Container data storage method based on artificial intelligence and cloud platform

Abstract

The invention is suitable for the technical field of data storage management, and particularly relates to a container data storage method based on artificial intelligence and a cloud platform, the method comprises the following steps: receiving a data storage request, obtaining to-be-stored data, and extracting metadata features; outputting optimal storage position decision information through the artificial intelligence model, and storing the to-be-stored data; performing real-time monitoring, collecting access information of each piece of stored data, re-classifying the stored data based on the access information, and generating optimal migration decision information; and obtaining state information of the storage server, generating a data migration instruction, and migrating the stored data needing to be migrated in batches. The optimal storage position is intelligently selected based on the real-time node state in the data writing stage, the access frequency is continuously monitored in the operation stage, the high-frequency access data is dynamically scheduled to the high-performance node, the I / O delay is reduced, meanwhile, the migration path is optimized through the network topology awareness algorithm, and resource occupation in the transmission aspect is reduced.

Description

TECHNICAL FIELD

[0001] The application belongs to the technical field of data storage management, and particularly relates to a container data storage method based on artificial intelligence and a cloud platform. BACKGROUND

[0002] Container data storage refers to a solution for providing persistent data storage for applications within containers when using container technologies such as Docker, Kubernetes, etc. Unlike the temporary nature of containers, container data storage ensures that data is not lost due to the stopping or deletion of containers. It can be implemented in various ways, including but not limited to mounting using host directories, utilizing network-attached storage (NAS), or through distributed file systems and cloud storage services.

[0003] In the current container storage process, container data is classified during storage, and the type of container data may change during subsequent use, which will cause inaccurate classification of container data storage, affect the effectiveness of data access, and reduce data storage efficiency. SUMMARY

[0004] The purpose of the present application is to provide a container data storage method based on artificial intelligence, which aims to solve the problem of inaccurate classification of container data storage in the current container storage process, which affects the effectiveness of data access and reduces data storage efficiency.

[0005] The present application is implemented as follows: a container data storage method based on artificial intelligence, the method comprising: receiving a data storage request sent from a container application, obtaining to-be-stored data, extracting metadata features of the to-be-stored data, the metadata features at least including data type, access frequency, and data priority; importing the metadata features corresponding to the to-be-stored data into a trained artificial intelligence model, outputting optimal storage location decision information through the artificial intelligence model, and storing the to-be-stored data; real-time monitoring of the stored data, collecting access information of each stored data, reclassifying the stored data based on the access information, and generating optimal migration decision information; obtaining storage server state information through the artificial intelligence model, generating data migration instructions, and migrating the stored data that needs to be migrated in batches.

[0006] Preferably, the step of importing the metadata features corresponding to the to-be-stored data into a trained artificial intelligence model, outputting optimal storage location decision information through the artificial intelligence model, and storing the to-be-stored data comprises: The normalized metadata features are imported into the pre-trained artificial intelligence model, and the preset optimization target is retrieved. The artificial intelligence model is invoked, and the optimal storage location decision information is output based on metadata features and server status information. The optimal storage location decision information includes the optimal storage node and the corresponding storage path. Determine the optimal location for storing each piece of data to be stored, and store each piece of data to be stored in the corresponding optimal storage node according to the corresponding storage path.

[0007] Preferably, the steps of real-time monitoring of stored data, collecting access information for each stored data item, reclassifying the stored data based on the access information, and generating optimal migration decision information specifically include: Real-time monitoring of stored data; access information for each stored data item is obtained at preset time intervals. A corresponding access frequency prediction function is constructed based on the access information, and the time interval for obtaining access information is dynamically adjusted based on the access frequency prediction function. Based on access information, the stored data is reclassified to determine the latest classification of each stored data item and generate optimal migration decision information.

