Cloud disk space planning method and device, equipment, storage medium and program product

Abstract

The invention discloses a cloud disk space planning method and device, equipment, a storage medium and a program product, and is applied to the technical field of storage resource management. Cloud disk feature data corresponding to multiple users and a current to-be-planned space are acquired; inputting the current to-be-planned space and the plurality of cloud disk feature data into a first cloud disk space planning model to obtain a to-be-planned space of the user group; inputting the to-be-planned space of the user group and the cloud disk feature data corresponding to each user into a second cloud disk space planning model to obtain the to-be-planned space of each user; obtaining cloud disk historical operation data of the user, inputting the cloud disk historical operation data of the user, a to-be-planned space of the user and cloud disk feature data corresponding to the user into a third cloud disk space planning model, and generating a space allocation execution strategy of the user; cloud disk space planning is performed according to the space allocation execution strategy, so that the cloud disk space is dynamically planned for the user, and waste of cloud disk space resources is avoided.

Description

Technical Field

[0001] This application belongs to the field of storage resource management technology, and in particular relates to a cloud disk space planning method, apparatus, equipment, storage medium and program product. Background Technology

[0002] With the rapid development of cloud computing technology, users are increasingly reliant on cloud storage services, and efficient management of cloud storage space has become a core element in ensuring service quality and improving resource utilization.

[0003] In existing technologies, cloud disk space planning mainly relies on manual operation and static resource allocation strategies. In other words, operations and maintenance personnel allocate fixed cloud disk space to users based on their storage needs, and only when the user's cloud disk space reaches the capacity threshold will the operations and maintenance personnel allocate new storage space to the user.

[0004] However, when faced with a large number of users and complex storage needs, existing technologies are inefficient in cloud disk space planning and are prone to uneven distribution of cloud disk space, resulting in a waste of cloud disk space resources. Summary of the Invention

[0005] This application provides a cloud disk space planning method, apparatus, device, storage medium, and program product, which can dynamically plan cloud disk space for different users in real time according to their storage needs, thereby avoiding waste of cloud disk space resources.

[0006] In a first aspect, embodiments of this application provide a cloud disk space planning method, the method comprising: Obtain cloud disk feature data for multiple users and the current unplanned space shared by multiple users; The current space to be planned and the cloud disk feature data corresponding to multiple users are input into the first cloud disk space planning model. The first cloud disk space planning model is used to plan space for user groups, resulting in one or more user groups to be planned. The user groups are classified based on the cloud disk feature data corresponding to multiple users. For each user group, the space to be planned for that user group and the cloud disk feature data corresponding to each user in that user group are input into the second cloud disk space planning model. The second cloud disk space planning model is used to plan the space for each user and obtain the space to be planned for each user. For each user, obtain the user's cloud disk historical operation data, input the user's cloud disk historical operation data, the user's space to be planned, and the user's corresponding cloud disk characteristic data into the third cloud disk space planning model, and generate the user's space allocation execution strategy through the third cloud disk space planning model. Cloud disk space planning is performed based on the space allocation strategy for each user.

[0007] Secondly, embodiments of this application provide a cloud storage space planning device, the device comprising: The acquisition module is used to acquire cloud disk feature data corresponding to multiple users and the current unplanned space corresponding to multiple users. The first processing module is used to input the current space to be planned and the cloud disk feature data corresponding to multiple users into the first cloud disk space planning model, and to perform space planning for user groups through the first cloud disk space planning model to obtain the space to be planned for one or more user groups. The user groups are classified based on the cloud disk feature data corresponding to multiple users. The second processing module is used to input the planned space of each user group and the cloud disk feature data corresponding to each user in the user group into the second cloud disk space planning model, and to perform space planning for each user through the second cloud disk space planning model to obtain the planned space for each user. The third processing module is used to obtain the user's cloud disk historical operation data for each user, input the user's cloud disk historical operation data, the user's space to be planned, and the user's corresponding cloud disk feature data into the third cloud disk space planning model, and generate the user's space allocation execution strategy through the third cloud disk space planning model. The planning module is used to plan cloud disk space based on the space allocation strategy for each user.

[0008] Thirdly, embodiments of this application provide an electronic device, the device comprising: A processor and a memory storing computer program instructions; a cloud disk space planning method that implements any of the above when the processor executes the computer program instructions.

[0009] Fourthly, embodiments of this application provide a computer storage medium on which computer program instructions are stored, and when the computer program instructions are executed by a processor, the cloud disk space planning method described above is implemented.

[0010] Fifthly, embodiments of this application provide a computer program product in which instructions, when executed by the processor of an electronic device, enable the electronic device to execute any of the cloud disk space planning methods described above.

