Operation and maintenance twin modeling method based on hierarchical entity decomposition and incremental data synchronization

Abstract

The application relates to the technical field of operation and maintenance twin modeling, and discloses an operation and maintenance twin modeling method based on hierarchical entity decomposition and incremental data synchronization, which comprises the following steps: collecting basic information of various entities in an operation and maintenance scene, dividing entity levels according to data update frequency and functional importance, matching corresponding modeling accuracy and data collection range for each level, building an initial operation and maintenance twin model, sorting out business association relationships among entities, and generating an entity association graph; collecting entity operation data in real time, comparing and extracting different parts as update data, and completing model synchronization update according to the transmission mode corresponding to the entity level; and statistically analyzing synchronization delay, mapping error and resource occupation data in the synchronization update process, adjusting the entity level according to the statistical result, and adjusting the update mode of upstream and downstream associated entities along the entity association graph when the level changes. The application can realize dynamic optimal balance between mapping accuracy of the twin model and operation resource consumption in a complex and heterogeneous operation and maintenance scene.

Description

Technical Field

[0001] This application relates to the technical field of operation and maintenance twin modeling, and in particular to an operation and maintenance twin modeling method based on hierarchical entity deconstruction and incremental data synchronization. Background Technology

[0002] As enterprises deepen their digital transformation, the complexity of information system operation and maintenance scenarios continues to increase. Digital twin technology, with its virtual-real mapping and dynamic simulation capabilities, is gradually becoming an important supporting technology for operation and maintenance status visualization, fault location, and process simulation. Operation and maintenance scenarios cover a wide range of objects, including business systems, data links, and service components. The data sources are diverse and their update characteristics vary greatly, which places higher demands on the efficiency of twin model construction, real-time updates, and resource adaptability.

[0003] Existing O&M twin modeling technologies mostly adopt a unified modeling standard and full data synchronization mode, which cannot be configured differently according to the operating characteristics and business value of different O&M objects. This can easily lead to redundant consumption of computing power and bandwidth resources. At the same time, it is difficult to adaptively adjust the model configuration with the dynamic changes of the O&M scenario, and it is impossible to balance model mapping accuracy and operating efficiency.

[0004] As can be seen from the above, how to achieve a dynamic optimal balance between the mapping accuracy of the twin model and the consumption of runtime resources in complex and heterogeneous operation and maintenance scenarios is a technical problem that urgently needs to be solved in the field of operation and maintenance twin modeling. Summary of the Invention

[0005] In order to achieve a dynamic optimal balance between the mapping accuracy of twin models and the consumption of runtime resources in complex and heterogeneous operation and maintenance scenarios, this application provides an operation and maintenance twin modeling method based on hierarchical entity deconstruction and incremental data synchronization.

[0006] Firstly, this application provides an operation and maintenance twin modeling method based on hierarchical entity deconstruction and incremental data synchronization, employing the following technical solution: An operation and maintenance twin modeling method based on hierarchical entity deconstruction and incremental data synchronization includes: Collect basic information on various entities within the operation and maintenance scenario, divide the entity into levels according to data update frequency and functional importance, match the corresponding modeling accuracy and data collection range for each entity level, build an initial operation and maintenance twin model, sort out the business relationships between entities, and generate an entity relationship graph. Real-time collection of operational data from each entity; comparison of operational data with corresponding content in the operation and maintenance twin model; extraction of discrepancies as update data; and synchronization of the operation and maintenance twin model based on the update data according to the transmission method corresponding to the level to which the entity belongs. Continuously collect statistics on synchronization delay, mapping error, and resource usage during the synchronous update process of the operation and maintenance twin model, obtain corresponding statistical results, adjust the hierarchical level of the corresponding entity based on the statistical results, and adjust the update method of upstream and downstream related entities along the entity association graph when the entity hierarchical level changes.

[0007] Optionally, the step of dividing entity levels according to data update frequency and functional importance, and matching corresponding modeling accuracy and data collection range to each entity level, includes: Entities are divided into high-frequency entities (update frequency above a preset threshold) and low-frequency entities (update frequency below a preset threshold) based on data update frequency. Entities are also divided into core business-related entities and auxiliary business-related entities based on functional importance. The two dimensions are combined to form four entity levels. Differentiated modeling granularity and data collection frequency are set for each group of entity levels. The modeling granularity and data collection frequency for core business-related entities are higher than those for auxiliary business-related entities, and the update transmission priority for high-frequency entities is higher than that for low-frequency entities.

[0008] Optionally, the step of comparing the operational data with the corresponding content in the operation and maintenance twin model and extracting the differing parts as update data includes: For high-frequency core business related entities, a continuous sliding window method is used to compare the running data with the corresponding content of the model segment by segment. Data segments that exceed the preset fluctuation range are judged as the difference part. For low-frequency core business related entities, a content segment hash comparison method is used to identify version changes, and the segmented content that has changed is determined to be the difference part; For entities related to auxiliary business, a periodic fixed-point comparison method is used to identify data changes, and the corresponding differences are extracted only at preset comparison points.

[0009] Optionally, the steps of adjusting the hierarchical level of the corresponding entity based on statistical results, and adjusting the update method of upstream and downstream related entities along the entity association graph when the entity hierarchical level changes, include: When an entity is adjusted from the auxiliary business association level to the core business association level, the upstream and downstream related entities are traced along the entity association graph. For related entities with association strength higher than the preset threshold, their modeling granularity is improved, transmission priority is updated, and data collection frequency is increased simultaneously. When an entity is adjusted from the core business association level to the auxiliary business association level, upstream and downstream related entities are traced along the entity association graph. For related entities with association strength higher than the preset threshold, their modeling granularity is reduced, transmission priority and data collection frequency are updated simultaneously.

[0010] Optionally, the step of sorting out the business relationships between entities and generating an entity relationship graph includes: Analyze the business dependency links and data interaction flows between various entities in the operation and maintenance scenario, and clarify the upstream and downstream interaction directions corresponding to each group of interacting entities; The total amount of data interaction and the number of business calls between entities within a statistical unit period are combined with the proportion of business dependencies and the proportion of data interaction to calculate the correlation strength between entities. For each group of entities, a corresponding association strength label is matched according to the upstream and downstream interaction direction to form an entity association map with association strength labels.

[0011] Optionally, the step of adjusting the hierarchy of the corresponding entity based on the statistical results includes: Preset mapping error threshold, synchronization delay threshold and resource consumption threshold, normalize the statistically obtained synchronization delay, mapping error and resource consumption data to obtain the comprehensive score of the corresponding entity's synchronization operation; The system continuously tracks the synchronous operation comprehensive score of entities over multiple statistical periods, and records entities whose scores are consistently lower than the preset downward adjustment threshold or consistently higher than the preset upward adjustment threshold. Entities whose scores consistently fall below the downgrade threshold will be upgraded to a higher level, while entities whose scores consistently rise above the upgrade threshold will be downgraded to a lower level.

[0012] Optionally, the step of synchronizing and updating the operation and maintenance twin model based on the updated data and the transmission method corresponding to the level to which the entity belongs includes: For high-frequency core business related entities, the extracted differential data is processed by lightweight encoding and then the synchronous update of the operation and maintenance twin model is completed through a low-latency transmission channel. For low-frequency core business related entities, the extracted differential data is processed by differential packaging and then the synchronous update of the operation and maintenance twin model is completed through a reliable transmission channel. For auxiliary business-related entities, after batch compression of all differential data generated within the same statistical period, the operation and maintenance twin model is synchronously updated through a general transmission channel.

