If the CV exceeds a preset threshold, it indicates that the rating data is too dispersed, and the system triggers the re-collection of rating data or the organization of expert discussions to reach a consensus. The verification process is implemented through a cloud computing platform, supporting multi-threaded processing and adapting to the rating verification needs of large-scale indicator systems.

[0115] It should be noted that standardized questionnaires or online scoring systems effectively reduce the subjectivity of expert scoring by using unified question templates, scoring standards, and constraints (scores for low-importance indicators should not exceed those for high-importance indicators). Coefficient of variation (CV) validation ensures the consistency and stability of the scoring data, providing reliable input for the reverse cloud model. The entire process is implemented in a cloud computing environment and is applicable to disaster recovery service quality evaluation scenarios in public, private, or hybrid clouds.

[0116] In one specific implementation scenario of this invention, seven industry experts were invited to rank the importance of the corresponding indicators in the cloud disaster recovery service quality evaluation system, and then assign a percentage-based score to each indicator based on its importance. The primary indicators are listed below. For example, the first-level indicator The rating vector can be denoted as Secondary indicators The rating row vector can be denoted as As shown in the following formula.

[0117]

[0118] This means that 7 experts commented on the indicators respectively. The importance of the score is represented by a row vector.

[0119] In one specific embodiment of the present invention, the statistical characteristic parameters of the scoring data are calculated using the inverse cloud model algorithm to generate the weight of each secondary evaluation indicator. Specifically,

[0120] The rating matrix Each row vector As input, the inverse cloud model algorithm is used to calculate statistical feature parameters, including,

[0121] Expected value, the formula is:

[0122]

[0123] in, , where is the expectation, represents the rating data. The average value reflects the core importance value of the i-th secondary evaluation indicator under the k-th primary evaluation dimension. The score given by the j-th expert to the i-th secondary evaluation indicator;

[0124] Variance, the formula is:

[0125]

[0126] in, Variance represents the volatility of the score data of the i-th secondary evaluation indicator under the k-th primary evaluation dimension, reflecting the degree of dispersion of the score data.

[0127] Entropy, the formula is:

[0128]

[0129] in, Entropy represents the uncertainty of the scoring data of the i-th secondary evaluation indicator under the k-th primary evaluation dimension, reflecting the randomness of the data;

[0130] Regarding the expected After normalization, the weight of the i-th secondary evaluation indicator under the k-th primary evaluation dimension is given by the formula:

[0131]

[0132] in, Let be the weight of the i-th secondary evaluation indicator under the k-th primary evaluation dimension. Let m be the expected value of the i-th secondary evaluation indicator under the k-th primary evaluation dimension, and m be the total number of secondary evaluation indicators under the k-th primary evaluation dimension.

[0133] Forming a weight vector ;in, Let be the weight vector of the k-th primary evaluation dimension.

[0134] Based on the variance Entropy The formula for calculating hyperentropy is:

[0135]

[0136] in, Hyperentropy represents the degree of consensus and stability of the rating data.

[0137] when When this occurs, hyperentropy diagnostic information is generated, indicating that the domain of the secondary evaluation index exceeds a preset threshold, triggering a refinement or redefinition operation of the secondary evaluation index.

[0138] The refinement includes decomposing secondary evaluation indicators whose scope exceeds a preset threshold into more specific sub-indicators.

[0139] In one specific embodiment of the present invention, ~ The importance score row vectors of the indicators are used as sample point inputs to the inverse cloud model, and the output results can be obtained. ~ A total of 6 sets of expectations ,variance ,entropy hyperentropy The parameters for each indicator are shown in the table below. Additional functions can be added to the table. Indicator parameter H e Value greater than It can be seen that the scope of this indicator is relatively large, and further refinement and evaluation adjustments can be made in the future.

[0140]

[0141] Furthermore, by normalizing the columns using the expected values ​​obtained from the inverse cloud model in Table 1, the importance weight vector of the secondary indicators is obtained. As shown below.