[0011] The cloud disk space planning method, apparatus, device, storage medium, and program product of this application embodiment obtain cloud disk feature data corresponding to multiple users and the current space to be planned corresponding to multiple users. Then, the current space to be planned and the cloud disk feature data corresponding to multiple users are input into a first cloud disk space planning model. The first cloud disk space planning model is used to plan space for user groups to obtain the space to be planned for one or more user groups. Then, for each user group, the space to be planned for the user group and the cloud disk feature data corresponding to each user in the user group are input into a second cloud disk space planning model. The second cloud disk space planning model is used to plan space for each user to obtain the space to be planned for each user. Finally, for each user, the user's cloud disk historical operation data is obtained. The user's cloud disk historical operation data, the user's space to be planned, and the user's corresponding cloud disk feature data are input into a third cloud disk space planning model. The third cloud disk space planning model is used to generate the space allocation execution strategy for the user. Finally, cloud disk space planning is performed based on the space allocation execution strategy for each user. Compared to existing technologies, cloud disk space planning is inefficient when facing a large number of users and complex storage needs. This application improves the efficiency of cloud disk space planning by acquiring users' cloud disk characteristic data and the space to be planned in real time, and automatically generating space allocation execution strategies for each user through a first cloud disk space planning model, a second cloud disk space planning model, and a third cloud disk space planning model. Furthermore, the first cloud disk space planning model first plans space for user groups to identify the cloud disk space needs of different groups, achieving differentiated allocation among user groups and reducing resource imbalance. Planning space for user groups first allows for the formulation of a unified space allocation strategy for the common needs of the same group, avoiding the repetitive operation of planning space individually for each user from the beginning, further improving the efficiency of cloud disk space planning. Then, based on the second cloud disk space planning model, space planning is performed for each user, which can improve the balance of space planning for each user. Then, based on the third cloud disk space planning model and the historical operation data of each user's cloud disk, the space allocation execution strategy for each user is output. By using the trend information provided by the historical operation data of the cloud disk, the space allocation execution strategy is transformed from static passive allocation to active prediction, which improves the foresight and accuracy of the space allocation execution strategy and reduces the waste of cloud disk space resources. Attached Figure Description

[0012] To more clearly illustrate the technical solutions of the embodiments of this application, the accompanying drawings used in the embodiments of this application will be briefly introduced below. For those skilled in the art, other drawings can be obtained based on these drawings without creative effort.

[0013] Figure 1A flowchart illustrating a cloud disk space planning method according to an embodiment of the present invention is shown. Figure 2 A flowchart illustrating a cloud disk space planning method according to another embodiment of the present invention is shown; Figure 3 A flowchart illustrating a cloud disk space planning method according to another embodiment of the present invention is shown; Figure 4 A flowchart illustrating a cloud disk space planning method according to another embodiment of the present invention is shown; Figure 5 This paper presents a schematic diagram of the overall process of a cloud disk space planning method according to this application; Figure 6 This is a schematic diagram of the structure of a cloud disk space planning device provided in another embodiment of this application; Figure 7 This is a schematic diagram of the structure of an electronic device provided in another embodiment of this application. Detailed Implementation

[0014] The features and exemplary embodiments of various aspects of this application will be described in detail below. To make the objectives, technical solutions, and advantages of this application clearer, the application will be further described in detail below with reference to the accompanying drawings and specific embodiments. It should be understood that the specific embodiments described herein are only intended to explain this application and not to limit it. For those skilled in the art, this application can be implemented without some of these specific details. The following description of the embodiments is merely to provide a better understanding of this application by illustrating examples.

[0015] It should be noted that, in this document, relational terms such as "first" and "second" are used merely to distinguish one entity or operation from another, and do not necessarily require or imply any such actual relationship or order between these entities or operations. Furthermore, the terms "comprising," "including," or any other variations thereof are intended to cover non-exclusive inclusion, such that a process, method, article, or apparatus that comprises a list of elements includes not only those elements but also other elements not expressly listed, or elements inherent to such a process, method, article, or apparatus. Without further limitations, an element defined by the phrase "comprising..." does not exclude the presence of additional identical elements in the process, method, article, or apparatus that includes the element.

[0016] It should be noted that the acquisition, storage, use, and processing of data in this application embodiment all comply with the relevant provisions of national laws and regulations.

[0017] It should be noted that in the embodiments of this application, certain software, components, models and other existing solutions in the industry may be mentioned. These should be regarded as exemplary and are only intended to illustrate the feasibility of implementing the technical solution of this application. However, it does not mean that the applicant has used or necessarily used the solution.

[0018] First, let me explain the terms used in this application: Hierarchical Reinforcement Learning (HRL) is an improved algorithm for reinforcement learning. Its characteristic is that it decomposes complex tasks into multiple levels and performs decision optimization at each level. In this invention, HRL is used to solve the multi-level allocation and optimization problem of cloud disk space. By setting different decision levels (first cloud disk space planning model, second cloud disk space planning model, and third cloud disk space planning model), HRL can dynamically adjust the strategy at each level based on real-time feedback to improve the utilization efficiency of cloud disk space resources.

[0019] Cloud Disk Dataset: The cloud disk dataset refers to the cloud disk characteristic data collected from user terminals, including: user attribute data, cloud disk usage data, and storage resource data. The cloud disk dataset includes key information such as storage space utilization, file access frequency, and data type, and is the core data input of the automated cloud disk space management algorithm of this invention.

[0020] User Segmentation Strategy: The User Segmentation Strategy is a strategy built based on user cloud drive feature data. It is used to divide users into different groups or levels, which facilitates the system to allocate resources to different user groups in a personalized manner. The strategy plays an important role in classifying and optimizing user behavior in hierarchical reinforcement learning models.

[0021] Resource Expansion: Resource expansion refers to the system automatically expanding a user's cloud disk storage space according to the user's actual usage needs when the user's cloud disk storage space is close to a preset threshold, so as to ensure that the user can use the cloud disk service smoothly during peak periods. This invention uses a reinforcement learning model to dynamically predict expansion needs and automatically execute expansion operations.

[0022] Resource Reclamation: Resource reclamation refers to the process by which the system automatically reclaims a user's storage space and cleans up data when the user becomes inactive or leaves the cloud storage service, in order to ensure the continued efficient use of resources. The reclaimed storage resources will be redistributed to other active users to prevent resource waste.

[0023] Currently, cloud disk space planning is based on static resource allocation strategies. For example, operations and maintenance personnel allocate a fixed amount of cloud disk space to users based on their current storage needs, and then allocate new cloud disk space when the user's cloud disk space reaches a capacity threshold. However, existing technology cannot dynamically allocate cloud disk space to users in real time based on their storage needs, resulting in some users having excess cloud disk space while others have insufficient cloud disk space, ultimately leading to an uneven distribution of cloud disk space.