[0008] Preferably, the step of obtaining storage server status information through an artificial intelligence model, generating data migration instructions, and migrating the stored data to be migrated in batches specifically includes: When it is determined to migrate the stored data, the storage server status information is obtained in real time. The storage server status information includes network topology, inter-node bandwidth status and load status. The server storage status information and corresponding optimal migration decision information are matched with the stored data that needs to be migrated to determine the migration location of the stored data and generate data migration instructions. According to the data migration instructions, access each stored data that needs to be migrated, perform location migration on the stored data, and update the location information of the stored data.

[0009] Preferably, the steps of receiving a data storage request from a container application, obtaining data to be stored, and extracting metadata features of the data to be stored include: receiving a data storage request from a container application, extracting metadata features of the data to be stored, including data type, estimated access frequency, and data priority; collecting resource status indicators of each node in the storage server in real time, including CPU utilization, remaining memory, storage I / O speed, and network bandwidth; and normalizing the metadata features to obtain normalized metadata features.

[0010] Another objective of this invention is to provide an artificial intelligence-based container data storage cloud platform, the cloud platform comprising: The data acquisition module is used to receive data storage requests sent from container applications, acquire data to be stored, and extract metadata features of the data to be stored. The metadata features include at least data type, access frequency, and data priority. The data storage module is used to import the metadata features corresponding to the data to be stored into the trained artificial intelligence model, and then use the artificial intelligence model to output the optimal storage location decision information to store the data to be stored. The migration decision module is used to monitor the stored data in real time, collect access information of each stored data, reclassify the stored data based on the access information, and generate optimal migration decision information. The data migration module is used to obtain storage server status information through artificial intelligence models, generate data migration instructions, and migrate the stored data that needs to be migrated in batches.

[0011] Preferably, the data storage module includes: The data input unit is used to import normalized metadata features into a pre-trained artificial intelligence model and retrieve preset optimization targets. The decision information generation unit is used to retrieve the artificial intelligence model and output the optimal storage location decision information based on metadata features and server status information. The optimal storage location decision information includes the optimal storage node and the corresponding storage path. The storage management unit is used to determine the optimal location for storing each piece of data to be stored, and to store each piece of data to be stored in the corresponding optimal storage node according to the corresponding storage path.

[0012] Preferably, the migration decision module includes: Access information acquisition unit is used to monitor the stored data in real time and acquire access information of each stored data according to a preset time interval; The time interval adjustment unit is used to construct a corresponding access frequency prediction function based on the access information, and to dynamically adjust the time interval for obtaining access information based on the access frequency prediction function. The decision information generation unit is used to reclassify the stored data based on the access information, determine the latest classification of each stored data, and generate the optimal migration decision information.

[0013] Preferably, the data migration module includes: The status information acquisition unit is used to acquire the storage server status information in real time when it is determined that the stored data needs to be migrated. The storage server status information includes network topology, inter-node bandwidth status and load status. The migration instruction generation unit is used to match the server storage status information and the corresponding optimal migration decision information with the stored data that needs to be migrated, determine the migration location of the stored data, and generate data migration instructions. The location update unit is used to access each stored data that needs to be migrated according to the data migration instruction, perform location migration on the stored data, and update the location information of the stored data.

[0014] Preferably, the steps of receiving a data storage request from a container application, obtaining data to be stored, and extracting metadata features of the data to be stored include: receiving a data storage request from a container application, extracting metadata features of the data to be stored, including data type, estimated access frequency, and data priority; collecting resource status indicators of each node in the storage server in real time, including CPU utilization, remaining memory, storage I / O speed, and network bandwidth; and normalizing the metadata features to obtain normalized metadata features.

