Distributed database management method and system based on artificial intelligence

Abstract

The present invention is applicable to the field of database management technology, and provides a distributed database management method and system based on artificial intelligence. The system includes: a database content analysis module, a storage layout adjustment module, an anomaly detection module, a drone storage communication module, an access request guidance module, a drone node recovery module, and a solution using drones as storage and communication nodes. It brings significant innovation and multiple advantages to the distributed database management method. First, the mobility of drones enables them to quickly take over data storage tasks when the network is interrupted or the node fails, ensuring the integrity and availability of the data, while this response capability is relatively limited in traditional architectures. In complex scenarios such as disaster recovery or deployment in remote areas, drones can break through network limitations through offline transmission, effectively ensuring data security and storage continuity.

Description

Technical Field

[0001] The present invention belongs to the technical field of database management, and in particular to a distributed database management method and system based on artificial intelligence. Background Art

[0002] Database management technology mainly studies how to organize, store, manage and access data efficiently and securely to meet the growing data processing needs in the modern information society. This field covers a variety of application scenarios such as distributed databases, cloud databases, and big data storage and processing, and is committed to solving core issues such as high-performance data access, reliable storage, load balancing, disaster recovery, and data security. With the rapid development of artificial intelligence, the Internet of Things, and cloud computing, database management technology has gradually evolved in the direction of intelligence and automation. By introducing technologies such as predictive analysis, dynamic resource scheduling, and intelligent optimization, the flexibility and adaptability of the system are improved to cope with complex and changing business needs and data structures. In addition, the field of database management continues to explore how to ensure data consistency, integrity, and high availability under challenges such as heterogeneous environments, network interruptions, and high concurrent access, providing strong technical support for various critical applications in modern society.

[0003] In the existing database management process, when a node is abnormal due to hardware failure, network interruption or resource overload, the traditional architecture is prone to data unavailability, affecting the stability of the system. At the same time, in disaster recovery, remote area deployment or special scenarios (such as military missions), traditional distributed databases that rely on fixed networks are difficult to ensure effective storage and transmission of data. In addition, in the face of sudden data access peaks, the traditional system's insufficient response capability is prone to performance degradation or access delays. The lack of flexible optimization allocation strategies for frequently accessed hot data will further reduce access efficiency and increase system load. Not only that, in a distributed environment, when different nodes process related data, cross-group data transmission often occurs, resulting in communication delays. At the same time, traditional methods are insufficient in data transmission and storage security, especially sensitive data is vulnerable to network attacks and hijacking. The limitation of lack of flexibility in resource allocation makes it difficult for traditional architectures to dynamically adapt to changes in demand, resulting in low resource utilization. These problems significantly restrict the performance and applicable scenarios of traditional distributed databases. Summary of the invention

[0004] The purpose of the present invention is to provide a distributed database management method based on artificial intelligence, aiming to solve the technical problems existing in the prior art identified in the background technology.

[0005] The present invention is implemented in this way: a distributed database management method based on artificial intelligence, the method comprising:

[0006] Read and analyze the database content, divide the database content into trust levels based on data importance, sensitivity and access frequency, analyze the logical associations between the data in the database, identify naturally aggregated data clusters with natural aggregation characteristics, and set data allocation rules to distribute the database content;

[0007] Evaluate the status of all existing distributed nodes in real time, and dynamically adjust the storage layout of data in distributed nodes based on the evaluation results;

[0008] Establish a stable analysis model, obtain the current network status and node status data, and analyze the current network status and node status based on the stable analysis model to determine whether there are any abnormalities;

[0009] Equip the drone with a storage communication module, and when any abnormality in the network status or node status is identified, distribute the data in the corresponding abnormal distributed node to the drone according to the storage result of the dynamic storage strategy, and group the drones;

[0010] Optimize the data storage mode according to the access request to guide the access request to the matching node, and adjust the number of drones in the corresponding group according to the data access frequency and network conditions;

[0011] Continuously monitor the network environment and node status that have been determined to be abnormal, and recycle the drone nodes when the abnormalities disappear.

[0012] As a further solution of the present invention, the trust level classification of the database content and the analysis of the logical associations between the data in the database to identify the natural aggregation data clusters with natural aggregation characteristics specifically include:

[0013] Read the database content and pre-process the read content;

[0014] Calculate the business relevance of the data to assess the importance of each data item, identify sensitive data based on information detection, and analyze historical access logs to calculate the access frequency of each data item or data shard;

[0015] Set up phase intervals respectively, including: importance phase interval, sensitivity phase interval and access frequency phase interval, and assign matching intervals to database contents;

[0016] Define the trust level standard and calculate the trust level of each data according to the stage interval of the data;

[0017] The data in the database is mapped into a graph structure, where nodes represent data items and edges represent logical associations between data. The relationship between data is analyzed based on the graph structure, and closely interacting subclusters are identified, which are defined as naturally aggregated data clusters.

[0018] As a further solution of the present invention, the setting of data distribution rules to perform distributed storage of database content specifically includes:

[0019] Uniformly divide the identified natural aggregate data clusters and trust levels, and package the data into several data shards;

[0020] Assign trust level weight and natural aggregation weight to each data shard. The higher the trust level and the higher the natural aggregation, the greater the weight value.

[0021] Set a weight threshold. For data shards where both the trust level weight and the natural aggregation weight are higher than the weight threshold, select nodes with high security and the same geographical location for storage.

[0022] For data with trust level weights higher than the weight threshold, high-security nodes are selected for storage;

[0023] For data whose natural aggregation weight is higher than the weight threshold, nodes with the same geographical location are selected for storage, and a backup storage node is set for each storage node, and the backup storage node is not the same node as the original storage node.