[0142]

[0143] Will The importance scores of the corresponding secondary indicators are used as input to the inverse cloud model to obtain the corresponding cloud model parameters, as shown in Table 2. The weight parameters of the primary and secondary indicators calculated from these indicators are shown in Table 3.

[0144]

[0145] A questionnaire was used to rate the service quality of the cloud disaster recovery service used by a certain organization. The scores for each indicator are shown in Table 3. The importance weights of the primary and secondary indicators were derived by normalizing the expected data column in Table 2. The data in the secondary indicator rating column are percentage scores given by users based on the actual situation of the cloud disaster recovery service used by the organization. ~ This refers to the membership value in the membership function of the secondary indicator score.

[0146]

[0147] In one specific embodiment of the present invention, a parabolic membership function is used to convert service quality score data into a fuzzy relation matrix, specifically, as follows:

[0148] Based on the set of indicator data of secondary evaluation indicators, quantitative data of cloud disaster recovery services are obtained through standardized testing tools or technical testing interfaces, and service quality scores for secondary evaluation indicators are obtained through user terminals. User score data is generated using preset evaluation levels, the highest and lowest scores are removed, and the average score is calculated as the service quality score of secondary evaluation indicators.

[0149] The scoring data is mapped into a fuzzy relation matrix using a K-order parabolic membership function:

[0150]

[0151] in, This indicates the membership value between the score and the preset evaluation level;

[0152] The parabolic membership function adapts to different user rating styles by adjusting parameters, including the function exponent K, the center value, and the range width.

[0153] It should be noted that the data set of indicators based on secondary evaluation metrics is obtained through standardized testing tools or technical testing interfaces to acquire quantitative data on cloud disaster recovery services. For example, RPO is calculated through data gap analysis, and RTO is calculated through disaster recovery test timing. This quantitative data is stored in the cloud computing platform's database, serving as the basis for user ratings. Users rate the service quality of secondary evaluation metrics through online questionnaires via client applications (such as web interfaces or mobile applications), using preset rating levels [Excellent, Good, Satisfactory, Poor, Very Poor], corresponding to score ranges [100-70, 90-50, 70-30, 50-10, 30-0]. To reduce subjective bias, the system removes the highest and lowest scores for each indicator and calculates the average of the remaining scores as the service quality score for that secondary evaluation indicator.

[0154] In a specific embodiment of the present invention, the evaluation method is as follows:

[0155] Determine the evaluation level. User technical personnel evaluate the quality of the cloud disaster recovery service adopted by the organization based on the secondary indicators. The levels include: [Excellent, Good, Satisfactory, Poor, Poor].

[0156] Detailed evaluation rating scores. User technical personnel should base their assessments on the actual quality of cloud disaster recovery services, as shown in the table below.

[0157]

[0158] The highest and lowest scores are discarded from the specific scores given by the user's technical staff, and the average of the remaining scores is the score value of that secondary indicator.

[0159] Furthermore, in the comprehensive quality evaluation of cloud disaster recovery services, the evaluation objects include both precisely measurable technical indicators (such as RPO, RTO, and bandwidth utilization) and experience indicators that rely on user subjective perception (such as switching convenience, operation and maintenance response speed, and overall satisfaction). The latter often appears in vague language such as "excellent," "good," and "qualified," which is difficult to directly map to precise values, and different user groups have significant differences in scoring scales. To ensure that the evaluation system can reflect the continuous changes in subjective perception and be uniformly quantified with objective indicators, this invention selects a K-order parabolic distribution as the fuzzy membership function, smoothly mapping the user's fuzzy language to the [0,1] interval, realizing a continuous transition from fuzzy semantics to the numerical domain.

[0160] This method can adapt to different user scoring styles by adjusting the exponent K, center value c, and half-width w of the parabolic function. For example, for a group that tends to give lenient scores, a larger distribution can be selected to expand the membership range of high-scoring segments; for a group with neutral scores, a symmetrical intermediate distribution can be used; and for a group with strict scores, a smaller distribution can be selected to enhance the discriminative power of low-scoring segments. In cloud disaster recovery evaluation scenarios, this adjustability can not only eliminate systematic biases caused by group differences, but also maintain the stability and comparability of results in multi-cycle, multi-batch quality monitoring.