[0013] Optionally, it also includes: Real-time detection of business topology changes within the operation and maintenance scenario; identification of all entities whose call relationships have been added, disconnected, or changed; and formation of a set of affected entities. Based on the changed business call flow and interaction frequency, the association strength values ​​between affected entities are recalculated, and the corresponding content of the entity association graph is updated synchronously. By combining the updated association strength and the importance of entity functions, the hierarchical level of the affected entities is re-determined, and the modeling accuracy and data collection range of the corresponding entities are adjusted accordingly.

[0014] Optionally, it also includes: Receive alarm information in the operation and maintenance scenario, locate the target entity corresponding to the alarm information, and retrieve the upstream and downstream related entities of the target entity in the entity association graph; Temporarily upgrade the target entity's hierarchical level, and simultaneously adjust the modeling accuracy and data collection range of upstream and downstream related entities whose association strength is higher than a preset threshold.

[0015] Once the alarm is cleared and the entity's operating status stabilizes, the target entity and its associated entities' hierarchical level, modeling accuracy, and data collection range will be restored to their pre-alarm state.

[0016] Secondly, this application provides an operation and maintenance twin modeling system based on hierarchical entity deconstruction and incremental data synchronization, which adopts the following technical solution: An operation and maintenance twin modeling system based on hierarchical entity deconstruction and incremental data synchronization includes: The hierarchical modeling module collects basic information of various entities in the operation and maintenance scenario, divides the entity into levels according to the data update frequency and functional importance, matches the corresponding modeling accuracy and data collection range for each entity level, builds an initial operation and maintenance twin model, sorts out the business relationship between each entity, and generates an entity relationship graph. The incremental synchronization module collects the operational data of each entity in real time, compares the operational data with the corresponding content in the operation and maintenance twin model, extracts the parts with differences as update data, and completes the synchronous update of the operation and maintenance twin model according to the transmission method corresponding to the level to which the entity belongs based on the update data. The dynamic optimization module continuously collects statistics on synchronization delay, mapping error, and resource usage during the synchronous update process of the operation and maintenance twin model, obtains corresponding statistical results, adjusts the hierarchical level of the corresponding entity based on the statistical results, and adjusts the update method of upstream and downstream related entities along the entity association graph when the entity hierarchical level changes.

[0017] In summary, this application includes at least one of the following beneficial technical effects: By categorizing entities in the operation and maintenance scenario into hierarchical levels based on data update frequency and functional importance, and matching differentiated modeling precision and data collection scope, precise allocation of resources is achieved during the modeling phase. This ensures the mapping accuracy of core business-related entities and reduces redundant resource consumption by non-core entities. Combined with an incremental synchronization mode that only transmits differentiated content, the computational power and bandwidth consumption for data transmission and model updates are significantly reduced. Furthermore, by continuously and dynamically adjusting entity levels through the statistical analysis of synchronized running data, and by relying on the entity association graph to synchronously link the update configurations of upstream and downstream entities, the model configuration adaptively adjusts according to the operational status of the operation and maintenance scenario. This achieves a dynamic optimal balance between the mapping accuracy of the twin model and the consumption of operational resources in complex and heterogeneous operation and maintenance scenarios.

[0018] Based on this, the efficiency and reliability of incremental synchronization are further improved by using differentiated detection methods and transmission channels for matching entities at different levels. For dynamic scenarios such as changes in business topology and operation and maintenance alarms, the corresponding adjustments to entity association graphs, entity levels and model configurations are automatically completed, enhancing the solution's adaptability to dynamic changes in operation and maintenance scenarios. Under the premise of ensuring consistency between virtual and real mapping, resource utilization efficiency and model response speed are continuously optimized. Attached Figure Description

[0019] Figure 1 This is a flowchart illustrating an operation and maintenance twin modeling method based on hierarchical entity deconstruction and incremental data synchronization, according to an exemplary embodiment.

[0020] Figure 2 This is a flowchart of steps S101 to S103 in an operation and maintenance twin modeling method based on hierarchical entity deconstruction and incremental data synchronization, according to an exemplary embodiment.

[0021] Figure 3 This is a flowchart of steps S201 to S203 in an operation and maintenance twin modeling method based on hierarchical entity deconstruction and incremental data synchronization, according to an exemplary embodiment.

[0022] Figure 4 This is a flowchart of steps S301 to S303 in an operation and maintenance twin modeling method based on hierarchical entity deconstruction and incremental data synchronization, according to an exemplary embodiment.

[0023] Figure 5 This is a structural block diagram of an operation and maintenance twin modeling system based on hierarchical entity deconstruction and incremental data synchronization, according to an exemplary embodiment. Detailed Implementation

[0024] The embodiments of this application are described in detail below, and examples of the embodiments are shown in the accompanying drawings.

[0025] In the description of this specification, the references to "certain embodiments," "one embodiment," "some embodiments," "illustrative embodiment," "example," "specific example," or "some examples" refer to specific features, structures, materials, or characteristics described in connection with the described embodiment or example, which are included in at least one embodiment or example of this application. In this specification, the illustrative expressions of the above terms do not necessarily refer to the same embodiment or example. Furthermore, the specific features, structures, materials, or characteristics described may be combined in any suitable manner in one or more embodiments or examples.

[0026] This application discloses an operation and maintenance twin modeling method based on hierarchical entity deconstruction and incremental data synchronization, mainly applied to scenarios such as enterprise-level information system operation and maintenance, digital business platform operation and maintenance, data center facility operation and maintenance, and park infrastructure operation and maintenance. With the continuous advancement of digital transformation, the number of objects covered by enterprise operation and maintenance scenarios, such as business systems, microservice components, data links, configuration files, and database instances, continues to grow. The data update characteristics, business value, and operation and maintenance requirements of different objects vary significantly, placing higher demands on the construction efficiency, real-time update performance, and resource adaptability of operation and maintenance twin models. Traditional operation and maintenance twin modeling generally adopts a unified modeling standard and a full data synchronization mode, which cannot differentiate configurations according to the operating characteristics and business value of different operation and maintenance objects. This easily leads to redundant consumption of computing power and bandwidth resources, and it is difficult to adaptively adjust the model configuration according to changes in the operation and maintenance scenario, failing to balance model mapping accuracy and operating efficiency.

[0027] In this embodiment of the application, an operation and maintenance twin modeling method based on hierarchical entity deconstruction and incremental data synchronization is described, referring to... Figure 1 It includes the following steps: In this embodiment of the application, an operation and maintenance twin modeling method based on hierarchical entity deconstruction and incremental data synchronization is described, referring to... Figure 1 It includes the following steps: S100 collects basic information of various entities in the operation and maintenance scenario, divides entity levels according to data update frequency and functional importance, matches the corresponding modeling accuracy and data collection range for each entity level, builds an initial operation and maintenance twin model, sorts out the business relationship between entities, and generates an entity relationship graph.

[0028] In this embodiment of the invention, reference is made to Figure 2 The specific execution process of S100 is as follows: S101 collects basic information on various entities within the operational scenario through multiple channels, including the configuration management database, service registry, call chain tracing system, and configuration management platform. The collection scope covers all digital operational objects, such as business systems, microservice components, data interfaces, configuration files, database instances, caching services, and message queue instances. The collection process employs various access methods, including standardized API interfaces, log collection probes, and configuration library synchronization interfaces, adapting different collection protocols to different data sources to ensure complete access to all entity information. For each entity, the collected content includes fields such as unique entity identifier, business domain, deployment location, historical update cycle records, business call relationships, historical fault records, type of business carried, and data volume. Different entity types correspond to different collection channels: basic information for business systems and microservice components is obtained from the configuration management database and service registry; basic information for configuration files and document-type entities is obtained from the configuration management platform and document management system; and basic information for middleware entities such as databases and caches is obtained from the middleware management platform.