[0015] This invention provides an artificial intelligence-based container data storage method that uses an artificial intelligence model to dynamically decide the storage location and migration strategy of container data, significantly improving storage performance and resource utilization. During the data writing phase, the optimal storage location is intelligently selected based on the real-time node status. During the operation phase, the access frequency is continuously monitored, and high-frequency access data is dynamically scheduled to high-performance nodes to reduce I / O latency. At the same time, the migration path is optimized using a network topology awareness algorithm to reduce the resource consumption in transmission. Attached Figure Description

[0016] Figure 1 A flowchart illustrating an artificial intelligence-based container data storage method provided in an embodiment of the present invention; Figure 2 The flowchart illustrates the steps of storing data to be stored, provided in this embodiment of the invention, which involve importing metadata features corresponding to the data to be stored into a trained artificial intelligence model and then outputting optimal storage location decision information through the artificial intelligence model. Figure 3 The flowchart illustrates the steps of real-time monitoring of stored data, collecting access information of each stored data item, reclassifying the stored data based on the access information, and generating optimal migration decision information, as provided in this embodiment of the invention. Figure 4 The flowchart illustrates the steps of obtaining storage server status information through an artificial intelligence model, generating data migration instructions, and migrating the stored data to be migrated in batches, as provided in this embodiment of the invention. Figure 5 An architecture diagram of an artificial intelligence-based container data storage cloud platform provided for an embodiment of the present invention; Figure 6 An architecture diagram of a data storage module provided in an embodiment of the present invention; Figure 7 An architecture diagram of a migration decision module provided in an embodiment of the present invention; Figure 8 This is an architecture diagram of a data migration module provided in an embodiment of the present invention. Detailed Implementation

[0017] To make the objectives, technical solutions, and advantages of this invention clearer, the invention will be further described in detail below with reference to the accompanying drawings and embodiments. It should be understood that the specific embodiments described herein are merely illustrative and not intended to limit the invention.

[0018] like Figure 1 The diagram shown is a flowchart of a container data storage method based on artificial intelligence provided in an embodiment of the present invention. The method includes: S100 receives a data storage request sent from a container application, obtains the data to be stored, and extracts the metadata features of the data to be stored. The metadata features include at least the data type, access frequency, and data priority.

[0019] In this step, data storage requests are received from container applications. Each container application actively initiates a data storage request when it needs to store data. The data storage request is received by the data management center, which can manage the storage server according to the data storage request to realize the storage of data. After receiving the data storage request, the corresponding data to be stored is received synchronously, and information is extracted from the data to be stored to determine the data type, access frequency and data priority. The system receives data storage requests from container applications, extracts metadata features of the data to be stored, including data type, estimated access frequency, and data priority; collects resource status indicators of each node in the storage server in real time, including CPU utilization, remaining memory, storage I / O speed, and network bandwidth; and normalizes the metadata features to obtain normalized metadata features.

[0020] S200 imports the metadata features corresponding to the data to be stored into the trained artificial intelligence model, and then stores the data by outputting the optimal storage location decision information through the artificial intelligence model.

[0021] In this step, the metadata features corresponding to the data to be stored are imported into the trained artificial intelligence model. The artificial intelligence model adopts a dual-model collaborative architecture of deep reinforcement learning model and graph neural network, which includes a DRL decision subject and a GNN encoder. The DRL decision subject adopts a dynamic learning storage strategy, optimizes the location decision based on long-term benefits, and processes the cluster topology relationship through the GNN encoder to capture the correlation features such as network latency and bandwidth between nodes. A fusion mechanism is used to combine the topology features extracted by the GNN with the DRL state vector to form the final optimal storage location decision information, and the data to be stored is stored accordingly.

[0022] The S300 performs real-time monitoring of stored data, collects access information for each stored data item, reclassifies the stored data based on the access information, and generates optimal migration decision information.

[0023] In this step, the stored data is monitored in real time. After each piece of stored data is stored, in order to determine the access status of each piece of data, the stored data is accessed at preset time intervals to read relevant information of the stored data, including access information, i.e., the access frequency of the stored data. Based on the access frequency of the stored data, it is determined whether the currently stored data can be classified into other types. If the preset migration conditions are met, optimal migration decision information is generated. The optimal migration decision information is used to determine the stored data that needs to be migrated.

[0024] The S400 uses an artificial intelligence model to obtain storage server status information, generates data migration instructions, and migrates the stored data that needs to be migrated in batches.