[0161] Preferably, this invention uses a K-order parabolic distribution as the fuzzy membership function. Experiments have verified that when K is 1.2, A... v The membership function exhibits good convergence. Table 3 shows A. v The expression is as follows:

[0162] The membership function for a skewed large distribution at level v1 is as follows:

[0163]

[0164] The intermediate distributions are v2, v3, and v4, with the following membership functions:

[0165]

[0166]

[0167]

[0168] For a small-scale distribution at level v5, the membership function can be expressed as follows:

[0169]

[0170] like Figure 3 As shown, this is a schematic diagram of the parabolic membership function mapping curve for the five evaluation levels of "Poor—Fairly Poor—Qualified—Good—Excellent" when K=1.2. Different colored curves in the figure correspond to the changes in membership degree for each level, with the dashed line representing the center value. This curve intuitively reflects the continuous mapping relationship from fuzzy semantics to the numerical domain and demonstrates the characteristics of the parabolic function: high membership degree in the central region and smooth decay at the boundary. This allows for a balance between subtle changes in user subjective perception and the stability of the results in cloud disaster recovery quality evaluation.

[0171] In one specific embodiment of the present invention, a hierarchical weighted calculation is performed by combining preset evaluation index weights with a fuzzy relation matrix to generate a quantitative comprehensive evaluation result of cloud disaster recovery services. Specifically,

[0172] Weight vector With fuzzy relation matrix Perform matrix multiplication to calculate the fuzzy evaluation vector, using the following formula:

[0173]

[0174] in, This indicates the evaluation results for each evaluation dimension;

[0175] Using a preset set of evaluation level scores E, the quantitative scores for each evaluation dimension are calculated using the following formula:

[0176]

[0177] in, This represents the overall score across the evaluation dimensions. The transpose of the evaluation grade score set E is used to convert the row vector E into a column vector;

[0178] Based on each primary evaluation dimension Calculate the comprehensive score of cloud disaster recovery services layer by layer. ;

[0179] according to and Identify weaknesses in service quality, including:

[0180] Compare the quantitative scores of each primary evaluation dimension with the preset thresholds, when When the score is below the threshold, the k-th primary evaluation dimension is marked as a weak point, and the overall score is taken into account. Assess the impact of weaknesses on overall service quality;

[0181] Based on the hyperentropy diagnostic information, secondary evaluation indicators whose domain range exceeds a preset threshold are identified, triggering refinement or redefinition operations to optimize the indicator system structure.

[0182] In a specific embodiment of the present invention, the fuzzy evaluation vector of C1 The calculations are as follows: 0.20 0.20 0.11 0.15 0.08

[0183]

[0184] Similarly, we can conclude that:

[0185]

[0186]

[0187]

[0188]

[0189] Based on the above derivation, we can derive the weighted evaluation value V of the primary indicator and the comprehensive evaluation score of a certain unit's cloud disaster recovery service system. As shown in the table below

[0190]

[0191] As can be seen from Table 5, the overall quality assessment value of this cloud disaster recovery service is... The score is 74.64, which falls under the "good" level of service quality. Performance efficiency requirements... Evaluation value Compared to other performance indicators, this is slightly weak. Overall network service efficiency is low, which can lead to a poor user experience and cause service lag or other issues. Future cloud disaster recovery services could improve performance efficiency. In terms of service performance and service quality. From the perspective of information security requirements indicators. A score of 76.53 indicates that the user base of this organization places considerable importance on the protection of personal information, and the organization is in a relatively secure state. The overall score... As can be seen from the above, there is still room for improvement in the quality of this cloud disaster recovery service; it is sufficient to meet the development needs of cloud disaster recovery services.