[0024] As a further solution of the present invention, the real-time evaluation of the status of all existing distributed nodes and the dynamic adjustment of the storage layout of data in the distributed nodes based on the evaluation results specifically include:

[0025] Collect the status data of all distributed nodes, and calculate the load status of each node based on the current resource usage of the node. At the same time, measure the network delay, bandwidth and packet loss rate between nodes to evaluate the communication quality between nodes.

[0026] When any node is inaccessible, the backup node of the storage node is retrieved according to the data allocation rule, and a communication connection is established with the original storage node;

[0027] When the performance of any node cannot meet the load requirement, the load status of other nodes under the same data allocation rule as the node is identified, and any other node that meets the migration condition is obtained;

[0028] The migration condition is: full load status of other nodes - load status of other nodes > load demand of the node;

[0029] Migrate data shards of nodes whose performance cannot meet the load requirements to other nodes that meet the migration conditions and are under the same allocation rules.

[0030] As a further solution of the present invention, the establishment of a stable analysis model, obtaining current network status and node status data, and analyzing the current network status and node status based on the stable analysis model to determine whether there is an abnormality specifically includes:

[0031] Collect the current network status and node status data in real time, establish a stable analysis model, use the stable analysis model to analyze the network status and node status data, and compare the difference between the real-time data and the predicted data;

[0032] The predicted data is calculated by the following formula:

[0033] ;

[0034] in, is the predicted current time data value, that is, the predicted data, is a constant term, representing the long-term average, is the natural regression coefficient, which is used to weigh the impact of historical data points on the current forecast. For time The data value at It is the moving average coefficient multiplied by the previous error term, which is used to adjust the predicted value and is the error term;

[0035] A difference threshold is set, and the content where the difference between the predicted data and the real-time data is greater than the difference threshold is defined as abnormal state data, and the nodes related to the data are marked.

[0036] As a further solution of the present invention, the drone is equipped with a storage communication module, and when any abnormality in the network state or node state is identified, the data in the corresponding abnormal distributed node is allocated to the drone according to the storage result of the dynamic storage strategy, and the drone is grouped, which specifically includes:

[0037] Based on the analysis results of the stable analysis model, determine whether there are marked nodes, and identify all marked nodes;

[0038] Analyze the content of the marked nodes, and allocate a corresponding number of drone groups to each marked node according to the storage content size, and synchronously transfer the content in the node to the corresponding drone group;

[0039] Based on the access frequency and access history of the data, the same grouping tags are added to the data shards that are accessed simultaneously, and the drone groups are regrouped according to the grouping tags.

[0040] As a further solution of the present invention, the method of optimizing the data storage mode according to the access request guides the access request to the matching node, and adjusts the number of drones in the corresponding group according to the data access frequency and network conditions, specifically includes:

[0041] Collect and analyze database access logs to determine the frequency, patterns, and trends of data access and identify hot spots of data access;

[0042] Establish an access prediction model to predict future access requests, and generate a dynamic optimization strategy based on the output of the access prediction model and the rule of balancing the distributed node load;

[0043] The prediction of future access requests is specifically as follows:

[0044] ;

[0045] in, is the predicted access request volume at the next moment, For in time The actual number of access requests at the moment, For time The weight of the moment, is the size of the historical time window;

[0046] Based on dynamic optimization strategies, the number and grouping of drones are dynamically adjusted.

[0047] Another object of the present invention is to provide a distributed database management system based on artificial intelligence, the system comprising:

[0048] The database content analysis module is used to read and analyze the database content, divide the database content into trust levels based on data importance, sensitivity and access frequency, analyze the logical associations between the data in the database, identify natural aggregation data clusters with natural aggregation characteristics, and set data allocation rules to distribute the database content;

[0049] The storage layout adjustment module is used to evaluate the status of all existing distributed nodes in real time and dynamically adjust the storage layout of data in the distributed nodes based on the evaluation results;

[0050] The anomaly detection module is used to establish a stable analysis model, obtain the current network status and node status data, and analyze the current network status and node status based on the stable analysis model to determine whether there is an anomaly;

[0051] The UAV storage communication module is used to equip the UAV with a storage communication module, and when any abnormality in the network state or node state is identified, the data in the corresponding abnormal distributed node is allocated to the UAV according to the storage result of the dynamic storage strategy, and the UAV is grouped;

[0052] An access request guiding module is used to guide access requests to matching nodes according to the access request optimization data storage mode, and adjust the number of drones in the corresponding group according to the data access frequency and network conditions;

[0053] The drone node recovery module is used to continuously monitor the network environment and node status that have been determined to be abnormal, and to recover the drone nodes when the abnormality disappears.

[0054] The beneficial effects of the present invention are:

[0055] The use of drones as storage and communication nodes brings significant innovation and multiple advantages to distributed database management methods. First, the mobility of drones enables them to quickly take over data storage tasks when the network is interrupted or the node fails, ensuring data integrity and availability, which is more limited in traditional architectures. In complex scenarios such as disaster recovery or remote area deployment, drones can break through network limitations through offline transmission, effectively ensuring data security and storage continuity.

[0056] Secondly, by dynamically adjusting the number and grouping of drones and optimizing resource allocation according to data access hotspots and access frequency, drones can significantly improve data access efficiency. For example, for peak access areas, the system can deploy more drones to carry relevant data, thereby reducing the load pressure on fixed nodes and shortening data transmission delays. At the same time, combined with the data access prediction model, drones can be deployed in advance to quickly respond to sudden demands, giving full play to their flexibility and efficiency.