[0024] Based on this, this application provides a cloud disk space planning method. First, a first cloud disk space planning model is used to plan space for user groups, identifying the cloud disk space needs of different groups, achieving differentiated allocation, and reducing resource imbalance. Then, a second cloud disk space planning model is used to plan space for each user, improving the balance of space planning for each user. Finally, a third cloud disk space planning model, combined with users' historical cloud disk operation data, outputs a space allocation execution strategy for each user. Based on the trend information provided by the historical cloud disk operation data, the forward-looking nature of the space allocation execution strategy can be improved, reducing cloud disk space resource waste.

[0025] To address the problems of the prior art, embodiments of the present invention provide a cloud disk space planning method, apparatus, device, storage medium, and program product.

[0026] The cloud disk space planning method provided in the embodiments of the present invention will be introduced first below.

[0027] Figure 1 A flowchart illustrating a cloud disk space planning method according to an embodiment of the present invention is shown. Figure 1 As shown, the method may include the following steps: S101: Obtain cloud disk feature data for multiple users and the current unplanned space for multiple users.

[0028] In this embodiment, cloud disk feature data and current planned space for multiple users are acquired in real time. Each user corresponds to one set of cloud disk feature data. In one example, the cloud disk feature data includes: user attribute data, cloud disk usage data, and storage resource data. The user attribute data includes: user identifier, user level, user account creation time, user's most recent cloud disk access time, and user account status. The cloud disk usage data includes: storage space utilization rate, file access frequency, file type, and the storage space occupied by each file type. The storage resource data includes: available space on the storage server, used space on the storage server, Central Processing Unit (CPU) utilization rate, and memory utilization rate. In one example, the current planned space can be calculated based on the available space on the storage servers corresponding to multiple users.

[0029] In one example, after acquiring user cloud drive feature data in real time, the data undergoes preprocessing. This includes formatting the data, converting it from different sources into a standardized format, and denoising the data to remove outliers and noise. For instance, threshold-based filtering methods are used to remove outliers. Specifically, a reasonable threshold is set for the user cloud drive feature data; if the data exceeds this threshold, it is considered an outlier and removed. This preprocessing ensures that the data accurately reflects the actual situation of the user's cloud drive, providing a data foundation for subsequent cloud drive space planning.

[0030] S102: Input the current space to be planned and the cloud disk feature data corresponding to multiple users into the first cloud disk space planning model, and perform space planning for user groups through the first cloud disk space planning model to obtain the space to be planned for one or more user groups. The user groups are classified based on the cloud disk feature data corresponding to multiple users.

[0031] In this embodiment of the application, the current space to be planned and the cloud disk feature data corresponding to each user are input into the first cloud disk space planning model. The space planning is performed for the user group through the first cloud disk space planning model to obtain the space to be planned for each user group.

[0032] In one example, a user group refers to a set of users divided based on cloud disk feature data. By dividing users into different user groups using cloud disk feature data, users within the same user group have similar cloud disk feature data and cloud disk space requirements, thereby improving the accuracy and efficiency of cloud disk space allocation. Before the first cloud disk space planning model performs space planning for user groups, users can first be planned into multiple user groups based on the following formula (1), so that the first cloud disk space planning model can plan the space to be planned for each user group based on the current space to be planned and the cloud disk feature data corresponding to each user.

[0033] One specific implementation method for user group segmentation is as follows: continuously segment the user group until the segmented user group maximizes the value of formula (1), then the user group segmentation result at this point is the final user group: (1) in, This represents the generated user group; Indicates user i Cloud disk feature data; This indicates an indicator function used to mark the affiliation of users and user groups. i The cloud drive feature data belongs to a group When the user's value is 1, the value is 1. i The cloud drive characteristic data does not belong to the group. When the value is 0; Indicates user group The center vector; Indicates user group The covariance matrix; Represents a matrix with covariance The Mahalanobis distance represents the matching degree between user cloud disk feature data and user group center; Represents the regularization coefficient; Indicates user i The similarity weight between the cloud disk feature data of user j and the cloud disk feature data of user j; Indicates user i Cloud drive feature data and users j The similarity function between cloud disk feature data; i represents the user index number, used to uniquely identify a single user; j Indicates the user index number, and i Used in conjunction to represent another, different user; k The index number represents the user group and is used to uniquely identify the user group; N This indicates the total number of users; K This indicates the total number of users.

[0034] In another example, the cloud disk feature data corresponding to each user can be directly input into the first cloud disk space planning model, and user groups can be divided based on the first cloud disk space planning model.

[0035] S103: For each user group, input the space to be planned for that user group and the cloud disk feature data corresponding to each user in that user group into the second cloud disk space planning model. The second cloud disk space planning model is used to plan the space for each user and obtain the space to be planned for each user.

[0036] In this embodiment, the space to be planned for each user group and the cloud disk feature data corresponding to each user are input into the second cloud disk space planning model. The space to be planned for each user in each user group can be planned through the second cloud disk space planning model, and the space to be planned for each user can be obtained. The cloud disk space allocation can be further refined to meet individual needs.

[0037] S104: For each user, obtain the user's cloud disk historical operation data, input the user's cloud disk historical operation data, the user's space to be planned, and the user's corresponding cloud disk characteristic data into the third cloud disk space planning model, and generate the user's space allocation execution strategy through the third cloud disk space planning model.

[0038] In this embodiment, by inputting each user's historical cloud disk operation data, each user's planned space, and each user's corresponding cloud disk characteristic data into a third cloud disk space planning model, the space allocation execution strategy for each user is obtained through the third cloud disk space planning model. Based on the periodic patterns in the historical cloud disk operation data, it can help predict each user's future cloud disk space needs, providing a forward-looking basis for each user's space allocation execution strategy. For example, it can identify impending load surges, avoiding passively triggering expansion, or promptly reclaim fragmented space when a user transitions from an active to a less active or inactive state, reducing cloud disk space waste. Fragmented space refers to scattered, underutilized storage space generated by file deletion, modification, and other operations.