[0025] In this step, the storage server status information is obtained through an artificial intelligence model. When a migration is determined, the migration location and path need to be determined based on the storage server status information to save server computing resources. The artificial intelligence model analyzes the migration information of each stored data that needs to be migrated, including the migration location and corresponding migration path, to generate corresponding data migration instructions. The data migration instructions are executed by the storage server, which completes the migration of the stored data that needs to be migrated and updates the location information of the stored data for subsequent access.

[0026] like Figure 2 As shown, in a preferred embodiment of the present invention, the step of importing the metadata features corresponding to the data to be stored into a trained artificial intelligence model, and storing the data to be stored by outputting the optimal storage location decision information through the artificial intelligence model, specifically includes: S201 imports normalized metadata features into a pre-trained artificial intelligence model and retrieves a preset optimization objective.

[0027] In this step, normalized metadata features are retrieved and used as input to the artificial intelligence model. The artificial intelligence model analyzes the input metadata features to determine the attributes of each piece of data to be stored and to determine the corresponding optimization objectives. The optimization objectives can be performance priority, cost priority, or load balancing priority.

[0028] S202, retrieve the artificial intelligence model, and output the optimal storage location decision information based on metadata features and server status information. The optimal storage location decision information includes the optimal storage node and the corresponding storage path.

[0029] In this step, the artificial intelligence model is invoked. The artificial intelligence model has been trained and can analyze the metadata characteristics of the data to be stored. Based on the corresponding optimization objectives, it outputs the optimal storage location decision information. The optimal storage location decision information includes the optimal storage node and the corresponding storage path, that is, determining the storage server and the corresponding storage node for the data to be stored. The storage path is the flow path when the data to be stored arrives at the storage node.

[0030] S203, determine the optimal location for storing each piece of data to be stored, and store each piece of data to be stored in the corresponding optimal storage node according to the corresponding storage path.

[0031] In this step, the optimal location storage decision information for each piece of data to be stored is determined. After analysis by an artificial intelligence model, the corresponding storage location and storage path for each piece of data to be stored are determined. The data to be stored is then stored according to the above optimal location storage decision information. During this process, the data to be stored is classified according to its data type, access frequency, and data priority.

[0032] like Figure 3 As shown, in a preferred embodiment of the present invention, the steps of real-time monitoring of stored data, collecting access information of each stored data item, reclassifying the stored data based on the access information, and generating optimal migration decision information specifically include: S301 performs real-time monitoring of stored data and retrieves access information for each stored data item at preset time intervals.

[0033] In this step, the stored data is monitored in real time. After storing the data to be stored based on the optimal storage location decision information, it is transformed into stored data. In order to facilitate the monitoring of each stored data, it is necessary to determine the access information of each stored data, that is, the access frequency.

[0034] S302, construct a corresponding access frequency prediction function based on the access information, and dynamically adjust the time interval for obtaining access information based on the access frequency prediction function.

[0035] In this step, a corresponding access frequency prediction function is constructed based on the access information. When determining the time interval, if a fixed time interval is used, then for stored data with little change in access frequency, it is considered an invalid access. However, for stored data with significant change in access frequency, a fixed time interval cannot detect the change in access frequency in time. Therefore, the time of obtaining access information and the corresponding access frequency within a short period of time are recorded, and a corresponding access coordinate (t, P) is constructed, where t represents the time of obtaining the access frequency of the stored data, and P is the access frequency read in this instance. Based on the access coordinate, a function is fitted, with access time as the independent variable and access frequency as the dependent variable, to obtain the access frequency prediction function. The access frequency prediction function is used to determine the change in access frequency within a short period of time. The faster the change in access frequency, the shorter the time interval; the slower the change, the longer the time interval.

[0036] S303: Based on the access information, the stored data is reclassified to determine the latest classification of each stored data item and generate optimal migration decision information.

[0037] In this step, the stored data is reclassified based on access information. Specifically, it is classified according to access frequency. The stored data is classified according to a preset access frequency range, such as dividing it into three ranges: AB, BC, and CD. Data falling into different ranges is classified into different types, thereby determining the stored data that needs to be migrated. For example, if a piece of stored data is located in the AB range, but after a period of time it falls into the CD range, then the stored data needs to be migrated.