[0057] In addition, the physical isolation of drones further enhances the security of data transmission, especially the protection of sensitive data. Through encrypted storage and dynamic grouping, drones can reduce security risks and improve the reliability of distributed databases in complex environments. Therefore, this solution not only improves the system's disaster recovery capability and access efficiency, but also broadens the applicable scenarios of distributed databases, which has extremely high practical value and forward-looking significance. BRIEF DESCRIPTION OF THE DRAWINGS

[0058] Figure 1 A flowchart of a distributed database management method based on artificial intelligence provided by an embodiment of the present invention;

[0059] Figure 2 A flowchart of reading and analyzing database content provided by an embodiment of the present invention;

[0060] Figure 3 A flowchart for setting data distribution rules and distributing the database content;

[0061] Figure 4 A flowchart of dynamically adjusting the storage layout of data in distributed nodes provided by an embodiment of the present invention;

[0062] Figure 5 A flowchart for analyzing the current network status and node status based on a stable analysis model to determine whether there is an abnormality provided by an embodiment of the present invention;

[0063] Figure 6 A flowchart of allocating data in corresponding abnormal distributed nodes to drones and grouping the drones provided in an embodiment of the present invention;

[0064] Figure 7 A flowchart for adjusting the number of drones in a corresponding group according to data access frequency and network conditions provided by an embodiment of the present invention;

[0065] Figure 8 A structural block diagram of an artificial intelligence-based distributed database management system provided in an embodiment of the present invention. DETAILED DESCRIPTION

[0066] In order to make the purpose, technical solution and advantages of the present invention more clearly understood, the present invention is further described in detail below in conjunction with the accompanying drawings and embodiments. It should be understood that the specific embodiments described herein are only used to explain the present invention and are not intended to limit the present invention.

[0067] It is understood that the terms "first", "second", etc. used in this application may be used herein to describe various elements, but unless otherwise specified, these elements are not limited by these terms. These terms are only used to distinguish a first element from another element. For example, without departing from the scope of this application, a first xx script may be referred to as a second xx script, and similarly, a second xx script may be referred to as a first xx script.

[0068] Figure 1 A flowchart of a distributed database management method based on artificial intelligence provided by an embodiment of the present invention, such as Figure 1 As shown, the method includes:

[0069] S100, read and analyze the database content, divide the database content into trust levels based on data importance, sensitivity and access frequency, analyze the logical associations between the data in the database, identify natural aggregation data clusters with natural aggregation characteristics, set data allocation rules, and perform distributed storage of the database content;

[0070] This step first reads all the content in the database and preprocesses it. The preprocessing process includes data cleaning, format standardization, and data integrity verification to ensure the accuracy and consistency of subsequent analysis. This step is crucial to improving data quality and analysis accuracy. The system then evaluates the importance of each data item by calculating the business relevance of each data item and combining the importance of the data in practical applications. For example, through in-depth analysis of the business model, key data points are extracted and given higher weights. At the same time, machine learning algorithms automatically detect and identify sensitive data that may be contained in the database, such as user privacy information or commercial confidential data. This sensitivity detection not only relies on predefined rules and keyword matching, but also combines natural language processing technology and the intelligent recognition capabilities of deep learning models to improve detection efficiency.

[0071] In addition, the system will combine historical access logs, use statistical analysis and frequent pattern mining technology to calculate the access frequency of each data item or data segment to quantify the activity of the data in actual use. Based on the analysis results, the system divides the database content into different intervals, covering importance stage intervals, sensitivity stage intervals, and access frequency stage intervals, and assigns matching interval labels to these data. By setting trust level standards, the system conducts a graded trust level assessment for each piece of data based on these interval labels. This process comprehensively considers the importance, sensitivity, and access frequency of the data, making the division of trust levels more scientific and reasonable.

[0072] After completing the trust level division, the system uses a graph structure to map the data in the database into a network, where nodes represent data items and edges represent logical associations between data. Through graph structure analysis, the system can accurately identify the relationship patterns between data, especially subclusters that interact closely. For these subclusters, the system further defines them as naturally aggregated data clusters, which is a clustering method based on the inherent association of data. For example, some highly related data items may be frequently called simultaneously in actual applications, and storing them in an aggregated manner can help reduce access latency and improve data management efficiency.

[0073] This step can ensure that the distributed storage of data is carried out according to the priority and security level through the division of trust levels, which not only improves the access efficiency of data, but also enhances the security of data. The introduction of graph structure analysis further optimizes data management, allowing closely interacting data to be stored centrally, thereby reducing network overhead and latency during distributed access. In addition, this process lays a solid foundation for subsequent steps, ensuring the rationality and flexibility of data distribution, and providing a guarantee for the efficient operation of the entire distributed database management system.

[0074] like Figure 2As shown, the trust level classification of the database content and the analysis of the logical associations between the data in the database to identify natural aggregation data clusters with natural aggregation characteristics specifically include:

[0075] S110, reading database content and preprocessing the read content;

[0076] S120, calculating the business relevance of the data, evaluating the importance of each data item accordingly, identifying sensitive data based on information detection, and analyzing historical access logs to calculate the access frequency of each data item or data shard;

[0077] S130, respectively setting stage intervals, including: importance stage intervals, sensitivity stage intervals, and access frequency stage intervals, and assigning matching intervals to database content;

[0078] S140, defining a trust level standard, and calculating the trust level of each data according to the stage interval in which the data is located;

[0079] S150, mapping the data in the database into a graph structure, where nodes represent data items and edges represent logical associations between data, and analyzing the relationships between data based on the graph structure to identify closely interacting subclusters, which are defined as naturally aggregated data clusters.