[0029] The collected basic information, after data cleaning and deduplication, is stored in a unified entity information database. The processing flow consists of three steps. The first step is format validation, removing invalid data with missing fields, duplicate identifiers, or incorrect formats. The second step is entity deduplication, merging data based on unique entity identifiers and cross-completing information from multiple sources for the same entity. The third step is standardized mapping, mapping field names from different data sources to the standard field system of the entity information database to ensure consistency in subsequent processing. For entities with missing attributes, cross-completion is performed using information from related entities; those that cannot be completed are marked as entities awaiting completion and temporarily placed in the lowest level of management until the information is complete, at which point the level is reassessed. For newly added entities to the operation and maintenance scenario, the basic information collection process is automatically triggered upon integration. Within several collection cycles, the basic information collection, initial level determination, and model generation for the new entity are completed, and the new entity is incorporated into the unified entity information database for management, ensuring that the database content is consistent with the actual operation and maintenance scenario and preventing monitoring blind spots. The initial execution cycle for data collection is a periodic full-scale verification, typically performed during off-peak business hours. Incremental collection is triggered in real-time when new or changed entities are detected during daily operations.

[0030] S102, after completing the basic entity information collection, all entities are hierarchically divided according to two dimensions: data update frequency and functional importance. These two dimensions combine to form four entity hierarchies. First, entities are classified according to data update frequency. The number of data updates for each entity over a past statistical period is counted, and the average update frequency per unit time is calculated. A preset update frequency threshold is established. Entities with an average update frequency higher than this threshold are classified as high-frequency entities, and those with an average update frequency lower than this threshold are classified as low-frequency entities. The update frequency threshold can be configured according to the overall data characteristics of the operation and maintenance scenario. It should be appropriately increased for business scenarios with rapid overall data changes and appropriately decreased for scenarios with relatively stable data changes. For newly integrated entities, when there is insufficient historical statistical data, an initial classification is made by referring to the average update frequency of similar entities, and then revised after sufficient operational data is accumulated.

[0031] Secondly, entities are categorized according to functional importance. Based on indicators such as the type of business an entity carries, the scope of its failure impact, business access volume, and core business dependency, entities carrying core business processes, whose failures would cause widespread business interruption, have a high percentage of access volume, and are depended on by multiple core businesses are classified as core business-related entities. Entities carrying auxiliary support businesses, whose failures only cause localized impact, have a low percentage of access volume, and are depended on only by non-core businesses are classified as auxiliary business-related entities. The determination of functional importance is combined with the enterprise's business architecture classification standards. Core transaction systems, core scheduling services, and main data transmission services in the production domain are classified as core business-related entities, while office support systems, log archiving services, test environment services, and non-core reporting services are classified as auxiliary business-related entities. After completing the initial hierarchical classification, a cross-validation is performed, reviewing the entity's historical failure impact scope and business SLA level. Entities whose classification results do not match their actual business importance are manually marked and adjusted to ensure the accuracy of the hierarchical classification and avoid core entities being mistakenly classified into auxiliary levels.

[0032] The classification results from the two dimensions are combined to form four entity hierarchies: high-frequency core business related entities, low-frequency core business related entities, high-frequency auxiliary business related entities, and low-frequency auxiliary business related entities. After hierarchical division, differentiated modeling granularity and data collection frequency are set for each entity hierarchy. The modeling granularity and data collection frequency for core business related entities are higher than those for auxiliary business related entities, and the update transmission priority for high-frequency entities is higher than that for low-frequency entities. In terms of modeling granularity, the model of high-frequency core business related entities includes multiple subdivided fields such as entity running status, resource consumption, interface calls, business indicators, and configuration parameters, with each field corresponding to an independent mapping update node; low-frequency core business related entities retain three types of fields: running status, core business indicators, and key configuration parameters; and auxiliary business related entities retain only two basic fields: running status and core indicators. The differentiation in modeling granularity is achieved through the differentiated control of the number of fields. In terms of data collection frequency, high-frequency core business related entities are set with a higher collection frequency, low-frequency core business related entities with a medium collection frequency, and auxiliary business related entities with a lower collection frequency. Each cycle automatically calibrates the collection frequency based on the actual update status of the entity in the previous stage. For entities whose actual update frequency is much higher than the current collection frequency, the collection frequency is automatically increased to avoid data sampling omissions; for entities whose actual update frequency is much lower than the current collection frequency, the collection frequency is automatically decreased to reduce invalid collection overhead.

[0033] S103. After completing the entity hierarchy, the business relationships between entities are analyzed to generate an entity relationship graph. First, the business dependency chains and data interaction flows between entities within the operation and maintenance scenario are analyzed. Through business call chain data, interface call logs, service registry data, configuration dependencies, and other channels, the call and called relationships between entities are identified, clarifying the upstream and downstream interaction directions for each group of interacting entities. For entities with bidirectional interactions, the interaction relationships in both directions are marked, corresponding to different upstream and downstream positioning. For indirectly dependent entities, the complete dependency path is clarified by tracing the call chain level by level.

[0034] Secondly, the total amount of data interaction and the number of business calls between entities within a unit period are statistically analyzed. The association strength between entities is calculated by combining the business dependency ratio and the data interaction ratio. Association strength is represented by a quantitative score; a higher score indicates a stronger association. This quantitative score is calculated using a weighted summation method: Pre-set the weight ratios of business dependencies and data interaction items, with the sum of these weight ratios being 100%; separately assess the business dependency score and the data interaction ratio score between two entities; multiply the business dependency score by the weight ratio of the business dependency item, and multiply the data interaction ratio score by the weight ratio of the data interaction item, then add the two results to obtain the final quantitative score of association strength between entities.

[0035] The business dependency ratio is determined based on the degree of business dependence between entities. If the core operation of one entity completely depends on the output of another entity, and the lack of the corresponding output will prevent normal service provision, then the business dependency ratio is high, and its corresponding weight is large. If there are only occasional, non-essential interactions between entities, and the lack of interaction does not affect the operation of the entity's core functions, then the business dependency ratio is low, and its corresponding weight is small. The data interaction ratio is determined based on the proportion of data interaction volume between entities to their respective total interaction volume. The higher the interaction volume proportion, the higher the corresponding data interaction ratio. The business dependency ratio and the data interaction ratio are weighted according to their corresponding weights to obtain the association strength value between entities. Under normal circumstances, the association strength is updated periodically. During daily operation, when a significant fluctuation in the interaction data between entities is detected, the strength of a single association edge is triggered to ensure that the association strength reflects the latest business interaction situation.

[0036] Finally, each group of entities is matched with a corresponding association strength tag according to the upstream and downstream interaction direction, forming an entity association graph with association strength tags. The graph is stored in a graph database. Each node stores basic attributes such as unique entity identifier, entity name, hierarchy, and business domain. Each edge stores attributes such as upstream entity identifier, downstream entity identifier, association strength, interaction direction, and statistical period, supporting fast upstream and downstream traversal based on nodes. When querying associated entities, a breadth-first traversal is used. Starting from the target entity, associated entities are traversed layer by layer along the upstream and downstream directions. The association strength threshold is checked at each layer. Branches below the threshold are stopped from traversing, and the traversal depth is recorded. Traversal stops when the maximum propagation depth is reached, ensuring query efficiency and avoiding meaningless depth traversal that consumes computing power. For newly added or removed entities, corresponding nodes are added or removed in the graph simultaneously, and the association edge information is updated.