[0038] like Figure 4 As shown, in a preferred embodiment of the present invention, the step of obtaining storage server status information through an artificial intelligence model, generating data migration instructions, and migrating the stored data to be migrated in batches specifically includes: S401: When it is determined that stored data needs to be migrated, the storage server status information is obtained in real time. The storage server status information includes network topology, inter-node bandwidth status and load status.

[0039] In this step, during data migration, each storage server is traversed. When accessing the storage server, the corresponding status information is retrieved, such as determining the network topology between nodes, the available transmission bandwidth between nodes, and the load status of each node.

[0040] S402, match the server storage status information and the corresponding optimal migration decision information with the stored data that needs to be migrated, determine the migration location of the stored data, and generate a data migration instruction.

[0041] S403: Access each stored data that needs to be migrated according to the data migration instruction, perform location migration on the stored data, and update the location information of the stored data.

[0042] In this step, the server storage status information and corresponding optimal migration decision information are matched with the stored data that needs to be migrated. After classifying the stored data, the matching relationship between two groups of stored data is determined based on the changes in the classification of the stored data. For example, if stored data a falls into the AB interval and stored data b falls into the CD interval, after reclassification, a falls into the CD interval and b falls into the AB interval. Then a and b are considered to be data with a matching relationship, and their data positions can be directly swapped. For stored data without a matching object, the corresponding storage position is reallocated. After determining the migration position of each stored data, a data migration instruction is generated. According to the data migration instruction, each stored data that needs to be migrated is accessed, the stored data is migrated, and the position information of the stored data is updated.

[0043] like Figure 5 As shown, an artificial intelligence-based container data storage cloud platform is provided in an embodiment of the present invention. The cloud platform includes: The data acquisition module 100 is used to receive data storage requests sent from container applications, acquire data to be stored, and extract metadata features of the data to be stored. The metadata features include at least data type, access frequency, and data priority.

[0044] In this cloud platform, the data acquisition module 100 receives data storage requests sent by container applications. Each container application actively initiates a data storage request when it needs to store data. The data storage request is received by the data management center, which can manage the storage server according to the data storage request to realize the storage of data. After receiving the data storage request, the module synchronously receives the corresponding data to be stored and extracts information from the data to be stored to determine the data type, access frequency and data priority.

[0045] The data storage module 200 is used to import the metadata features corresponding to the data to be stored into the trained artificial intelligence model, and then use the artificial intelligence model to output the optimal storage location decision information to store the data to be stored.

[0046] In this cloud platform, the data storage module 200 imports the metadata features corresponding to the data to be stored into the trained artificial intelligence model. The artificial intelligence model adopts a dual-model collaborative architecture of deep reinforcement learning model and graph neural network, which includes a DRL decision subject and a GNN encoder. The DRL decision subject adopts a dynamic learning storage strategy, optimizes the location decision based on long-term benefits, and processes the cluster topology relationship through the GNN encoder, capturing the correlation features such as network latency and bandwidth between nodes. A fusion mechanism is used to combine the topology features extracted by the GNN with the DRL state vector to form the final optimal storage location decision information, and stores the data to be stored accordingly.

[0047] The migration decision module 300 is used to monitor the stored data in real time, collect access information of each stored data, reclassify the stored data based on the access information, and generate optimal migration decision information.

[0048] In this cloud platform, the migration decision module 300 monitors the stored data in real time. After each piece of stored data is stored, in order to determine the access status of each piece of data, the module accesses the stored data at preset time intervals to read relevant information of the stored data, including access information, i.e., the access frequency of the stored data. Based on the access frequency of the stored data, the module determines whether the currently stored data can be classified into other types. If the preset migration conditions are met, the module generates optimal migration decision information. The optimal migration decision information is used to determine the stored data that needs to be migrated.