[0039] In one example, a user's cloud drive historical operation data includes, but is not limited to: cloud drive write / delete sequence, burst access peaks, and file popularity decay curves. Specifically, the cloud drive write / delete sequence refers to the chronological record of write and delete operations performed on cloud drive files within a statistical period; burst access peaks refer to the times and frequencies of access to files in the cloud drive exceeding a preset threshold; and the file popularity decay curve is the curve showing how the access frequency of a file changes over time from its creation to the current time.

[0040] In another example, the cloud disk write / delete sequence can be based on the cloud disk system's operation log interface to collect users' write / delete operation records at fixed intervals and plan them into a time-series sequence according to the period; sudden access peaks can be detected by real-time monitoring of users' access requests and data transfer volume to the cloud disk, and by setting a threshold, when the monitored file access frequency exceeds the threshold, it is marked as a sudden access peak, and the peak period and specific file access frequency value are recorded; the file popularity decay curve can be obtained by recording the creation time and access time of each file, counting the number of file accesses according to the time window, and obtaining the trend of popularity over time through curve fitting.

[0041] S150: Performs cloud disk space planning based on the space allocation strategy for each user.

[0042] In this embodiment, cloud disk space planning is performed based on a space allocation execution strategy for each user. This strategy includes, but is not limited to, cloud disk space allocation, delayed writes, hot / cold tiering, or fragmented space defragmentation. In one example, the planned space for each user is the calculated theoretical space allocation quota. The space allocation execution strategy derived from the third cloud disk space planning model transforms each user's planned space into executable actions. For instance, if a user has 20GB of planned space, based on their cloud disk historical operation data, it is known that 20GB of cold data (infrequently accessed files) exists within their cloud disk space. Therefore, by migrating this 20GB of cold data to low-speed storage, the freed-up 20GB of high-frequency storage remains within the user's planned space and is used to store new hot data (frequently accessed files). Thus, by implementing a hot / cold tiered space allocation strategy for this user, space planning for that user can be achieved.

[0043] In this embodiment, cloud disk feature data corresponding to multiple users and the current unplanned space shared by multiple users are obtained. Then, the current unplanned space and the cloud disk feature data corresponding to multiple users are input into a first cloud disk space planning model. The first cloud disk space planning model is used to plan space for user groups, resulting in unplanned spaces for one or more user groups. For each user group, the unplanned space of the user group and the cloud disk feature data corresponding to each user in the user group are input into a second cloud disk space planning model. The second cloud disk space planning model is used to plan space for each user, resulting in unplanned space for each user. Finally, for each user, the user's cloud disk historical operation data is obtained. The user's cloud disk historical operation data, the user's unplanned space, and the user's corresponding cloud disk feature data are input into a third cloud disk space planning model. The third cloud disk space planning model is used to generate the user's space allocation execution strategy. Finally, cloud disk space planning is performed based on each user's space allocation execution strategy. Compared to existing technologies, cloud disk space planning is inefficient when facing a large number of users and complex storage needs. This application improves the efficiency of cloud disk space planning by acquiring users' cloud disk characteristic data and the space to be planned in real time, and automatically generating space allocation execution strategies for each user through a first cloud disk space planning model, a second cloud disk space planning model, and a third cloud disk space planning model. Furthermore, the first cloud disk space planning model first plans space for user groups to identify the cloud disk space needs of different groups, achieving differentiated allocation among user groups and reducing resource imbalance. Planning space for user groups first allows for the formulation of a unified space allocation strategy for the common needs of the same group, avoiding the repetitive operation of planning space individually for each user from the beginning, further improving the efficiency of cloud disk space planning. Then, based on the second cloud disk space planning model, space planning is performed for each user, which can improve the balance of space planning for each user. Then, based on the third cloud disk space planning model and the historical operation data of each user's cloud disk, the space allocation execution strategy for each user is output. By using the trend information provided by the historical operation data of the cloud disk, the space allocation execution strategy is transformed from static passive allocation to active prediction, which improves the foresight and accuracy of the space allocation execution strategy and reduces the waste of cloud disk space resources.

[0044] Figure 2 A flowchart illustrating a cloud disk space planning method according to another embodiment of the present invention is shown. Figure 2 As shown above, in the above Figure 1 Based on the illustrated embodiment, the method further includes the following step before step S102: S201: Obtain the first training sample set, which includes multiple first training samples. Each first training sample includes: the current space to be planned, the cloud disk feature data sample corresponding to each user, and the space to be planned for each user group.

[0045] In this embodiment of the application, the first training sample set provides the correspondence between the current space to be planned, user cloud disk feature data, and the space to be planned of the user group for training the first cloud disk space planning model, and provides a data foundation for the first cloud disk space planning model to learn the group space planning rules.

[0046] S202: Based on the first training sample set, train the first cloud disk space planning model to be trained to obtain the trained first cloud disk space planning model.

[0047] In this embodiment of the application, the first cloud disk space planning model to be trained is trained based on the first training sample set. By using the correspondence between the current space to be planned, the user's cloud disk feature data and the space to be planned of the user group in the first training sample, the model continuously adjusts the parameters to obtain the trained first cloud disk space planning model.

[0048] S203: Obtain the second training sample set, which includes multiple second training samples. Each second training sample includes: the space to be planned for each user group, the cloud disk feature data sample corresponding to each user, and the space to be planned for each user.