[0037] Through the above entity layering and association graph construction process, differentiated modeling and data collection configurations can be matched according to the actual operating characteristics and business value of entities in the initial modeling stage. This breaks the problem of uneven resource allocation in the traditional unified modeling model, concentrates limited computing power and bandwidth resources on core business entities, and at the same time, fully sorts out the business relationships between entities, providing complete basic entity information support for subsequent synchronous updates and adaptive adjustments.

[0038] S200 collects operational data from each entity in real time, compares the operational data with the corresponding content in the operation and maintenance twin model, extracts the parts with differences as update data, and completes the synchronous update of the operation and maintenance twin model based on the updated data and the transmission method corresponding to the level to which the entity belongs.

[0039] In this embodiment of the invention, reference is made to Figure 3 The specific execution process of S200 is as follows: S201, during the operation and maintenance twin model, collects operational data of each entity in real time according to the data collection frequency corresponding to the level to which each entity belongs. The collected operational data includes entity operational status parameters, configuration information, business indicator data, interface call data, resource usage data, etc. The specific collected fields correspond to the modeling granularity; the finer the modeling granularity, the more comprehensive the collected fields. During the collection process, differentiated collection scheduling methods are adopted for entities at different levels. For high-frequency core business-related entities, a continuous collection mode is used, continuously acquiring entity operational data at the set maximum collection frequency; for low-frequency core business-related entities, a timed collection mode is used, periodically acquiring entity operational data at the set medium collection frequency; for auxiliary business-related entities, an interval collection mode is used, acquiring entity operational data at a lower collection frequency. For some non-critical auxiliary entities, a sampling collection method is used, performing a valid data read once every few collections to further reduce collection overhead.

[0040] The collected operational data is standardized and temporarily stored in the data cache, awaiting subsequent difference comparison and synchronization. Standardization includes uniform data field naming, data format, timestamp precision, and units of measurement to ensure accurate comparison between the collected data and the corresponding content in the twin model. The data cache employs tiered management, setting different cache durations and capacities for entities at different levels. High-frequency core business entities have longer cache durations and capacities sufficient for storing all data for extended periods, supporting historical data backtracking and comparison. Low-frequency core business entities have moderate cache durations, storing only key node data. Auxiliary business entities have longer cache durations, storing only the results of each comparison. The cache uses a first-in, first-out (FIFO) eviction mechanism, automatically cleaning up data exceeding its expiration time to free up cache space.

[0041] For collected abnormal data, such as data with values ​​exceeding reasonable ranges, incorrect formats, or abnormal timestamps, a secondary verification is performed before direct rejection. Data from multiple collection periods before and after the entity is retrieved for cross-validation. If data from multiple consecutive periods falls within the abnormal range, it is determined to be a genuine anomaly and included in subsequent comparisons; if it is a single-point jump, it is determined to be collection noise and is rejected to avoid collection errors affecting the difference detection results and causing unnecessary model updates. The anomaly data determination adopts statistical distribution criteria. Data points exceeding the normal fluctuation range are marked, and the corresponding time points are padded by interpolation between two consecutive normal data points to ensure the continuity of the data sequence.

[0042] S202: After completing the collection of operational data, compare the operational data with the corresponding content in the operation and maintenance twin model, and extract the parts with differences as update data. Different difference detection methods are used for entities at different levels to balance detection accuracy and computational efficiency.

[0043] For high-frequency core business-related entities, a continuous sliding window approach is used to compare operational data with the corresponding content in the model segment by segment. A fixed-length sliding window is set, with a step size equal to a single acquisition cycle. The latest acquired operational data is segmented according to the window length, and each segment is compared with the corresponding time period content in the twin model. The fluctuation range of each indicator within the data segment is calculated, and data segments exceeding the preset fluctuation range are identified as discrepancies. When a single sliding window detects data exceeding the fluctuation range, it is not immediately identified as a discrepancy; instead, it continues to track the data in the next window. If multiple consecutive windows exceed the fluctuation range, it is formally identified as a discrepancy, triggering a synchronous update. This avoids unnecessary model updates due to instantaneous data fluctuations and reduces ineffective synchronization overhead. The sliding window continuously slides forward with data acquisition, enabling continuous comparison of operational data and timely capture of sudden state changes in high-frequency entities. The fluctuation range is adjusted according to the entity's operational characteristics. For entities with stable operational states, a narrower fluctuation range is set to ensure the identification of subtle changes; for entities with significant operational fluctuations, a wider fluctuation range is set to avoid misjudging normal fluctuations as discrepancies.

[0044] For low-frequency core business-related entities, a content-segmented hash comparison method is used to identify version changes. The model content corresponding to the entity is divided into multiple data segments of a fixed size, and a corresponding hash value is calculated for each data segment. The segment size is adaptively adjusted for entities of different sizes: smaller entities use smaller segment granularity, medium-sized entities use medium segment granularity, and larger entities use larger segment granularity, achieving a balance between difference location accuracy and computational efficiency. Each time new runtime data is collected, the hash value of each segment is calculated according to the same segmentation rules. The two sets of hash values ​​are compared one by one; segments with inconsistent hash values ​​are the changed content and are identified as the difference. Using segmented hash comparison eliminates the need to compare all content; only the hash value is needed to quickly locate the changed segment, significantly reducing the computational load. This method is suitable for static entities with low update frequency and large data volume.

[0045] For auxiliary business entities, a periodic fixed-point comparison method is used to identify data changes. Fixed comparison points are preset, and the collected runtime data is compared with the corresponding model content only at these preset comparison points to extract the corresponding differences. No difference detection is performed outside of these comparison points. The comparison period is configured according to the importance of the auxiliary entity; shorter comparison periods are set for more important auxiliary entities, and longer comparison periods are set for less important auxiliary entities. The comparison points for auxiliary business entities are staggered to avoid all auxiliary entities performing comparison operations at the same time, which would cause fluctuations in peak computing power. Staggered scheduling allocates different comparison points according to the characteristics of the entity identifier, evenly distributing the comparison operations throughout the comparison period and smoothing the computing power load.

[0046] S203: After extracting the updated data of the differences, the operation and maintenance twin model is synchronously updated based on the updated data and the transmission method corresponding to the level to which the entity belongs. Different data processing methods and transmission channels are matched for entities at different levels, balancing update real-time performance and transmission efficiency.

[0047] For frequently accessed core business entities, the extracted discrepancies are processed using lightweight encoding and then synchronized with the operational twin model via a low-latency transmission channel. Lightweight encoding employs a compact data format to compress the discrepancies, removing redundant format fields and duplicate content, significantly reducing the volume of transmitted data. The low-latency transmission channel utilizes a connection-oriented low-latency transmission protocol to ensure that updated data for core entities can be quickly transmitted to the twin model, reducing model update latency and ensuring the real-time mapping of core entities, thus meeting the high real-time requirements of core business operations.

[0048] For low-frequency core business-related entities, the extracted differential data is differentially packaged and then synchronously updated through a reliable transmission channel to maintain the operation and maintenance twin model. Differential packaging generates a differential update package based on the previous version's model content, containing only the version differences and verification information, further compressing the transmitted data volume, resulting in a significant reduction compared to full data transmission. The reliable transmission channel employs a transmission protocol with data verification and retransmission mechanisms to ensure that configuration and document data of core entities can be completely and accurately synchronized to the twin model, avoiding data loss or errors during transmission, ensuring the mapping accuracy of core static data, and maintaining a high level of transmission accuracy.