[0192] In a second embodiment of the present invention, the present invention provides a cloud disaster recovery service quality evaluation system based on expected weighting and fuzzy mathematics, such as... Figure 4 As shown, the system includes an indicator system construction module 1, a weight generation module, a fuzzy evaluation module 2, and a comprehensive calculation module 3;

[0193] The indicator system construction module 1 is used to respond to a user-initiated cloud disaster recovery service quality evaluation request or the business system detecting a preset disaster recovery service evaluation trigger signal, and obtain the service quality score data of the business system based on the preset evaluation indicators.

[0194] The weight generation module 2 is used to convert service quality score data into a fuzzy relation matrix using a parabolic membership function;

[0195] The comprehensive calculation module 3 is used to perform hierarchical weighted calculation by combining preset evaluation index weights and fuzzy relation matrices to generate a quantitative comprehensive evaluation result of cloud disaster recovery services.

[0196] In a third embodiment of the present invention, the present invention provides an electronic device, including a memory and a processor, characterized in that the memory stores a computer program that can run on the processor, and the program executed on the processor implements the steps in the cloud disaster recovery service quality evaluation method of expected weighting and fuzzy mathematics as described above.

[0197] In Embodiment 4 of the present invention, the present invention provides a storage medium storing a computer program, characterized in that, when the computer program is executed by a processor, it implements the steps in the cloud disaster recovery service quality evaluation method of expected weighting and fuzzy mathematics as described above.

[0198] In summary, the invention provides a cloud disaster recovery service quality evaluation method based on expectation weighting and fuzzy mathematics. This method comprehensively covers the functionality, reliability, and user experience of cloud disaster recovery services by constructing a multi-dimensional, hierarchical indicator system. It objectively determines indicator weights using a reverse cloud model, reducing subjective bias. A K-order parabolic membership function is employed to effectively model user fuzzy scores, improving evaluation accuracy. Hierarchical weighted calculations generate a traceable comprehensive score, identifying service bottlenecks and optimizing disaster recovery performance. This method demonstrates high reproducibility and scalability in public, private, and hybrid cloud scenarios, providing a scientific basis for user selection and service optimization.

[0199] Those skilled in the art will understand that, for the sake of convenience and brevity, the specific working process of the modules described above can be referred to the corresponding process in the aforementioned method implementation, and will not be repeated here.

[0200] The modules described as separate components may or may not be physically separate. Similarly, the components shown as modules may or may not be physical modules; they may be located in one place or distributed across multiple network modules. Some or all of the modules can be selected to achieve the purpose of this embodiment, depending on actual needs.

[0201] Furthermore, the functional modules in the various embodiments of this application can be integrated into one processing module, or each module can exist physically separately, or two or more modules can be integrated into one module. The integrated module can be implemented in hardware or in a combination of hardware and software functional modules.

[0202] The integrated modules implemented as software functional modules described above can be stored in a computer-readable storage medium. These software functional modules, stored in a storage medium, include several instructions to cause a computer system (which may be a personal computer, server, or network system, etc.) or processor to execute some steps of the methods described in the various embodiments of this application. The aforementioned storage medium includes various media capable of storing program code, such as USB flash drives, portable hard drives, read-only memory (ROM), random access memory (RAM), magnetic disks, or optical disks.

[0203] Finally, it should be noted that the above embodiments are only used to illustrate the technical solutions of this application, and are not intended to limit them. Although this application has been described in detail with reference to the foregoing embodiments, those skilled in the art should understand that modifications can still be made to the technical solutions described in the foregoing embodiments, or equivalent substitutions can be made to some of the technical features. Such modifications or substitutions do not cause the essence of the corresponding technical solutions to deviate from the spirit and scope of the technical solutions of the embodiments of this application.

Claims

1. A cloud disaster recovery service quality evaluation method based on expectation weighting and fuzzy mathematics, characterized in that: include, In response to a user-initiated cloud disaster recovery service quality evaluation request or a business system detecting a preset disaster recovery service evaluation trigger signal, the service quality score data of the business system is obtained based on preset evaluation indicators. A parabolic membership function is used to convert service quality score data into a fuzzy relation matrix. By combining the preset evaluation index weights with the fuzzy relation matrix, a hierarchical weighted calculation is performed to generate a quantitative comprehensive evaluation result of cloud disaster recovery services.