[0049] In this embodiment, the second training sample set provides the correspondence between the user group's unplanned space, the user's corresponding cloud disk feature data, and the user's unplanned space for training the second cloud disk space planning model. It is the data foundation for the second cloud disk space planning model to learn the user space planning rules.

[0050] S204: Based on the second training sample set, train the second cloud disk space planning model to be trained to obtain the trained second cloud disk space planning model.

[0051] In this embodiment of the application, the second cloud disk space planning model to be trained is trained based on the second training sample set. By using the space to be planned for each user group in the second training sample, the cloud disk feature data corresponding to each user, and the correspondence between the space to be planned for each user, the model continuously adjusts the parameters to obtain the trained second cloud disk space planning model.

[0052] S205: Obtain the third training sample set, which includes multiple third training samples. Each third training sample includes: cloud disk historical operation data samples for each user, space samples to be planned for each user, cloud disk feature data samples corresponding to each user, and space allocation execution strategy samples for each user.

[0053] In this embodiment of the application, the third training sample set provides the corresponding relationship between each user's cloud disk historical operation data, each user's space to be planned, each user's corresponding cloud disk feature data, and each user's space allocation execution strategy for training the third cloud disk space planning model. It is the data foundation for the third cloud disk space planning model to learn each user's space allocation execution strategy.

[0054] S206: Based on the third training sample set, train the third cloud disk space planning model to be trained to obtain the trained third cloud disk space planning model.

[0055] In this embodiment of the application, the third cloud disk space planning model to be trained is trained based on the third training sample set. By using the correspondence between each user's cloud disk historical operation data, each user's space to be planned, each user's corresponding cloud disk feature data, and each user's space allocation execution strategy in the third training sample, the model continuously adjusts the parameters to obtain the trained third cloud disk space planning model.

[0056] In this embodiment, a first cloud disk space planning model, a second cloud disk space planning model, and a third cloud disk space planning model are trained using a first training sample set, a second training sample set, and a third training sample set, respectively. This enables the first, second, and third cloud disk space planning models to possess initial policy network parameters for executable cloud disk space planning, avoiding cloud disk space planning errors caused by incomplete initial policy network parameters in the early stages of online training. Furthermore, by initializing the first, second, and third cloud disk space planning models to possess initial policy network parameters, the speed at which cloud disk space allocation execution strategies are derived based on the first, second, and third cloud disk space planning models in the online stage can be improved. The online stage refers to the process of directly obtaining the cloud disk space allocation execution strategy using the first, second, and third cloud disk space planning models within the cloud disk space planning cycle.

[0057] To improve the accuracy and real-time performance of space allocation execution strategies, and to enhance the accuracy of space planning in the third-party cloud disk space planning model, a space allocation execution strategy for each user is generated through the third-party cloud disk space planning model, including: Within each planning cycle, acquire each user's historical cloud disk operation data, each user's unplanned space, and each user's corresponding cloud disk characteristic data.

[0058] In this embodiment, within each planning period, the historical operation data of each user's cloud disk, the space to be planned for each user, and the cloud disk characteristic data corresponding to each user are obtained. In one example, the planning period can be a set fixed time interval, that is, cloud disk space planning is performed once at a preset time interval.

[0059] Input each user's cloud disk historical operation data, each user's unplanned space, and each user's corresponding cloud disk characteristic data into the third cloud disk space planning model to obtain the initial space allocation execution strategy for each user.

[0060] In this embodiment, each user's historical cloud disk operation data, each user's unplanned space, and each user's corresponding cloud disk feature data are input into the third cloud disk space planning model to obtain each user's initial space allocation execution strategy. In one example, the third cloud disk space planning model is trained offline based on the third training sample set.

[0061] Based on the initial space allocation execution strategy for each user, update the cloud disk feature data corresponding to each user.

[0062] In this embodiment, based on the initial space allocation execution strategy obtained for each user, the cloud disk feature data corresponding to each user is updated to provide data for subsequent reward function calculation and ensure the authenticity of the space allocation execution strategy feedback.

[0063] Based on the updated cloud disk feature data corresponding to each user, the first reward function value is calculated. The first reward function value is obtained by the third-level reward function and the cross-level reward function. The third-level reward function is calculated based on the updated cloud disk feature data corresponding to each user, the current available storage space, and the total storage space. The cross-level reward function calculates the overall resource utilization rate, space planning cost, and service level achievement rate based on the updated cloud disk feature data. The service level achievement rate is the service level achievement rate of each user and / or each user group.

[0064] In this embodiment, a first reward function value is calculated based on the updated cloud disk feature data corresponding to each user. The first reward function value is obtained by calculating the third-layer reward function and the cross-layer reward function.

[0065] In one example, the formula for calculating the third-level reward function is: (2) in, Indicates the space allocation execution strategy; Indicates user i Priority weights; Indicates user i Storage space utilization; This represents the maximum storage space utilization rate among all users; Indicates the currently available storage space; This indicates the total storage space.

[0066] In one example, the formula for calculating user priority weight is: (3) in, Indicates user i Priority weights; Indicates user i Storage space utilization; Indicates user i File access frequency; Indicates user i The user's most recent cloud drive access time; Indicates user i CPU utilization; Indicates user i Memory usage; This represents the highest storage space utilization rate among all users; This indicates the highest file access frequency among all users; This indicates the time when the most recent user accessed the cloud drive among all users; This indicates the highest CPU utilization among all users; This indicates the highest memory usage among all users; This represents the weight parameter corresponding to the storage space utilization rate; The weight parameter represents the frequency of file access. This indicates the weight parameter corresponding to the user's most recent access time to the cloud drive; This represents the weighting parameter corresponding to CPU utilization. This represents the weighting parameter corresponding to memory usage.