[0049] For auxiliary business entities, all discrepancies generated within the same statistical period are batch compressed and then synchronized with the operational twin model via a general transmission channel. Batch compression integrates discrepancies from multiple auxiliary entities within the same period into a single update package, using a general compression algorithm to reduce the number of transmissions and the volume of data transmitted. The general transmission channel employs a conventional transmission protocol with lower priority than the channel corresponding to the core entity. A total bandwidth threshold is set; when the total bandwidth usage is below the threshold, update data for auxiliary business entities is transmitted normally; when the total bandwidth usage exceeds the threshold, the transmission rate of auxiliary business entities is automatically reduced to prioritize the transmission bandwidth of the core business entity, ensuring that updates to the core entity are not affected by bandwidth fluctuations.

[0050] Each update data entry is accompanied by a corresponding checksum. Upon receiving the update data, the twin model first verifies its integrity using the checksum. If the verification passes, the model update is executed. If the verification fails, a retransmission request is triggered to ensure the accuracy of the updated data and prevent model mapping errors due to transmission mistakes. Each model update is executed atomically, either completing the entire update or reverting to the state before the update. This avoids inconsistencies caused by partial updates and guarantees the integrity and consistency of the twin model.

[0051] By using a layered and adaptive difference detection and transmission method, the optimal update scheme can be selected for the characteristics of entities at different levels. While ensuring the real-time and accuracy of core entity updates, it significantly reduces invalid data transmission and computational overhead. Compared with the traditional full synchronization method, it significantly reduces bandwidth and computing power consumption during the synchronization process, while ensuring the consistency of the state between the twin model and the physical entity.

[0052] S300 continuously collects statistics on synchronization delay, mapping error, and resource usage during the synchronous update process of the operation and maintenance twin model, obtains corresponding statistical results, adjusts the hierarchical level of the corresponding entity based on the statistical results, and adjusts the update method of upstream and downstream related entities along the entity association graph when the entity hierarchical level changes.

[0053] In this embodiment of the invention, reference is made to Figure 4 The specific execution process of S300 is as follows: S301 continuously collects statistics on synchronization latency, mapping error, and resource consumption data during the synchronous update process of the operation and maintenance twin model, and obtains corresponding statistical results. Synchronization latency data is obtained by recording the time of update data generation and the time the model completes the update, calculating the time difference between the two. During the statistical process, the latency value of all update operations for each entity within the statistical period is calculated, and the extreme values ​​at both ends are removed before taking the average to avoid the influence of extreme values ​​on the statistical results. The statistics are performed at a fixed period. Mapping error data is obtained by comparing the status data output by the twin model with the actual operating data of the entity, calculating the degree of deviation between the two, and calculating the average deviation rate of all core indicators for each entity. The overall error is calculated by weighting the indicators according to their importance. The higher the degree of deviation, the larger the mapping error and the lower the model accuracy. A full mapping error statistical analysis of all entities is completed at a fixed period. Resource consumption data includes bandwidth resources and computing power resources used during synchronization. The bandwidth and computing power consumption ratios of a single entity during synchronization are calculated separately, converted into a unified resource consumption score, and the resource consumption degree of a single entity is quantified. Resource consumption data is summarized at a fixed period.

[0054] Preset thresholds for mapping error, synchronization delay, and resource usage are used to normalize the statistically obtained synchronization delay, mapping error, and resource usage data, resulting in a comprehensive synchronization score for the corresponding entity. Linear normalization is used, with preset thresholds as a baseline. A standardized score corresponds to an actual value equal to the threshold; a score increases when the actual value is better than the threshold; and a score decreases when the actual value is worse than the threshold. Specifically, the standardized values ​​for mapping error, synchronization delay, and resource usage decrease as the actual error increases. The standardized values ​​of the three indicators are then weighted according to preset weighting coefficients to obtain the final comprehensive synchronization score, which is expressed as a quantitative score. A higher score indicates better synchronization performance; a lower score indicates worse synchronization performance. The weighting coefficients can be adjusted according to the emphasis of the operation and maintenance scenario. Under the default configuration, mapping error has the highest weight, followed by synchronization latency and resource consumption. If the focus is on model mapping accuracy, the weight of the mapping error indicator can be increased; if the focus is on real-time updates, the weight of the synchronization latency indicator can be increased; if the focus is on resource conservation, the weight of the resource consumption indicator can be increased. At regular intervals, the scoring thresholds and weights are calibrated based on the overall operation of the entire platform to ensure that the scoring accurately reflects the actual synchronization performance of entities and avoids distortion of the scoring system due to changes in business scale.

[0055] S302: After obtaining the comprehensive score for synchronized operation, continuously track the comprehensive score of entities over multiple statistical periods, recording entities whose scores are consistently below a preset downward adjustment threshold or consistently above a preset upward adjustment threshold, and adjust the entity's level accordingly. When an entity's comprehensive score for synchronized operation is consistently below the downward adjustment threshold for multiple consecutive statistical periods, the entity's level is increased, correspondingly increasing the entity's modeling granularity and data collection frequency, and matching it with a better transmission channel and higher transmission priority. Level increases are performed step-by-step, increasing one level at a time, and the synchronization effect is continuously tracked after adjustment. If it still does not meet the standard, it continues to increase until the highest level is reached. When an entity's comprehensive score for synchronized operation is consistently above the upward adjustment threshold for multiple consecutive statistical periods, the entity's level is decreased, correspondingly decreasing the entity's modeling granularity and data collection frequency, and matching it with a lower priority transmission channel to release excess computing power and bandwidth resources. Level decreases are also performed step-by-step, decreasing one level at a time, to avoid a sudden drop in synchronization effect due to a large decrease at once.

[0056] After an entity completes a hierarchy adjustment, it enters a cooling-off period. During this period, even if the score does not meet the requirements, no further hierarchy adjustments will be made to avoid frequent hierarchy changes due to short-term data fluctuations, reduce system instability, and ensure operational stability. At the end of each statistical cycle, all entities requiring hierarchy adjustments are aggregated, sorted by importance, and adjusted in batches, with core entities adjusted first and auxiliary entities adjusted later to avoid sudden increases in system load caused by concentrated adjustments. After the hierarchy adjustment is completed, the hierarchy markers in the entity information database are updated synchronously, and the corresponding collection, detection, and transmission configurations are adjusted to ensure that the adjusted configurations take effect immediately. The time, reason, and hierarchy information before and after the adjustment are recorded and stored in the operations and maintenance log database. After the hierarchical adjustment is completed, the comprehensive score of synchronous operation is continuously tracked for multiple statistical periods. Taking a conventional scoring calculation scheme as an example: Mapping error, synchronization delay, and resource consumption are pre-assigned corresponding weight percentages, and the sum of these three weight percentages is 100%. The actual statistical value of each indicator is compared with the pre-set corresponding threshold, and converted into a single score within the range of 0 to 100 points according to linear rules. The better the actual performance of the indicator, the higher the corresponding single score. The single scores of the three indicators are then multiplied by their respective weight percentages and summed to obtain the comprehensive score of synchronous operation for the entity. If the adjusted score improves to a reasonable range, the current level is maintained; if the adjusted score still does not meet the standard, the next level of adjustment is executed until the score meets the standard or reaches the highest / lowest level.

[0057] S303: When an entity's hierarchy changes, the update method for upstream and downstream related entities is adjusted along the entity association graph. When an entity is adjusted from an auxiliary business association level to a core business association level, the upstream and downstream related entities of that entity are traced along the entity association graph. The association strength value of each related entity is checked. For related entities with an association strength higher than a preset threshold, their modeling granularity, update transmission priority, and data collection frequency are increased simultaneously. The cascading increase is determined based on the association strength value. Related entities with high association strength are simultaneously promoted one level, while related entities with medium association strength only have their transmission priority and collection frequency increased, without adjusting their hierarchy. Related entities with an association strength lower than the threshold are not adjusted.