[0080] like Figure 3 As shown, the setting of data distribution rules to distribute the database content specifically includes:

[0081] S160, uniformly dividing the identified natural aggregate data clusters and trust levels, and packaging the data into a number of data fragments;

[0082] S170, assigning a trust level weight and a natural aggregation weight to each data shard, wherein the higher the trust level and the higher the natural aggregation, the greater the weight value;

[0083] S180, setting a weight threshold, for data shards where both the trust level weight and the natural aggregation weight are higher than the weight threshold, selecting nodes with high security and the same geographical location for storage;

[0084] For data with trust level weights higher than the weight threshold, high-security nodes are selected for storage;

[0085] For data whose natural aggregation weight is higher than the weight threshold, nodes with the same geographical location are selected for storage, and a backup storage node is set for each storage node, and the backup storage node is not the same node as the original storage node.

[0086] The division of data shards not only considers the logical correlation between data, but also combines the importance and sensitivity of data, so that each shard is both independent and able to reflect the aggregation characteristics of related data. Each data shard is given a trust level weight and a natural aggregation weight. These two weights are based on the core indicators of system evaluation and reflect the importance of data and its inherent correlation. The size of the weight value directly determines the choice of subsequent storage strategy and is an important basis for optimizing storage efficiency and security.

[0087] In the specific storage strategy setting, the system introduces the concept of weight threshold. For data shards whose trust level weight and natural aggregation weight both exceed the weight threshold, the system gives priority to nodes with high security and the same geographical location for storage. The design of this strategy not only meets the highest demand for data security, but also reduces the delay of cross-regional data transmission through geographical aggregation, thereby improving access efficiency. For data whose trust level weight alone exceeds the weight threshold, the focus is on security, and high-security nodes are selected for storage to ensure that the data is protected from potential attacks. For data whose natural aggregation weight alone is higher than the weight threshold, the system selects nodes with the same geographical location for storage, which can maximize the performance advantages of naturally aggregated data and reduce network overhead and access time.

[0088] In addition, the system sets backup storage nodes for all storage nodes, and stipulates that the backup storage node cannot be the same as the original storage node. This design significantly improves the disaster recovery capability of data. When a node fails or is unavailable, the backup node can quickly take over the data service to ensure the continuity and stability of the system. This backup mechanism avoids the risk of traditional single point failure through the characteristics of distributed architecture, and provides guarantee for high availability of data.

[0089] S200, evaluate the status of all existing distributed nodes in real time, and dynamically adjust the storage layout of data in the distributed nodes based on the evaluation results;

[0090] This step uses the monitoring module within the system to comprehensively collect the status data of all distributed nodes. These data include but are not limited to hardware performance indicators such as the current resource utilization rate, CPU occupancy rate, memory usage, and disk space usage of the node. In addition, it is necessary to measure the network performance data between nodes, such as network latency, bandwidth, and packet loss rate, to form a comprehensive evaluation report on the communication quality between nodes. These data are integrated and passed to the analysis module, and the real-time load status of each node is calculated through a specific algorithm. The calculation formula for the load status may comprehensively consider multi-dimensional indicators such as the current resource utilization rate, bandwidth, and communication quality of the node to ensure the comprehensiveness and accuracy of the evaluation.

[0091] In actual operation, when the system detects that a node is inaccessible, the fault tolerance mechanism will be triggered immediately. According to the previously set distributed storage data allocation rules, the data backup node of the node will be quickly located, and a communication connection with the backup node will be attempted to ensure data availability. If the performance of a node cannot meet the current load requirements, the system will further evaluate the load status of other nodes under the same data allocation rules as the node, and select them according to the migration conditions. The setting of migration conditions is aimed at ensuring that the migrated node can operate stably, that is, data can only be migrated to the node if the migration target node can provide sufficient free resources. After completing the evaluation, the system will migrate the data shards stored on the node with insufficient performance, and migrate these data to other nodes that meet the migration conditions to optimize the global load distribution of the cluster.

[0092] By real-time monitoring and analyzing the status data of nodes, the system can quickly identify and locate problem nodes, thereby reducing service interruption time caused by single point failures. This capability is critical for distributed database environments, because node failures may occur at any time in highly distributed data storage. The S200 ensures high data availability and system stability by dynamically adjusting the storage layout. Secondly, in terms of load balancing, this step effectively avoids a single node becoming a performance bottleneck due to excessive load through load status evaluation and intelligent migration mechanisms. This dynamic load adjustment mechanism improves the utilization of system resources, while reducing the impact of performance fluctuations on data access, ensuring a consistent user experience.

[0093] like Figure 4 As shown, the real-time evaluation of the status of all existing distributed nodes and the dynamic adjustment of the storage layout of data in the distributed nodes based on the evaluation results specifically include:

[0094] S210, collect status data of all distributed nodes, and calculate the load status of each node based on the current resource usage of the node, and measure the network delay, bandwidth and packet loss rate between the nodes to evaluate the communication quality between the nodes;

[0095] S220, when any node is inaccessible, a backup node of the storage node is retrieved according to the data allocation rule, and a communication connection is established with the original storage node;

[0096] S230, when the performance of any node cannot meet the load requirement, identify the load status of other nodes under the same data allocation rule as the node, and obtain any other node that meets the migration condition;

[0097] The migration condition is: full load status of other nodes - load status of other nodes > load demand of the node;

[0098] S240, migrate the data shards of the nodes whose performance cannot meet the load requirements to other nodes that meet the migration conditions and are under the same allocation rule.