[0067] in, , It can adaptively adjust in the training procedures of the first, second, and third cloud disk space planning models to balance multi-dimensional requirements such as storage space, access frequency, and computing power overhead. By dividing by the maximum value of different indicators, the values ​​of different indicators are limited to the range of 0 to 1, avoiding drastic fluctuations in gradient calculation due to excessive differences in the value range of different indicators, which would affect the convergence of the training procedures of the first, second, and third cloud disk space planning models.

[0068] In another example, the formula for calculating the cross-layer reward function is: (4) in, This represents the cross-layer reward function value; This represents the value of the global reward function; This represents the local reward function value; This represents the weighting coefficients of the local reward function.

[0069] In one example, It is calculated from the overall resource utilization rate and spatial planning cost. It is calculated based on the service level achievement rate of each user or each user group.

[0070] The overall resource utilization rate is calculated from the used space and available space of each user's storage server in the cloud disk feature data.

[0071] Space planning costs include: expansion costs, migration costs, and defragmentation costs. Expansion costs are calculated by the increase in available space on storage servers and the changes in CPU and memory usage. Migration costs are calculated by the size of the migrated files and the changes in CPU and memory usage. The size of the migrated files is calculated from the file types in the cloud disk feature data, the storage space occupied by each file type, and the used space on each storage server. Defragmentation costs can be calculated using CPU and memory usage, for example, by calculating the duration for which CPU or memory usage exceeds a preset threshold.

[0072] Service level achievement rate can be calculated by pre-setting service level achievement conditions for each user, such as storage space utilization being less than or equal to a preset threshold and account status being normal. The service level achievement rate for each user can be obtained by counting the total number of visits for each user and the number of visits that meet the service level achievement conditions, and then calculating the total number of visits for each user. Alternatively, within each user group, the service level achievement rate can be obtained by counting the number of people who have achieved the service level achievement rate and dividing that number by the total number of people in each user group.

[0073] The first preset training stopping condition is determined based on the value of the first reward function. The first preset training stopping condition is that the value of the first reward function is the minimum.

[0074] In this embodiment of the application, when the value of the first reward function is minimized, it indicates that the model parameters of the current third cloud disk space planning model have converged.

[0075] If the conditions are not met, the model parameters of the third cloud disk space planning model are adjusted using the first preset step size. The adjusted third cloud disk space planning model is then trained using each user's cloud disk historical operation data, each user's space to be planned, and each user's corresponding cloud disk feature data until the first preset training stop condition is met, thus obtaining the target third cloud disk space planning model and the space allocation execution strategy for each user.

[0076] In this embodiment, if the current first reward function value is not minimum, the model parameters of the third cloud disk space planning model are adjusted using a first preset step size. The first preset step size is larger than the second and third preset step sizes because a larger parameter adjustment step size allows the third cloud disk space planning model to respond more quickly to the real-time needs of individual users. Then, based on each user's historical cloud disk operation data, each user's unplanned space, and each user's corresponding cloud disk feature data, the adjusted third cloud disk space planning model is trained again until the first preset training stop condition is met, resulting in the target third cloud disk space planning model. Finally, based on the target third cloud disk space planning model, the space allocation execution strategy for each user is derived.

[0077] In this embodiment, by acquiring each user's historical cloud disk operation data, planned space, and corresponding cloud disk characteristic data for each user within each planning cycle, and deriving a space allocation execution strategy for each user based on the third cloud disk space planning model, it is possible to ensure that the derived space allocation execution strategy adapts to the user's real-time behavior in a timely manner. Specifically, the third-layer reward function minimizes storage resource waste for each user, while the cross-layer reward function considers overall resource utilization, cost, and service level achievement rate. This ensures that while minimizing storage resource waste for each user, it also guarantees the synergy of global resource optimization. It avoids the problem of excessively reducing storage resource waste for each user and lowering overall resource utilization by solely optimizing the model parameters of the third cloud disk space planning model through the third-layer reward function. By comprehensively optimizing the third-layer reward function and the cross-layer reward function, it is possible to improve overall resource utilization efficiency while controlling the overall cost of space planning and ensuring that the service level requirements of different users and groups are met.

[0078] To improve the accuracy of obtaining the planned space for each user and simultaneously enhance the planning accuracy of the second cloud disk space planning model, the planned space for each user is generated through the second cloud disk space planning model, including: After multiple planning cycles, obtain the planned space for each user group and the cloud disk feature data for each user within each planning cycle.

[0079] In this embodiment of the application, after multiple planning cycles, the space to be planned for each user group and the cloud disk feature data corresponding to each user are obtained within each planning cycle.

[0080] Input the planned space for each user group and the cloud disk feature data for each user into the second cloud disk space planning model within each planning cycle to obtain the initial planned space for each user.

[0081] In this embodiment, the planned space for each user group within each planning period and the cloud disk feature data for each user are input into the second cloud disk space planning model to obtain the initial planned space for each user. In one example, the planned space for each user group within each planning period and the cloud disk feature data for each user can be aggregated and then input into the second cloud disk space planning model to obtain the initial planned space for each user.

[0082] Obtain the updated cloud disk feature data for each user and calculate the second reward function value. The second reward function value is calculated by the second-level reward function and the cross-level reward function. The second-level reward function is calculated based on the updated cloud disk feature data for each user.

[0083] In this embodiment, updated cloud disk feature data for each user is obtained, and a second reward function value is calculated. The second reward function value is calculated from a second-level reward function and a cross-level reward function. In one example, the updated cloud disk feature data for each user is obtained based on the space allocation execution strategy derived from the third cloud disk space planning model within the latest planning period.