[0058] When an entity is reassigned from the core business association level to the auxiliary business association level, its upstream and downstream related entities are traced along the entity association graph. For related entities with association strength exceeding a preset threshold, their modeling granularity, transmission priority, and data collection frequency are simultaneously reduced. The magnitude of the cascading reduction is also determined based on the association strength value. Related entities with high association strength are simultaneously reduced by one level, while related entities with medium association strength only have their transmission priority and collection frequency reduced, avoiding excessive adjustments that could affect the overall accuracy of the model. Cascading adjustments are directionally controlled: when an entity's level is increased, the cascading adjustment simultaneously covers upstream and downstream related entities; when an entity's level is decreased, the cascading adjustment only covers directly related downstream entities, preventing upward propagation that could lead to an erroneous reduction in the core entity level and ensuring that the model accuracy of core business operations is not affected.

[0059] During cascading adjustments, a maximum propagation depth is set, meaning the impact of hierarchical changes only reaches directly related entities and entities with several levels of separation. Entities exceeding this propagation depth are excluded from the cascading adjustment scope, ensuring system stability. When multiple entities on the same association link trigger hierarchical adjustments simultaneously, the adjustment result of the highest-level entity takes precedence. Adjustments to lower-level entities follow the cascading adjustment results of higher-level entities, avoiding adjustment conflicts and ensuring the consistency of the hierarchical system. After cascading adjustments are completed, the synchronous operation effect of affected entities is continuously monitored for a period to verify the rationality of the adjustments. If over-adjustment leads to resource waste or under-adjustment leads to decreased accuracy, timely fine-tuning is performed.

[0060] Through the synchronization effect-driven hierarchical adaptive adjustment and associated entity verification adjustment, the configuration of the twin model can be continuously optimized according to the actual operation of the entity. It can automatically adapt to changes in the operation and maintenance scenario without manual intervention. At the same time, through the collaborative adjustment of associated entities, the business link model accuracy imbalance caused by changes in the hierarchical level of a single entity is avoided, and a continuous dynamic balance between model accuracy and resource consumption is achieved.

[0061] The three steps from S100 to S300, progressing sequentially and supporting each other, form a complete closed-loop operation and maintenance twin modeling system. Among them, the entity hierarchical deconstruction and relational graph construction in S100 are the foundation of the entire solution. Through two-dimensional hierarchical division and relational analysis, it provides a hierarchical basis and relational link support for subsequent differentiated incremental synchronization strategies. Without accurate hierarchical division and relational graphs, hierarchical synchronization and cascading adjustments would lack an execution benchmark. Hierarchical incremental synchronization in S200 is the core operating mechanism of the solution. Based on the preceding hierarchical configuration, it achieves differentiated data collection, difference detection, and transmission updates, reducing synchronization overhead while ensuring the accuracy of core entities. Simultaneously, the operational data generated during synchronization, such as latency, errors, and resource consumption, provides data input for the dynamic optimization of S300. The synchronization effect-driven hierarchical dynamic adjustment of S300 is a closed-loop optimization link of the solution. Based on the statistical results of the previous synchronous operation, the entity hierarchical configuration is adjusted in reverse, and cascading linkage is achieved through the association graph. The adjusted hierarchical configuration, in turn, affects the synchronization strategy of S200, while continuously correcting the initial hierarchical division results of S100. This enables the entire model system to dynamically iterate and optimize with the operation and maintenance scenario. The three work together to achieve a dynamic balance between model accuracy and resource consumption.

[0062] In this embodiment of the application, a corresponding linkage update process is set up to address the characteristic of continuous changes in business topology within the operation and maintenance scenario. The specific execution process is as follows: First, business topology change detection and affected entity identification are performed. Real-time detection of business topology changes within the operation and maintenance scenario is conducted through change notifications from channels such as the service registry, configuration management platform, and call chain tracing system. Topology change signals are captured immediately, and all entities with added, disconnected, or changed call relationships are identified, forming a set of affected entities. During change detection, the specific type and content of the change are recorded simultaneously. Common topology change scenarios include microservice instance scaling, service decommissioning, interface version upgrades, business link reconstruction, and database instance migration. The method for identifying affected entities differs depending on the scenario. For microservice instance scaling, affected entities include newly added service instances during scaling, service instances removed during scaling, and the corresponding upstream and downstream calling services. Newly added instances inherit the initial hierarchical configuration of their respective services, and removed instances are simultaneously removed from the twin model. For service decommissioning, affected entities include the decommissioned service itself and all upstream entities calling that service. Nodes of the decommissioned service are marked as invalid, simultaneously triggering updates to the association relationships of upstream entities. For API version upgrades, affected entities include the service entity to which the API belongs, as well as all upstream service entities that call the API. The attribute information and relationships corresponding to the API will be updated synchronously. For database instance migrations, affected entities include the migrated database instance, as well as all business service entities that access the database. The deployment information and associated links of the database instance will be updated synchronously.

[0063] Secondly, incremental updates are performed on the entity association graph. Based on the changed business call flow and interaction frequency, the association strength values ​​between affected entities are recalculated, and the corresponding content of the entity association graph is updated synchronously. For newly added call relationships, corresponding directed edges are added to the graph, and the association strength is calculated according to the business dependency of the new link and the initial interaction data, completing the attribute assignment of the edges. For disconnected call relationships, the corresponding directed edges are removed from the graph, and the association strength attributes are cleaned up synchronously. If the entity has no associated edges after removal, it is marked as an isolated entity. For changed call relationships, the association strength values ​​and interaction directions of the corresponding directed edges are updated, and historical association data is retained as a backup. During the graph update process, only the nodes and edges corresponding to the affected entities are adjusted, without modifying the unaffected parts. All update operations are committed transactionally in the graph database, either all updates are completed or all are rolled back, ensuring the consistency of the graph. By adopting an incremental update method, there is no need to recalculate the entire graph, which significantly reduces the computing power consumption caused by topology changes, improves the graph update efficiency, and ensures that graph synchronization can be completed in a short time after topology changes.

[0064] Finally, the affected entities' hierarchical reassessment and configuration synchronization are performed. Based on the updated association strength and entity functional importance, the affected entities' hierarchical level is reassessed, and the corresponding modeling accuracy and data collection scope are adjusted accordingly. If the topology change increases the entity's business importance (e.g., adding a core business call chain or a significant increase in the entity's business access volume), the entity's hierarchical level is increased, matching higher modeling accuracy and data collection frequency. Conversely, if the topology change decreases the entity's business importance (e.g., removing a core business call chain or a significant decrease in the entity's business volume), the entity's hierarchical level is decreased, reducing modeling fields and data collection frequency. After the hierarchical adjustment is completed, a cascading adjustment process for related entities is triggered synchronously. Based on the association strength threshold and propagation depth limit, the corresponding configurations of upstream and downstream related entities are adjusted to ensure that the overall model configuration after the topology change matches the actual business state. After adjustment, the synchronization effect is continuously tracked for multiple statistical periods to verify the rationality of the adjusted configuration. If the synchronization effect is substandard, a secondary correction is performed promptly to ensure the model's operational quality after the topology change.