[0099] S300, establishing a stable analysis model, obtaining current network status and node status data, and analyzing the current network status and node status based on the stable analysis model to determine whether there is an abnormality;

[0100] This step collects status data from all distributed nodes, including the current resource usage of each node (such as CPU, memory, storage capacity, etc.), and calculates the load status of each node based on this data. The evaluation of the load status adopts a dynamic calculation method, which is closely related to the resource situation and real-time workload of the node. In addition, the system also measures the quality of network communication between nodes, including key indicators such as network latency, bandwidth utilization, and packet loss rate. The monitoring of these network parameters can fully reflect the communication status between nodes and provide data support for subsequent decision-making.

[0101] Once it is detected that a node is inaccessible, the system will immediately start the fault tolerance mechanism.

[0102] There are two fault-tolerant mechanisms for this method. One is to call the backup node, and the other is to use drones to build nodes:

[0103] The backup node is the default fault-tolerance mechanism in the system, which is used to handle situations where the node cannot operate normally or access failures occur in a short period of time. When a distributed node is temporarily unavailable due to a failure, network interruption or other short-term problems, the system will first find the backup storage node corresponding to the node according to the data allocation rules, and quickly establish a communication connection with the backup node. In this case, the main role of the backup node is to replace the original unavailable node to provide data access services, ensuring the continuity of the system and the high availability of data.

[0104] The drone storage solution is an innovative solution for the system when facing more complex or extremely abnormal conditions. The drone call is initiated only when the network status or node status is seriously abnormal, so that the backup node cannot meet the storage or access requirements. For example, when the distributed nodes cannot communicate normally or data access is seriously blocked due to network interruption, node overload or regional failure (such as large-scale power outages, natural disasters, etc.), drones are introduced into the system as dynamic storage and transmission tools.

[0105] In terms of load balancing, when it is detected that the performance of a node cannot meet its current load requirements (such as exceeding the processing capacity of the CPU or memory), the system will further identify and analyze the available resources of other nodes under the same data allocation rules as the node. To ensure the effectiveness and security of the migration process, the system introduces a calculation formula for the migration conditions: the full load status value of other nodes minus the current load status value must be greater than the load requirement value of the node to be migrated. This means that the target node will only be selected for data migration if the remaining resources are sufficient to carry it. In addition, the target node must follow the same data allocation rules as the source node to ensure logical consistency and access efficiency after data migration.

[0106] This step provides accurate status data support for the system by calculating the load status and evaluating the quality of network communication, so that the system can quickly identify potential problem nodes. Secondly, when a node is inaccessible, the fault-tolerant mechanism effectively avoids the risk of data loss and service interruption through the immediate takeover of the backup node, greatly improving the reliability of the system. In the case of insufficient node performance, the load balancing mechanism not only reasonably allocates system resources, but also ensures the consistency and security of the data migration process. Through the scientific setting of migration conditions, the system can effectively avoid the problem of target node overload or migration failure. In addition, the application of multi-threaded parallel transmission technology makes data migration more efficient and has little impact on the overall performance of the system.

[0107] like Figure 5 As shown, the establishment of a stable analysis model, obtaining the current network status and node status data, and analyzing the current network status and node status based on the stable analysis model to determine whether there is an abnormality, specifically includes:

[0108] S310, collecting current network status and node status data in real time, and establishing a stable analysis model, using the stable analysis model to analyze the network status and node status data, and comparing the difference between the real-time data and the predicted data;

[0109] The predicted data is calculated by the following formula:

[0110] ;

[0111] in, is the predicted current time data value, that is, the predicted data, is a constant term, representing the long-term average, is the natural regression coefficient, which is used to weigh the impact of historical data points on the current forecast. For time The data value at It is the moving average coefficient multiplied by the previous error term, which is used to adjust the predicted value and is the error term;

[0112] S320, setting a difference threshold, defining the content where the difference between the predicted data and the real-time data is greater than the difference threshold as abnormal state data, and marking the nodes related to the data.

[0113] S400, equipping the drone with a storage communication module, and when any abnormality in the network state or node state is identified, allocating data in the corresponding abnormal distributed node to the drone according to the storage result of the dynamic storage strategy, and grouping the drones;

[0114] For nodes marked as abnormal, the system will distribute the key data stored in the node to the drone based on the dynamic storage strategy. The drone is equipped with an efficient storage communication module in this system, which can quickly receive data from abnormal nodes and group the data according to the distribution strategy. After grouping, these drones can be divided into two categories: one is responsible for temporary data storage, and the other is responsible for transferring data to other available nodes or safe areas for storage. Drones solve the problem of data unavailability caused by abnormal conditions by temporarily supplementing storage resources.

[0115] When the model detects that some nodes are marked as abnormal (such as node disconnection, performance overload, storage device damage, etc.), the system will immediately classify these nodes as "marked nodes" and make them the focus of subsequent operations. At this time, the system will accurately identify all marked nodes and extract detailed information about their storage content, including data size, type, access frequency, and logical association with other data.

[0116] Subsequently, based on the storage content size of each abnormal node, the system will scientifically allocate a corresponding number of drone groups to each abnormal node. The allocation of drone groups is based on the matching of data volume and drone storage capacity, while also considering physical characteristics such as communication capabilities between drones and flight duration. After the allocation is completed, the drone group establishes a communication connection with the abnormal node and performs data synchronization transfer. In this process, the system will take a variety of measures to ensure the integrity and consistency of the data, including data verification, redundant backup and other technical means.