[0084] In one example, the formula for calculating the second-level reward function is: (5) in, This indicates the user's planned space allocation strategy; Indicates user group k Priority weights; Indicates user i Storage space utilization; This represents the average storage space utilization rate of user group k; Indicates user group k The variance of storage space utilization; Indicates user group k ; i Indicates user identifier; k Indicates user group identification; among which, through the Squaring the curve makes the second-layer reward function curve smoother, thus making the gradient more stable during parameter updates of the second cloud disk space planning model and ensuring the convergence of the training process; by using Normalization avoids amplifying the weight of user groups with large differences in storage space utilization in the second-level reward function, thus preventing them from excessively dominating the optimization direction of the second cloud disk space planning model.

[0085] In one example, the formula for calculating the priority weight of user groups is: (6) in, User group priority weight This represents the number of users included in the k-th user group; This represents the priority weight of user i; This indicates that user i belongs to the k-th user group.

[0086] This indicates an indicator function used to mark the affiliation of users and user groups. i The cloud drive feature data belongs to a group When the user's value is 1, the value is 1. i The cloud drive characteristic data does not belong to the group. At that time, the value is 0, because It only represents the static clustering results, while As time progresses, when users enter a low-activity or logout state, Set to 0 if necessary, otherwise keep it at 1. This means that within the same user group, it is necessary to further distinguish between users to be cleaned up and active users, and to treat users at different lifecycle stages within the user group differently.

[0087] In one example, the formula for calculating the average storage space utilization rate of a user group is: (7) in, This represents the average storage space utilization rate of user group k; Indicates the first k The number of users included in each user group; This represents the storage space utilization rate of user i; Indicates user i Belongs to the k A user group. In another example, Used in the calculation process This is based on data monitored in real time during each planning cycle, not on updated cloud disk feature data. .

[0088] In one example, the formula for calculating the variance of storage space utilization for a user group is: (8) in, This represents the variance of storage space utilization for user group k; This represents the number of users included in the k-th user group; This represents the storage space utilization rate of user i; This represents the average storage space utilization rate of user group k; This indicates that user i belongs to the k-th user group. In one example, Used in the calculation process This is based on data monitored in real time during each planning cycle, not on updated cloud disk feature data. .

[0089] The second preset training stopping condition is determined based on the value of the second reward function. The second preset training stopping condition is that the value of the second reward function is minimized.

[0090] In this embodiment of the application, when the value of the second reward function is minimized, it indicates that the model parameters of the current third cloud disk space planning model have converged.

[0091] If the conditions are not met, the model parameters of the second cloud disk space planning model are adjusted using the second preset step size, and the adjusted second cloud disk space planning model is trained again using the space to be planned for each user group and the cloud disk feature data corresponding to each user group until the second preset training stop condition is met, so as to obtain the target second cloud disk space planning model and the space to be planned for each user.

[0092] In this embodiment, if the current value of the second reward function is not the minimum, the model parameters of the second cloud disk space planning model are adjusted using a second preset step size. The second preset step size is smaller than the first preset step size but larger than the third preset step size. This is because, compared to the first cloud disk space planning model, the second cloud disk space planning model can adjust its strategy more flexibly to a certain extent, so the parameter adjustment step size can be larger than the third preset step size. Furthermore, it does not need to respond to individual user needs in real time, and therefore is smaller than the first preset step size. The adjusted second cloud disk space planning model is then trained again based on the planned space for each user group and the corresponding cloud disk feature data for each user group, until the second preset training stopping condition is met, resulting in the target second cloud disk space planning model. The planned space for each user is then derived based on the target second cloud disk space planning model.

[0093] In this embodiment, since the goal of the second cloud disk space planning model is to achieve balanced resource allocation within a user group and priority differences between groups, the second cloud disk space planning model does not rely on instantaneous data, but needs to be based on the cumulative trend of user group behavior. Therefore, by obtaining the space to be planned for each user group and the cloud disk feature data corresponding to each user in each planning period after multiple planning periods, and then training the second cloud disk space planning model, instantaneous noise caused by accidental fluctuations can be filtered out, thereby avoiding the problem of over-adjustment of the second cloud disk space planning model strategy due to frequent training of the second cloud disk space planning model caused by instantaneous noise. Furthermore, the second-layer reward function minimizes the dispersion after standard deviation normalization, ensuring that the same user group obtains a consistent amount of planning space. The user group priority ensures the priority difference between each user group. Then, the second reward function value is calculated together with the cross-layer reward function, avoiding the problem of local optimization being disconnected from the global goal due to optimizing the model parameters of the second cloud disk space planning model solely through the second-layer reward function. Therefore, by comprehensively optimizing the second-layer reward function and the cross-layer reward function, it is possible to ensure the consistency of resource allocation within the same user group while matching the priority difference between groups with the global resource planning goal. This maintains the coherence of strategies between different levels of models and achieves a balance between local and global optimization.

[0094] To improve the accuracy of obtaining the planned space for each user group and simultaneously enhance the planning accuracy of the first cloud disk space planning model, the planned space for each user group is generated through the first cloud disk space planning model, including: When the cloud disk feature data of multiple users is detected to meet the preset trigger conditions, the current space to be planned and the cloud disk feature data of each user are obtained for each planning cycle before the preset trigger conditions are met.

[0095] In this embodiment, when it is detected that the cloud disk feature data of multiple users meets a preset trigger condition, the current space to be planned and the cloud disk feature data of each user for each planning period before the preset trigger condition is met are obtained. In one example, the cloud disk feature data of multiple users meeting the preset trigger condition includes: a sudden change in the storage space utilization rate of multiple users, a sudden change in the available cloud disk space of multiple users, or an abnormal CPU utilization rate or memory utilization rate of multiple users.

[0096] Input the current unplanned space and cloud disk feature data corresponding to each user within each planning cycle into the first cloud disk space planning model to obtain the initial unplanned space for each user group.