[0049] The data migration module 400 is used to obtain storage server status information through an artificial intelligence model, generate data migration instructions, and migrate the stored data that needs to be migrated in batches.

[0050] In this cloud platform, the data migration module 400 obtains storage server status information through an artificial intelligence model. When a migration is determined, the migration location and path are determined based on the storage server status information to save server computing resources. The artificial intelligence model analyzes the migration information of each stored data that needs to be migrated, including the migration location and corresponding migration path, to generate corresponding data migration instructions. The data migration instructions are executed by the storage server, which completes the migration of the stored data that needs to be migrated and updates the location information of the stored data for subsequent access.

[0051] like Figure 6 As shown, in a preferred embodiment of the present invention, the data storage module 200 includes: The data input unit 201 is used to import normalized metadata features into a pre-trained artificial intelligence model and retrieve a preset optimization target.

[0052] In this module, the data input unit 201 retrieves the normalized metadata features. The normalized metadata features serve as the input to the artificial intelligence model. The artificial intelligence model analyzes the input metadata features to determine the attributes of each piece of data to be stored and to determine the corresponding optimization objectives. The optimization objectives can be performance priority, cost priority, or load balancing priority.

[0053] The decision information generation unit 202 is used to retrieve the artificial intelligence model and output the optimal storage location decision information based on metadata features and server status information. The optimal storage location decision information includes the optimal storage node and the corresponding storage path.

[0054] In this module, the decision information generation unit 202 retrieves the artificial intelligence model. The artificial intelligence model has been trained and can analyze the metadata characteristics of the data to be stored. Based on the corresponding optimization objectives, it outputs the optimal storage location decision information. The optimal storage location decision information includes the optimal storage node and the corresponding storage path, that is, determining the storage server and the corresponding storage node for the data to be stored. The storage path is the flow path when the data to be stored arrives at the storage node.

[0055] The storage management unit 203 is used to determine the optimal location storage decision information for each piece of data to be stored, and to store each piece of data to be stored to the corresponding optimal storage node according to the corresponding storage path.

[0056] In this module, the storage management unit 203 determines the optimal location storage decision information for each piece of data to be stored. After analysis by an artificial intelligence model, the corresponding storage location and storage path for each piece of data to be stored are determined. The data to be stored is then stored according to the optimal location storage decision information. During this process, the data to be stored is classified according to its data type, access frequency, and data priority.

[0057] like Figure 7 As shown, in a preferred embodiment of the present invention, the migration decision module 300 includes: Access information acquisition unit 301 is used to monitor the stored data in real time and acquire access information of each stored data according to a preset time interval.

[0058] In this module, the access information acquisition unit 301 monitors the stored data in real time. After storing the data to be stored based on the optimal location storage decision information, it is transformed into stored data. In order to facilitate the monitoring of each stored data, it is necessary to determine the access information of each stored data, that is, the access frequency.

[0059] The time interval adjustment unit 302 is used to construct a corresponding access frequency prediction function based on the access information, and to dynamically adjust the time interval for obtaining access information based on the access frequency prediction function.

[0060] In this module, the time interval adjustment unit 302 constructs a corresponding access frequency prediction function based on the access information. When determining the time interval, if a fixed time interval is used, then for stored data with little change in access frequency, it is considered an invalid access. However, for stored data with significant change in access frequency, a fixed time interval cannot detect the change in access frequency in time. Therefore, the time of obtaining access information and the corresponding access frequency within a short period of time are recorded, and a corresponding access coordinate (t, P) is constructed, where t represents the time of obtaining the access frequency of the stored data, and P is the access frequency read in this instance. Based on the access coordinate, a function is fitted to determine that access time is the independent variable and access frequency is the dependent variable, and the access frequency prediction function is obtained. The access frequency prediction function is used to determine the change in access frequency within a short period of time. The faster the change in access frequency, the shorter the time interval; the slower the change, the longer the time interval.

[0061] The decision information generation unit 303 is used to reclassify the stored data based on the access information, determine the latest classification of each stored data, and generate the optimal migration decision information.