[0117] After data synchronization is completed, the system further analyzes the access frequency and access history of the data and logically shards the synchronously transferred data. Specifically, for data shards that are frequently accessed simultaneously, the system will add the same grouping tags to them. Based on these grouping tags, the drone groups will be regrouped again to ensure that drones in the same group can carry the same data shards with high access correlation. Such a grouping strategy can significantly improve the efficiency of subsequent data access and reduce communication delays or resource waste caused by cross-group access.

[0118] By equipping drones with storage and communication modules, flexibility and mobility that cannot be achieved in traditional distributed database architectures are achieved. When a node anomaly occurs, the drone can respond quickly and replace the abnormal node to take on the storage task, significantly improving the system's disaster recovery capability and data reliability. Secondly, the dynamic allocation and secondary grouping operations of drone groups based on the storage content size and data access characteristics maximize the efficiency of data storage and access, ensuring that the system can still operate efficiently under abnormal circumstances. In addition, by utilizing the mobility of drones, the system realizes flexible migration of data in space, which not only reduces the waste of resources during the repair of abnormal nodes, but also provides innovative solutions for further protection and management of data.

[0119] The main tasks of drones include extracting data from abnormal nodes and distributing the data to other available nodes according to dynamic storage strategies. At the same time, drones can be used as mobile storage devices to perform temporary data storage and access functions when the distributed network cannot provide stable services.

[0120] like Figure 6 As shown, the storage communication module is equipped for the drone, and when any abnormality in the network state or node state is identified, the data in the corresponding abnormal distributed node is allocated to the drone according to the storage result of the dynamic storage strategy, and the drone is grouped, specifically including:

[0121] S410, based on the analysis result of the stable analysis model, determining whether there are any marked nodes, and identifying all marked nodes;

[0122] S420, analyzing the content of the marked nodes, and assigning a corresponding number of drone groups to each marked node according to the size of the stored content, and synchronously transferring the content in the node to the corresponding drone group;

[0123] S430, based on the access frequency and access history of the data, the same grouping tag is added to the data fragments that are accessed simultaneously, and the drone group is regrouped according to the grouping tag.

[0124] S500, optimizing the data storage mode according to the access request to guide the access request to the matching node, and adjusting the number of drones in the corresponding group according to the data access frequency and network conditions;

[0125] This step collects comprehensive access log information from the distributed database. The log data includes the timestamp of each access, the source of access, the data object accessed, the access frequency, and other operation records (such as specific operation types such as read, write, and update). Based on these historical access data, the system uses artificial intelligence and data mining technology to deeply analyze the access frequency and pattern, such as which data is frequently accessed hot data, which time periods generate the largest number of visits, and the distribution pattern of requests from different access sources. At the same time, the geographical distribution characteristics of the distributed database are analyzed to identify the hot areas of data access and mark the regional characteristics of access requests, such as which regions of users are more inclined to access specific types of data. The main goal of this step is to clarify current and past access characteristics and provide reliable basic data for subsequent optimization.

[0126] Next, based on the analysis results, the system introduces an advanced access prediction model. Through model training and application, the system can make high-precision predictions on access requests that may be generated in the future. These prediction outputs include not only the access frequency and trend of specific data, but also the potential access peak time period and regional distribution characteristics. The results of the prediction model provide a key basis for the system to generate dynamic optimization strategies.

[0127] Based on the prediction results of the access prediction model, the system automatically generates a dynamic optimization strategy with the core goal of balancing the load of distributed nodes. Specifically, the dynamic optimization strategy includes two aspects: first, the adjustment of the data storage layout to ensure that high-frequency access data is distributed on nodes with lower loads, thereby reducing access conflicts and data blocking; second, the flexible adjustment of the number and grouping of drones, that is, dynamically increase or decrease the number of drones responsible for carrying and transmitting data according to the predicted access needs, and re-optimize the grouping structure of drones. For example, when the access demand of a data hotspot area is predicted to increase suddenly, the system will allocate more drone groups in the area, and concentrate on carrying data related to the access demand of the area, thereby improving the response speed of data access.

[0128] In actual operation, the dynamic adjustment process of drones is particularly meticulous. The adjustment of the number of drones not only takes into account the current access needs, but also comprehensively evaluates multiple factors such as network bandwidth, drone storage capacity, flight duration, and node load, so as to achieve the optimal configuration of drone resources. At the same time, the system will re-divide drone groups for different access request types. By matching highly correlated data shards with requests, these data shards are allocated to the same drone group, thereby reducing the delay problem caused by cross-group data access.

[0129] This step introduces access prediction models and dynamic optimization strategies, which greatly improves the response speed and accuracy of the system to access needs. Especially in the face of a surge in access requests or sudden access to hot spots, S500 effectively avoids data access delays and blocking problems through advance prediction and optimization adjustments, providing users with a smooth user experience. In addition, the dynamic optimization strategy avoids the overload problem of a single node by reasonably allocating the load of distributed nodes, further improving the stability and reliability of the system. At the same time, the flexible adjustment of the number and grouping of drones fully utilizes the mobility and storage flexibility of drones, greatly improving the resource utilization efficiency and emergency response capabilities of the system.

[0130] like Figure 7 As shown, the method optimizes the data storage mode according to the access request to guide the access request to the matching node, and adjusts the number of drones in the corresponding group according to the data access frequency and network conditions, specifically including:

[0131] S510, collecting and analyzing database access logs, determining the frequency, pattern, and trend of data access, and identifying hot spots of data access;

[0132] S520, establishing an access prediction model to predict future access requests, and generating a dynamic optimization strategy based on the output of the access prediction model and taking distributed node load balancing as a rule;

[0133] The prediction of future access requests is specifically as follows:

[0134] ;

[0135] in, is the predicted access request volume at the next moment, For in time The actual number of access requests at the moment, For time The weight of the moment, is the size of the historical time window;

[0136] S530, dynamically adjusts the number and grouping of drones based on dynamic optimization strategies.