2. The cloud disaster recovery service quality evaluation method based on expectation weighting and fuzzy mathematics according to claim 1, characterized in that: It also includes, A multi-dimensional, hierarchical cloud disaster recovery service quality evaluation index system is constructed during system startup, specifically including: Based on the preset cloud disaster recovery service quality evaluation standard, the cloud disaster recovery service quality is decomposed into multiple primary evaluation dimensions, which cover the functionality, non-functionality and user experience characteristics of the cloud disaster recovery system. At least one secondary evaluation indicator is configured for each primary evaluation dimension. The secondary evaluation indicators obtain quantitative data through standardized testing tools or technical detection interfaces to generate an indicator data set that can be used for expert importance scoring and user service quality scoring. The secondary evaluation indicators are mapped to specific components or operation and maintenance processes of the cloud disaster recovery system to support the localization of service quality issues; The weights of the primary evaluation dimensions are calculated using the analytic hierarchy process (AHP) to determine their priority. The correlation between the primary evaluation dimensions and the secondary evaluation indicators is also analyzed to ensure that the indicator dataset supports the accuracy of subsequent expert and user ratings.

3. The cloud disaster recovery service quality evaluation method based on expectation weighting and fuzzy mathematics according to claim 2, characterized in that: Obtain the importance score data of the expert group for the secondary evaluation indicators, including, Based on the set of indicator data of secondary evaluation indicators, multiple experts rank the secondary evaluation indicators under the same primary evaluation dimension by importance through the scoring interface provided by a standardized online scoring system or cloud computing platform, and assign scores on a percentage scale based on the importance ranking to generate scoring data. The scores of low-importance secondary evaluation indicators should not be higher than those of high-importance secondary evaluation indicators to ensure scoring consistency. Construct a rating matrix based on the rating data: ； in, This represents the rating matrix for the k-th primary evaluation dimension. This represents the scoring data for the m-th secondary evaluation indicator under the k-th primary evaluation dimension. This represents the score data of the nth expert for the mth secondary evaluation indicator, where m represents the number of secondary evaluation indicators and n represents the number of experts. By calculating each row vector of the scoring matrix coefficient of variation To verify the consistency of the scoring data, the formula is: ； in, for The mean, This indicates the relative dispersion of the rating data; like If the preset threshold is exceeded, the scoring data collection or scoring data will be re-executed through the online scoring system.

4. The cloud disaster recovery service quality evaluation method based on expectation weighting and fuzzy mathematics according to claim 3, characterized in that: The statistical characteristic parameters of the scoring data are calculated using the inverse cloud model algorithm to generate the weight of each secondary evaluation indicator. include, The scoring matrix Each row vector As input, the inverse cloud model algorithm is used to calculate statistical feature parameters, including, Expected value, the formula is: ； in, The expected value is the core value reflecting the importance of the i-th secondary evaluation indicator under the k-th primary evaluation dimension. The score given by the j-th expert to the i-th secondary evaluation indicator; Variance, the formula is: ； in, Variance represents the volatility of the score data of the i-th secondary evaluation indicator under the k-th primary evaluation dimension, reflecting the degree of dispersion of the score data. Entropy, the formula is: ； in, Entropy represents the uncertainty of the scoring data of the i-th secondary evaluation indicator under the k-th primary evaluation dimension, reflecting the randomness of the data; Regarding the expected After normalization, the weight of the i-th secondary evaluation indicator under the k-th primary evaluation dimension is given by the formula: ； in, Let be the weight of the i-th secondary evaluation indicator under the k-th primary evaluation dimension. Let m be the expected value of the i-th secondary evaluation indicator under the k-th primary evaluation dimension, and m be the total number of secondary evaluation indicators under the k-th primary evaluation dimension. Forming a weight vector ;in, Let be the weight vector of the k-th primary evaluation dimension.