[0065] In this embodiment of the application, a corresponding temporary level adjustment process is set for operation and maintenance fault alarm scenarios, and the specific execution process is as follows: First, alarm information reception and impact range positioning are performed. Alarm information from the operation and maintenance scenario is received, originating from multiple channels such as the operation and maintenance monitoring platform and alarm management system. All alarm information is uniformly accessed by the alarm processing module. After deduplication and noise reduction, the entity identifier and alarm level corresponding to the alarm are extracted. Based on the entity identifier in the alarm information, the target entity corresponding to the alarm information is located. The upstream and downstream related entities of the target entity in the entity relationship graph are retrieved to clarify the entity range affected by the alarm. Alarm levels are divided into three levels: Level 1, Level 2, and Level 3. Level 1 alarms correspond to severe faults with business interruption, have a wide impact range, and have the highest priority for handling; Level 2 alarms correspond to relatively severe faults with business performance degradation, have a limited impact range, and have a lower priority for handling; Level 3 alarms correspond to minor anomalies or warning information, do not affect normal business operation, and have the lowest priority for handling. When locating the impact range, the traversal depth is determined according to the alarm level. Severe alarms traverse multiple layers of related entities, medium alarms traverse two layers of related entities, and minor alarms only locate the target entity itself, ensuring the accuracy of impact range positioning and avoiding unnecessary expansion of the scope.

[0066] Secondly, temporary entity level upgrades and configuration adaptations are performed. Based on the alarm level, the target entity's level is temporarily upgraded, and the modeling accuracy and data collection range of upstream and downstream related entities with a correlation strength exceeding a preset threshold are adjusted simultaneously. Different alarm levels correspond to different temporary upgrade magnitudes. The most severe alarms are upgraded by two levels, even if the target entity was originally at the lowest level, it is directly upgraded to the highest level to ensure model support for fault diagnosis; moderately severe alarms are upgraded by one level, increasing the level from the original level; minor alarms do not trigger level upgrades, but only increase the data collection frequency of the entity. After the temporary level upgrade, the data collection frequency of the target entity and related entities increases, the fluctuation range of difference detection narrows, and the transmission priority is upgraded to the highest level, prioritizing the use of transmission bandwidth. This ensures that the twin model can reflect the changes in the entity's operating status in real time and accurately, providing accurate model data support for maintenance personnel to troubleshoot and locate problems. Temporary upgrade settings have a default validity period, which is calculated from the time the alarm is triggered. If the alarm is not cleared within the validity period, it will be automatically renewed until the alarm is officially cleared, thus avoiding the premature expiration of temporary configurations and affecting fault handling.

[0067] Finally, alarm cancellation and configuration restoration are performed. After the alarm information is cleared and the entity's operating status stabilizes, the target entity and its corresponding associated entities' hierarchy, modeling accuracy, and data collection range are restored to their pre-alarm state. Multiple conditions are set for determining stable status: first, the alarm information is officially cleared, and the operations platform no longer pushes corresponding alarms for that entity; second, the entity's operating parameters are within the normal range for several consecutive collection cycles, and all indicators return to normal intervals; and finally, the entity's business call volume and call success rate return to normal levels before the alarm. Only when all three conditions are met is the operating status considered stable, triggering configuration restoration to avoid premature restoration that could lead to recurring faults and affect the handling effect. The restoration operation adopts a gradual approach, first reducing the collection frequency and transmission priority, then adjusting the modeling accuracy and hierarchy, completing the restoration in two stages, with each stage separated by a statistical cycle, to avoid sudden configuration changes causing model fluctuations. After restoration is complete, the system returns to its normal operating flow, continuously optimizing the hierarchy configuration based on the synchronous operating results.

[0068] This paper illustrates the practical operation of a provincial-level digital operation and maintenance platform. This platform supports dozens of business systems and hundreds of microservice components, encompassing various types of operation and maintenance entities, including core business scheduling, data transmission, office support, log management, and testing environments. These entities exhibit significantly different data update characteristics. The core transaction scheduling system requires high real-time model mapping, while auxiliary systems such as log archiving and testing environments have lower real-time requirements. Traditional modeling solutions employ a unified full-data synchronization mode. To ensure the accuracy of core business operations, a high overall data collection and synchronization frequency is set, resulting in persistently high bandwidth and computing resources, with non-core entities consuming a large amount of redundant resources. Reducing the overall synchronization frequency significantly increases the model mapping latency for core entities, failing to meet the operational requirements for fault location and situational awareness, and ultimately failing to balance model accuracy and resource efficiency.

[0069] After implementing this solution, the platform first completes the hierarchical division and initial model building of all operational entities. Based on data update frequency and functional importance, all entities are divided into four levels. The most granular modeling and high collection frequency are configured for high-frequency core entities such as core transaction scheduling services and main data transmission services; medium granularity and medium collection frequency are configured for low-frequency core entities such as office support systems and reporting services; and simplified granularity and lower collection frequency are configured for auxiliary entities such as log archiving and test environment services. Simultaneously, differentiated difference detection and transmission methods are matched for different levels of entities. Core entities use sliding window detection and low-latency channel transmission, while auxiliary entities use periodic comparison and batch compression transmission. After deployment, the model mapping accuracy and real-time performance of core entities fully meet operational requirements. Overall bandwidth and computing power consumption are significantly reduced compared to the original full-synchronization mode. Precise resource allocation is achieved in the initial modeling stage, greatly reducing redundant resource consumption by non-core entities.

[0070] During continuous operation, the system continuously and adaptively adjusts entity levels through synchronization effect statistics and comprehensive scoring, and achieves cascading adjustments of related entities based on the entity association graph. When a core business link adds an interface or the business topology changes, the system automatically updates the entity association graph and reassesses the level of the corresponding entity, synchronously adjusting the configuration of related entities to ensure that the model accuracy of the new business link meets the standards. When a serious business alarm occurs, the system temporarily raises the level of the target entity and related entities to provide high-precision model support for fault diagnosis. After the alarm is cleared, the original configuration is gradually restored. Data statistics after a period of operation show that the average mapping error of core entities is consistently controlled at a low level, the synchronization latency is stable at the millisecond level, fully meeting the operation and maintenance requirements, while the average utilization rate of overall bandwidth and computing resources remains within a reasonable range, significantly improving resource utilization efficiency compared to traditional solutions. Through continuous adaptive iterative optimization, the model configuration is always dynamically adapted to the business operation status, achieving an optimal balance between the mapping accuracy of the twin model and the consumption of operating resources without increasing the overall resource investment.

[0071] This application discloses an operation and maintenance twin modeling system based on hierarchical entity deconstruction and incremental data synchronization, referring to... Figure 5 ,include: The hierarchical modeling module 001 collects basic information of various entities in the operation and maintenance scenario, divides the entity hierarchy according to the data update frequency and functional importance, matches the corresponding modeling accuracy and data collection range for each entity hierarchy, builds the initial operation and maintenance twin model, sorts out the business relationship between each entity, and generates an entity relationship graph. The incremental synchronization module 002 collects the running data of each entity in real time, compares the running data with the corresponding content in the operation and maintenance twin model, extracts the parts with differences as update data, and completes the synchronous update of the operation and maintenance twin model according to the transmission method corresponding to the level to which the entity belongs based on the update data. The dynamic optimization module 003 continuously collects statistics on synchronization delay, mapping error, and resource usage during the synchronous update process of the operation and maintenance twin model, obtains the corresponding statistical results, adjusts the hierarchical level of the corresponding entity based on the statistical results, and adjusts the update method of upstream and downstream related entities along the entity association graph when the entity hierarchical level changes.

[0072] This application also discloses an operation and maintenance twin modeling method based on hierarchical entity deconstruction and incremental data synchronization, including a processor, wherein the processor runs a program of the operation and maintenance twin modeling method based on hierarchical entity deconstruction and incremental data synchronization described in any one of the above embodiments.

[0073] This application also discloses a storage medium storing the program of the operation and maintenance twin modeling method based on hierarchical entity deconstruction and incremental data synchronization as described in any one of the above embodiments.

[0074] Although embodiments of this application have been shown and described above, it is understood that the above embodiments are exemplary and should not be construed as limiting this application. Those skilled in the art can make changes, modifications, substitutions and variations to the above embodiments within the scope of this application.