[0137] S600, continuously monitor the network environment and node status that have been determined to be abnormal, and recycle the drone node when the abnormality disappears.

[0138] Figure 8 The structural block diagram of the distributed database management system based on artificial intelligence provided by the embodiment of the present invention is as follows: Figure 8 As shown, the distributed database management system based on artificial intelligence includes:

[0139] The database content analysis module 100 is used to read and analyze the database content, classify the database content into trust levels based on the importance, sensitivity and access frequency of the data, analyze the logical associations between the data in the database, identify the natural aggregation data clusters with natural aggregation characteristics, and set data allocation rules to perform distributed storage of the database content;

[0140] The storage layout adjustment module 200 is used to evaluate the status of all existing distributed nodes in real time, and dynamically adjust the storage layout of data in the distributed nodes based on the evaluation results;

[0141] The anomaly detection module 300 is used to establish a stable analysis model, obtain the current network status and node status data, and analyze the current network status and node status based on the stable analysis model to determine whether there is an anomaly;

[0142] The drone storage communication module 400 is used to equip the drone with a storage communication module, and when an abnormality is identified in either the network state or the node state, the data in the corresponding abnormal distributed node is allocated to the drone according to the storage result of the dynamic storage strategy, and the drone is grouped;

[0143] The access request guiding module 500 is used to guide the access request to the matching node according to the access request optimization data storage mode, and adjust the number of drones in the corresponding group according to the data access frequency and network conditions;

[0144] The drone node recovery module 600 is used to continuously monitor the network environment and node status that have been determined to be abnormal, and to recover the drone node when the abnormality disappears.

[0145] It should be understood that, although each step in the flow chart of each embodiment of the present invention is shown in sequence according to the indication of the arrow, these steps are not necessarily performed in sequence according to the order indicated by the arrow. Unless there is a clear explanation in this article, the execution of these steps does not have a strict order restriction, and these steps can be performed in other orders. Moreover, at least a portion of the steps in each embodiment may include a plurality of sub-steps or a plurality of stages, and these sub-steps or stages are not necessarily performed at the same time, but can be performed at different times, and the execution order of these sub-steps or stages is not necessarily performed in sequence, but can be performed in turn or alternately with at least a portion of other steps or sub-steps or stages of other steps.

[0146] Those skilled in the art can understand that all or part of the processes in the above-mentioned embodiment methods can be completed by instructing the relevant hardware through a computer program, and the program can be stored in a non-volatile computer-readable storage medium. When the program is executed, it can include the processes of the embodiments of the above-mentioned methods. Among them, any reference to memory, storage, database or other media used in the embodiments provided in this application may include non-volatile and / or volatile memory. Non-volatile memory may include read-only memory (ROM), programmable ROM (PROM), electrically programmable ROM (EPROM), electrically erasable programmable ROM (EEPROM) or flash memory. Volatile memory may include random access memory (RAM) or external cache memory. As an illustration and not limitation, RAM is available in many forms, such as static RAM (SRAM), dynamic RAM (DRAM), synchronous DRAM (SDRAM), double data rate SDRAM (DDRSDRAM), enhanced SDRAM (ESDRAM), synchronous link (Synchlink) DRAM (SLDRAM), memory bus (Rambus) direct RAM (RDRAM), direct memory bus dynamic RAM (DRDRAM), and memory bus dynamic RAM (RDRAM).

[0147] The technical features of the above-described embodiments may be arbitrarily combined. To make the description concise, not all possible combinations of the technical features in the above-described embodiments are described. However, as long as there is no contradiction in the combination of these technical features, they should be considered to be within the scope of this specification.

[0148] The above-mentioned embodiments only express several implementation methods of the present invention, and the description thereof is relatively specific and detailed, but it cannot be understood as limiting the scope of the patent of the present invention. It should be pointed out that, for ordinary technicians in this field, several variations and improvements can be made without departing from the concept of the present invention, which all belong to the protection scope of the present invention. Therefore, the protection scope of the patent of the present invention shall be subject to the attached claims.

[0149] The above description is only a preferred embodiment of the present invention and is not intended to limit the present invention. Any modifications, equivalent substitutions and improvements made within the spirit and principles of the present invention should be included in the protection scope of the present invention.

Claims

1. A distributed database management method based on artificial intelligence, characterized in that: The method comprises: Read and analyze the database content, divide the database content into trust levels based on data importance, sensitivity and access frequency, analyze the logical associations between the data in the database, identify naturally aggregated data clusters with natural aggregation characteristics, and set data allocation rules to distribute the database content; Evaluate the status of all existing distributed nodes in real time, and dynamically adjust the storage layout of data in distributed nodes based on the evaluation results; Establish a stable analysis model, obtain the current network status and node status data, and analyze the current network status and node status based on the stable analysis model to determine whether there are any abnormalities; Equip the drone with a storage communication module, and when any abnormality in the network status or node status is identified, distribute the data in the corresponding abnormal distributed node to the drone according to the storage result of the dynamic storage strategy, and group the drones; Optimize the data storage mode according to the access request to guide the access request to the matching node, and adjust the number of drones in the corresponding group according to the data access frequency and network conditions; Continuously monitor the network environment and node status that have been determined to be abnormal, and recycle the drone nodes when the abnormality disappears; Among them, the storage communication module is equipped for the drone, and when any abnormality in the network state or node state is identified, the data in the corresponding abnormal distributed node is allocated to the drone according to the storage result of the dynamic storage strategy, and the drone is grouped, which specifically includes: Based on the analysis results of the stable analysis model, determine whether there are marked nodes, and identify all marked nodes; Analyze the content of the marked nodes, and allocate a corresponding number of drone groups to each marked node according to the storage content size, and synchronously transfer the content in the node to the corresponding drone group; Based on the access frequency and access history of the data, the same grouping tags are added to the data shards that are accessed simultaneously, and the drone groups are regrouped according to the grouping tags.