5. The cloud disaster recovery service quality evaluation method based on expected weighting and fuzzy mathematics according to claim 4, characterized in that: The calculated statistical characteristic parameters also include, Based on the variance Entropy The formula for calculating hyperentropy is: ； in, Hyperentropy represents the degree of consensus and stability of the rating data. when When this occurs, hyperentropy diagnostic information is generated, indicating that the domain of the secondary evaluation index exceeds a preset threshold, triggering a refinement or redefinition operation of the secondary evaluation index. The refinement includes decomposing secondary evaluation indicators whose scope exceeds a preset threshold into more specific sub-indicators.

6. The cloud disaster recovery service quality evaluation method based on expectation weighting and fuzzy mathematics according to claim 2, characterized in that: The process of converting service quality score data into a fuzzy relation matrix using a parabolic membership function includes... Based on the set of indicator data of secondary evaluation indicators, quantitative data of cloud disaster recovery services are obtained through standardized testing tools or technical testing interfaces, and service quality scores for secondary evaluation indicators are obtained through user terminals. User score data is generated using preset evaluation levels, the highest and lowest scores are removed, and the average score is calculated as the service quality score of secondary evaluation indicators. The scoring data is mapped into a fuzzy relation matrix using a K-order parabolic membership function: ； in, This indicates the membership value between the score and the preset evaluation level; The parabolic membership function adapts to different user rating styles by adjusting parameters, including the function exponent K, the center value, and the range width.

7. The cloud disaster recovery service quality evaluation method based on expected weighting and fuzzy mathematics according to claim 5 or 6, characterized in that: The step of combining preset evaluation index weights with a fuzzy relation matrix to perform hierarchical weighted calculations to generate a quantitative comprehensive evaluation result of cloud disaster recovery services includes: Weight vector With fuzzy relation matrix Perform matrix multiplication to calculate the fuzzy evaluation vector, using the following formula: ； in, This indicates the evaluation results for each evaluation dimension; Using a preset set of evaluation level scores E, the quantitative scores for each evaluation dimension are calculated using the following formula: ； in, This represents the overall score across the evaluation dimensions. The transpose of the evaluation grade score set E is used to convert the row vector E into a column vector; Based on each primary evaluation dimension Calculate the comprehensive score of cloud disaster recovery services layer by layer. ; according to and Identify weaknesses in service quality, including: Compare the quantitative scores of each primary evaluation dimension with the preset thresholds, when When the score is below the threshold, the k-th primary evaluation dimension is marked as a weak point, and the overall score is taken into account. Assess the impact of weaknesses on overall service quality; Based on the hyperentropy diagnostic information, secondary evaluation indicators whose domain range exceeds a preset threshold are identified, triggering refinement or redefinition operations to optimize the indicator system structure.

8. A cloud disaster recovery service quality evaluation system based on expectation weighting and fuzzy mathematics, characterized in that: It includes an indicator system construction module, a weight generation module, and a fuzzy evaluation and comprehensive calculation module; The indicator system construction module is used to respond to user-initiated cloud disaster recovery service quality evaluation requests or business systems detecting preset disaster recovery service evaluation trigger signals, and to obtain service quality score data of the business system based on preset evaluation indicators. The weight generation module is used to convert service quality score data into a fuzzy relation matrix using a parabolic membership function. The comprehensive calculation module is used to perform hierarchical weighted calculation by combining preset evaluation index weights and fuzzy relation matrices to generate a quantitative comprehensive evaluation result of cloud disaster recovery services.

9. An electronic device, comprising a memory and a processor, characterized in that: The memory stores a computer program that can run on the processor, and when the program is executed on the processor, it implements the steps in the cloud disaster recovery service quality evaluation method based on expected weighting and fuzzy mathematics as described in any one of claims 1-7.

10. A storage medium storing a computer program, characterized in that: When the computer program is executed by the processor, it implements the steps in the cloud disaster recovery service quality evaluation method based on expected weighting and fuzzy mathematics as described in any one of claims 1-7.

Images