[0062] In this module, the decision information generation unit 303 reclassifies the stored data based on access information. Specifically, it classifies the data according to access frequency and according to a preset access frequency range, such as dividing it into three ranges: AB, BC, and CD. Data falling into different ranges is classified into different types, thereby determining the stored data that needs to be migrated. For example, if a piece of stored data is located in the AB range, but after a period of time it falls into the CD range, then the stored data needs to be migrated.

[0063] like Figure 8 As shown, in a preferred embodiment of the present invention, the data migration module 400 includes: The status information acquisition unit 401 is used to acquire the storage server status information in real time when it is determined that the stored data needs to be migrated. The storage server status information includes network topology, inter-node bandwidth status and load status.

[0064] In this module, when the status information acquisition unit 401 performs data migration, it traverses each storage server and retrieves the corresponding status information when accessing the storage server, such as determining the network topology between each node, the available transmission bandwidth between nodes, and the load status of each node.

[0065] The migration instruction generation unit 402 is used to match the server storage status information and the corresponding optimal migration decision information with the stored data that needs to be migrated, determine the migration location of the stored data, and generate a data migration instruction.

[0066] The location update unit 403 is used to access each stored data that needs to be migrated according to the data migration instruction, perform location migration on the stored data, and update the location information of the stored data.

[0067] In this module, server storage status information and corresponding optimal migration decision information are matched with the stored data that needs to be migrated. After classifying the stored data, the matching relationship between two groups of stored data is determined based on the changes in the classification of the stored data. For example, if stored data a falls into the AB interval and stored data b falls into the CD interval, after reclassification, a falls into the CD interval and b falls into the AB interval. Then a and b are considered to be data with a matching relationship, and their data positions can be directly swapped. For stored data without a matching object, the corresponding storage position is reallocated. After determining the migration position of each stored data, a data migration instruction is generated. According to the data migration instruction, each stored data that needs to be migrated is accessed, the stored data is migrated, and the position information of the stored data is updated.

[0068] 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 within the protection scope of the present invention.

Claims

1. A container data storage method based on artificial intelligence, characterized in that, The method includes: Receive data storage requests from container applications, obtain the data to be stored, and extract the metadata features of the data to be stored. The metadata features include at least the data type, access frequency, and data priority. The metadata features corresponding to the data to be stored are imported into the trained artificial intelligence model. The artificial intelligence model outputs the optimal storage location decision information and stores the data to be stored. Real-time monitoring of stored data; collection of access information for each stored data item; reclassification of stored data based on access information; and generation of optimal migration decision information. The system uses an artificial intelligence model to obtain storage server status information, generates data migration instructions, and migrates the stored data that needs to be migrated in batches.

2. The container data storage method based on artificial intelligence according to claim 1, characterized in that, The step of importing the metadata features corresponding to the data to be stored into a trained artificial intelligence model, and storing the data by outputting the optimal storage location decision information through the artificial intelligence model, specifically includes: The normalized metadata features are imported into the pre-trained artificial intelligence model, and the preset optimization target is retrieved. The artificial intelligence model is invoked, and the optimal storage location decision information is output based on metadata features and server status information. The optimal storage location decision information includes the optimal storage node and the corresponding storage path. Determine the optimal location for storing each piece of data to be stored, and store each piece of data to be stored in the corresponding optimal storage node according to the corresponding storage path.

3. The container data storage method based on artificial intelligence according to claim 2, characterized in that, The steps of real-time monitoring of stored data, collecting access information for each stored data item, reclassifying the stored data based on the access information, and generating optimal migration decision information specifically include: Real-time monitoring of stored data; access information for each stored data item is obtained at preset time intervals. A corresponding access frequency prediction function is constructed based on the access information, and the time interval for obtaining access information is dynamically adjusted based on the access frequency prediction function. Based on access information, the stored data is reclassified to determine the latest classification of each stored data item and generate optimal migration decision information.