2. The method according to claim 1, characterized in that The trust level classification of the database content and the analysis of the logical associations between the data in the database to identify natural aggregation data clusters with natural aggregation characteristics specifically include: Read the database content and pre-process the read content; Calculate the business relevance of the data to assess the importance of each data item, identify sensitive data based on information detection, and analyze historical access logs to calculate the access frequency of each data item or data shard; Set up phase intervals respectively, including: importance phase interval, sensitivity phase interval and access frequency phase interval, and assign matching intervals to database contents; Define the trust level standard and calculate the trust level of each data according to the stage interval of the data; The data in the database is mapped into a graph structure, where nodes represent data items and edges represent logical associations between data. The relationship between data is analyzed based on the graph structure, and closely interacting subclusters are identified, which are defined as naturally aggregated data clusters.

3. The method according to claim 2, characterized in that The setting of data distribution rules to distribute the database content specifically includes: Uniformly divide the identified natural aggregate data clusters and trust levels, and package the data into several data shards; Assign trust level weight and natural aggregation weight to each data shard. The higher the trust level and the higher the natural aggregation, the greater the weight value. Set a weight threshold. For data shards where both the trust level weight and the natural aggregation weight are higher than the weight threshold, select nodes with high security and the same geographical location for storage. For data with trust level weights higher than the weight threshold, high-security nodes are selected for storage; For data whose natural aggregation weight is higher than the weight threshold, nodes with the same geographical location are selected for storage, and a backup storage node is set for each storage node, and the backup storage node is not the same node as the original storage node.

4. The method according to claim 3, characterized in that The real-time evaluation of the status of all existing distributed nodes and the dynamic adjustment of the storage layout of data in the distributed nodes based on the evaluation results specifically include: Collect the status data of all distributed nodes, and calculate the load status of each node based on the current resource usage of the node. At the same time, measure the network delay, bandwidth and packet loss rate between nodes to evaluate the communication quality between nodes. When any node is inaccessible, the backup node of the storage node is retrieved according to the data allocation rule, and a communication connection is established with the original storage node; When the performance of any node cannot meet the load requirement, the load status of other nodes under the same data allocation rule as the node is identified, and any other node that meets the migration condition is obtained; The migration condition is: full load status of other nodes - load status of other nodes > load demand of the node; Migrate data shards of nodes whose performance cannot meet the load requirements to other nodes that meet the migration conditions and are under the same allocation rules.

5. The method according to claim 4, characterized in that The establishment of a stable analysis model, obtaining current network status and node status data, and analyzing the current network status and node status based on the stable analysis model to determine whether there is an abnormality, specifically includes: Collect the current network status and node status data in real time, establish a stable analysis model, use the stable analysis model to analyze the network status and node status data, and compare the difference between the real-time data and the predicted data; The predicted data is calculated by the following formula: ; in, is the predicted current time data value, that is, the predicted data, is a constant term, representing the long-term average, is the natural regression coefficient, which is used to weigh the impact of historical data points on the current forecast. For time The data value at It is the moving average coefficient multiplied by the previous error term, which is used to adjust the predicted value and is the error term; A difference threshold is set, and the content where the difference between the predicted data and the real-time data is greater than the difference threshold is defined as abnormal state data, and the nodes related to the data are marked.

6. The method according to claim 5, characterized in that The method optimizes the data storage mode according to the access request to guide the access request to the matching node, and adjusts the number of drones in the corresponding group according to the data access frequency and network conditions, specifically including: Collect and analyze database access logs to determine the frequency, patterns, and trends of data access and identify hot spots of data access; Establish an access prediction model to predict future access requests, and generate a dynamic optimization strategy based on the output of the access prediction model and the rule of balancing the distributed node load; The prediction of future access requests is specifically as follows: ; in, is the predicted access request volume at the next moment, For in time The actual number of access requests at the moment, For time The weight of the moment, is the size of the historical time window; Based on dynamic optimization strategies, the number and grouping of drones are dynamically adjusted.

7. The method according to claim 1, characterized in that The artificial intelligence-based distributed database management system implementing the method includes: The database content analysis module is used to read and analyze the database content, divide the database content into trust levels based on data importance, sensitivity and access frequency, analyze the logical associations between the data in the database, identify natural aggregation data clusters with natural aggregation characteristics, and set data allocation rules to distribute the database content; The storage layout adjustment module is used to evaluate the status of all existing distributed nodes in real time and dynamically adjust the storage layout of data in the distributed nodes based on the evaluation results; The anomaly detection module is used to establish a stable analysis model, obtain the current network status and node status data, and analyze the current network status and node status based on the stable analysis model to determine whether there is an anomaly; The UAV storage communication module is used to equip the UAV with a storage communication module, and when any abnormality in the network state or node state is identified, the data in the corresponding abnormal distributed node is allocated to the UAV according to the storage result of the dynamic storage strategy, and the UAV is grouped; An access request guiding module is used to guide access requests to matching nodes according to the access request optimization data storage mode, and adjust the number of drones in the corresponding group according to the data access frequency and network conditions; The drone node recovery module is used to continuously monitor the network environment and node status that have been determined to be abnormal, and to recover the drone nodes when the abnormality disappears.

Images