Claims

1. A method for operation and maintenance twin modeling based on hierarchical entity deconstruction and incremental data synchronization, characterized in that, include: Collect basic information on various entities within the operation and maintenance scenario, divide the entity into levels according to data update frequency and functional importance, match the corresponding modeling accuracy and data collection range for each entity level, build an initial operation and maintenance twin model, sort out the business relationships between entities, and generate an entity relationship graph. Real-time collection of operational data from each entity; comparison of operational data with corresponding content in the operation and maintenance twin model; extraction of discrepancies as update data; and synchronization of the operation and maintenance twin model based on the update data according to the transmission method corresponding to the level to which the entity belongs. Continuously collect statistics on synchronization delay, mapping error, and resource usage during the synchronous update process of the operation and maintenance twin model, obtain corresponding statistical results, adjust the hierarchical level of the corresponding entity based on the statistical results, and adjust the update method of upstream and downstream related entities along the entity association graph when the entity hierarchical level changes.

2. The operation and maintenance twin modeling method based on hierarchical entity deconstruction and incremental data synchronization according to claim 1, characterized in that, The steps of dividing entities into hierarchical levels based on data update frequency and functional importance, and matching corresponding modeling accuracy and data collection range to each entity level, include: Entities are divided into high-frequency entities (update frequency above a preset threshold) and low-frequency entities (update frequency below a preset threshold) based on data update frequency. Entities are also divided into core business-related entities and auxiliary business-related entities based on functional importance. The two dimensions are combined to form four entity levels. Differentiated modeling granularity and data collection frequency are set for each group of entity levels. The modeling granularity and data collection frequency for core business-related entities are higher than those for auxiliary business-related entities, and the update transmission priority for high-frequency entities is higher than that for low-frequency entities.

3. The operation and maintenance twin modeling method based on hierarchical entity deconstruction and incremental data synchronization according to claim 2, characterized in that, The step of comparing the operational data with the corresponding content in the operation and maintenance twin model and extracting the differences as update data includes: For high-frequency core business related entities, a continuous sliding window method is used to compare the running data with the corresponding content of the model segment by segment. Data segments that exceed the preset fluctuation range are judged as the difference part. For low-frequency core business related entities, a content segment hash comparison method is used to identify version changes, and the segmented content that has changed is determined to be the difference part; For entities related to auxiliary business, a periodic fixed-point comparison method is used to identify data changes, and the corresponding differences are extracted only at preset comparison points.

4. The operation and maintenance twin modeling method based on hierarchical entity deconstruction and incremental data synchronization according to claim 2, characterized in that, The steps of adjusting the hierarchical level of the corresponding entity based on statistical results, and adjusting the update method of upstream and downstream related entities along the entity relationship graph when the entity level changes, include: When an entity is adjusted from the auxiliary business association level to the core business association level, the upstream and downstream related entities are traced along the entity association graph. For related entities with association strength higher than the preset threshold, their modeling granularity is improved, transmission priority is updated, and data collection frequency is increased simultaneously. When an entity is adjusted from the core business association level to the auxiliary business association level, upstream and downstream related entities are traced along the entity association graph. For related entities with association strength higher than the preset threshold, their modeling granularity is reduced, transmission priority and data collection frequency are updated simultaneously.

5. The operation and maintenance twin modeling method based on hierarchical entity deconstruction and incremental data synchronization according to claim 1, characterized in that, The steps involved in identifying the business relationships between entities and generating an entity relationship graph include: Analyze the business dependency links and data interaction flows between various entities in the operation and maintenance scenario, and clarify the upstream and downstream interaction directions corresponding to each group of interacting entities; The total amount of data interaction and the number of business calls between entities within a statistical unit period are combined with the proportion of business dependencies and the proportion of data interaction to calculate the correlation strength between entities. For each group of entities, a corresponding association strength label is matched according to the upstream and downstream interaction direction to form an entity association map with association strength labels.

6. The operation and maintenance twin modeling method based on hierarchical entity deconstruction and incremental data synchronization according to claim 1, characterized in that, The steps for adjusting the hierarchical level of the corresponding entity based on statistical results include: Preset mapping error threshold, synchronization delay threshold and resource consumption threshold, normalize the statistically obtained synchronization delay, mapping error and resource consumption data to obtain the comprehensive score of the corresponding entity's synchronization operation; The system continuously tracks the synchronous operation comprehensive score of entities over multiple statistical periods, and records entities whose scores are consistently lower than the preset downward adjustment threshold or consistently higher than the preset upward adjustment threshold. Entities whose scores consistently fall below the downgrade threshold will be upgraded to a higher level, while entities whose scores consistently rise above the upgrade threshold will be downgraded to a lower level.

7. The operation and maintenance twin modeling method based on hierarchical entity deconstruction and incremental data synchronization according to claim 2, characterized in that, The steps for synchronizing and updating the operation and maintenance twin model based on the updated data and the transmission method corresponding to the level to which the entity belongs include: For high-frequency core business related entities, the extracted differential data is processed by lightweight encoding and then the synchronous update of the operation and maintenance twin model is completed through a low-latency transmission channel. For low-frequency core business related entities, the extracted differential data is processed by differential packaging and then the synchronous update of the operation and maintenance twin model is completed through a reliable transmission channel. For auxiliary business-related entities, after batch compression of all differential data generated within the same statistical period, the operation and maintenance twin model is synchronously updated through a general transmission channel.

8. The operation and maintenance twin modeling method based on hierarchical entity deconstruction and incremental data synchronization according to claim 1, characterized in that, Also includes: Real-time detection of business topology changes within the operation and maintenance scenario; identification of all entities whose call relationships have been added, disconnected, or changed; and formation of a set of affected entities. Based on the changed business call flow and interaction frequency, the association strength values ​​between affected entities are recalculated, and the corresponding content of the entity association graph is updated synchronously. By combining the updated association strength and the importance of entity functions, the hierarchical level of the affected entities is re-determined, and the modeling accuracy and data collection range of the corresponding entities are adjusted accordingly.

9. The operation and maintenance twin modeling method based on hierarchical entity deconstruction and incremental data synchronization according to claim 1, characterized in that, Also includes: Receive alarm information in the operation and maintenance scenario, locate the target entity corresponding to the alarm information, and retrieve the upstream and downstream related entities of the target entity in the entity association graph; Temporarily upgrade the target entity's hierarchy and simultaneously adjust the modeling accuracy and data collection range of upstream and downstream related entities whose association strength is higher than a preset threshold; Once the alarm is cleared and the entity's operating status stabilizes, the target entity and its associated entities' hierarchical level, modeling accuracy, and data collection range will be restored to their pre-alarm state.

10. An operation and maintenance twin modeling system based on hierarchical entity deconstruction and incremental data synchronization, characterized in that, include: The hierarchical modeling module collects basic information of various entities in the operation and maintenance scenario, divides the entity into levels according to the data update frequency and functional importance, matches the corresponding modeling accuracy and data collection range for each entity level, builds an initial operation and maintenance twin model, sorts out the business relationship between each entity, and generates an entity relationship graph. The incremental synchronization module collects the operational data of each entity in real time, compares the operational data with the corresponding content in the operation and maintenance twin model, extracts the parts with differences as update data, and completes the synchronous update of the operation and maintenance twin model according to the transmission method corresponding to the level to which the entity belongs based on the update data. The dynamic optimization module continuously collects statistics on synchronization delay, mapping error, and resource usage during the synchronous update process of the operation and maintenance twin model, obtains corresponding statistical results, adjusts the hierarchical level of the corresponding entity based on the statistical results, and adjusts the update method of upstream and downstream related entities along the entity association graph when the entity hierarchical level changes.

Images