4. The container data storage method based on artificial intelligence according to claim 2, characterized in that, The steps of obtaining storage server status information through an artificial intelligence model, generating data migration instructions, and migrating the stored data to be migrated in batches specifically include: When it is determined to migrate the stored data, the storage server status information is obtained in real time. The storage server status information includes network topology, inter-node bandwidth status and load status. The server storage status information and corresponding optimal migration decision information are matched with the stored data that needs to be migrated to determine the migration location of the stored data and generate data migration instructions. According to the data migration instructions, access each stored data that needs to be migrated, perform location migration on the stored data, and update the location information of the stored data.

5. The container data storage method based on artificial intelligence according to claim 1, characterized in that, The steps of receiving a data storage request from a container application, obtaining the data to be stored, and extracting the metadata features of the data to be stored include: receiving a data storage request from a container application, extracting the metadata features of the data to be stored, including data type, estimated access frequency, and data priority; collecting resource status indicators of each node in the storage server in real time, including CPU utilization, remaining memory, storage I / O speed, and network bandwidth; and normalizing the metadata features to obtain normalized metadata features.

6. A container data storage cloud platform based on artificial intelligence, characterized in that, The cloud platform includes: The data acquisition module is used to receive data storage requests sent from container applications, acquire data to be stored, and extract metadata features of the data to be stored. The metadata features include at least data type, access frequency, and data priority. The data storage module is used to import the metadata features corresponding to the data to be stored into the trained artificial intelligence model, and then use the artificial intelligence model to output the optimal storage location decision information to store the data to be stored. The migration decision module is used to monitor the stored data in real time, collect access information of each stored data, reclassify the stored data based on the access information, and generate optimal migration decision information. The data migration module is used to obtain storage server status information through artificial intelligence models, generate data migration instructions, and migrate the stored data that needs to be migrated in batches.

7. The artificial intelligence-based container data storage cloud platform according to claim 6, characterized in that, The data storage module includes: The data input unit is used to import normalized metadata features into a pre-trained artificial intelligence model and retrieve preset optimization targets. The decision information generation unit is used to retrieve the artificial intelligence model and output the optimal storage location decision information based on metadata features and server status information. The optimal storage location decision information includes the optimal storage node and the corresponding storage path. The storage management unit is used to determine the optimal location for storing each piece of data to be stored, and to store each piece of data to be stored in the corresponding optimal storage node according to the corresponding storage path.

8. The artificial intelligence-based container data storage cloud platform according to claim 7, characterized in that, The migration decision module includes: Access information acquisition unit is used to monitor the stored data in real time and acquire access information of each stored data according to a preset time interval; The time interval adjustment unit is used to construct a corresponding access frequency prediction function based on the access information, and to dynamically adjust the time interval for obtaining access information based on the access frequency prediction function. The decision information generation unit is used to reclassify the stored data based on the access information, determine the latest classification of each stored data, and generate the optimal migration decision information.

9. The artificial intelligence-based container data storage cloud platform according to claim 7, characterized in that, The data migration module includes: The status information acquisition unit is used to acquire the storage server status information in real time when it is determined that the stored data needs to be migrated. The storage server status information includes network topology, inter-node bandwidth status and load status. The migration instruction generation unit is used to match the server storage status information and the corresponding optimal migration decision information with the stored data that needs to be migrated, determine the migration location of the stored data, and generate data migration instructions. The location update unit is used to access each stored data that needs to be migrated according to the data migration instruction, perform location migration on the stored data, and update the location information of the stored data.

10. The artificial intelligence-based container data storage cloud platform according to claim 6, characterized in that, The steps of receiving a data storage request from a container application, obtaining the data to be stored, and extracting the metadata features of the data to be stored include: receiving a data storage request from a container application, extracting the metadata features of the data to be stored, including data type, estimated access frequency, and data priority; collecting resource status indicators of each node in the storage server in real time, including CPU utilization, remaining memory, storage I / O speed, and network bandwidth; and normalizing the metadata features to obtain normalized metadata features.

Images