[0097] In this embodiment, before the cloud disk feature data of multiple users meet the preset triggering conditions, the current space to be planned for each planning period and the cloud disk feature data corresponding to each user are input into the first cloud disk space planning model to obtain the initial space to be planned for each user group. In one example, the current space to be planned for each planning period and the cloud disk feature data corresponding to each user can be aggregated and then input into the first cloud disk space planning model to obtain the initial space to be planned for each user.

[0098] Obtain the updated cloud disk feature data for each user and calculate the third reward function value. The third reward function value is calculated by the first-level reward function and the cross-level reward function. The first-level reward function is calculated based on the updated cloud disk feature data for each user.

[0099] In this embodiment, updated cloud disk feature data for each user is obtained, and a third reward function value is calculated. The updated cloud disk feature data for each user is obtained by updating the space allocation execution strategy based on the third cloud disk space planning model in the latest planning cycle.

[0100] In one example, the formula for calculating the first-level reward function is: (9) in, This indicates the planned space allocation strategy for the user group; Indicates user group k Priority weights; Indicates user i Storage space utilization; This represents the maximum storage space utilization rate among all users; Indicates user i Priority weights; Indicates user i Belongs to the k A user group; k This indicates the user group identifier.

[0101] The third preset training stopping condition is determined based on the value of the third reward function. The third preset training stopping condition is that the value of the third reward function is the maximum.

[0102] In this embodiment of the application, when the value of the third reward function is maximized, it indicates that the model parameters of the current first cloud disk space planning model have converged.

[0103] If the conditions are not met, the model parameters of the first cloud disk space planning model are adjusted using the third preset step size, and the adjusted first cloud disk space planning model is trained again using the current space to be planned and the cloud disk feature data corresponding to each user until the third preset training stop condition is met, so as to obtain the target first cloud disk space planning model and the space to be planned for each user group.

[0104] In this embodiment, if the current value of the third reward function is not the maximum, the model parameters of the third cloud disk space planning model are adjusted using a third preset step size. This third preset step size is smaller than both the first and second preset step sizes. If the parameter adjustment step size is too large, it may cause severe oscillations in the global strategy, disrupting the synergy between strategies at different levels. The adjusted first cloud disk space planning model is then trained again based on the current space to be planned and the cloud disk feature data corresponding to each user, until the third preset training stopping condition is met, resulting in the target first cloud disk space planning model. Based on the target first cloud disk space planning model, the space to be planned for each user group is then derived.

[0105] In this embodiment, since the first cloud disk space planning model is responsible for the overall planning of cloud disk space, frequent adjustments in response to local or short-term fluctuations would cause repeated oscillations in the overall cloud disk space planning, thereby disrupting the synergy of strategies at all levels. Therefore, when the cloud disk feature data of multiple users meets the preset triggering conditions, i.e. when a significant change in the overall cloud disk space is detected, the current space to be planned for each planning cycle before the preset triggering conditions are met, as well as the cloud disk feature data corresponding to each user, are obtained for model training of the first cloud disk space planning model. This ensures the stability of the first cloud disk space planning model, allows for timely adaptation when the global state undergoes substantial changes, and provides a reliable macro framework for the second and third cloud disk space planning models, thereby maintaining the synergy of strategies at all levels. Furthermore, the first-layer reward function maximizes the storage space utilization of all users, prompting the first cloud disk space planning model to prioritize the allocation of space to users with high storage space utilization and high priority, thus avoiding global resource idleness or inefficient utilization. Then, the third reward function value is calculated together with the cross-layer reward function, avoiding the problem of excessively pursuing the maximization of global storage space utilization while ignoring the feasibility of implementing strategies in the middle and low layers, which is caused by only optimizing the model parameters of the first cloud disk space planning model through the first-layer reward function. In addition, the update of the first cloud disk space planning model depends on the cumulative advantage returned by the lower layers rather than the gradual gradient, avoiding the policy drift caused by training the higher layers first, then the middle layers, and finally the lower layers.

[0106] Figure 3 A flowchart illustrating a cloud disk space planning method according to another embodiment of the present invention is shown. Figure 3 As shown above, in the above Figure 1 Based on the illustrated embodiment, the method further includes the following after step S101: S301: Real-time monitoring of the cloud disk space usage status of users. When the cloud disk space usage status of a user meets the preset trigger conditions, the current space to be planned and multiple cloud disk feature data are input into the first cloud disk space planning model. The first cloud disk space planning model determines whether the current space to be planned meets the expansion requirements. Here, the cloud disk space usage status of a user meets the preset trigger conditions, which means that the cloud disk space usage status of any user meets the preset trigger conditions, and / or the cloud disk space usage status of any user group meets the preset trigger conditions.

[0107] In this embodiment, by monitoring the cloud disk space usage status of each user in real time, when the cloud disk space usage status meets the preset triggering conditions, the current space to be planned and multiple cloud disk feature data are input into the first cloud disk space planning model. The first cloud disk space planning model determines whether the current space to be planned meets the expansion requirements, that is, whether the cloud disk space usage status of the user does not meet the preset triggering conditions after the current space to be planned is additionally allocated to the user.

[0108] In one example, a user's cloud disk space usage status meeting the preset trigger condition means that the cloud disk space usage status of any user meets the preset trigger condition, and / or the cloud disk space usage status of any user group meets the preset trigger condition.

[0109] In another example, cloud disk space usage meeting preset trigger conditions refers to the storage space utilization rate of a user or user group. Reaching the storage space expansion trigger threshold .

[0110] S302: If the condition is not met, apply for additional space from the storage resource pool through the first cloud disk space planning model.

[0111] In this embodiment of the application, if it is determined that the current space to be planned cannot meet the user's current space needs, then additional space is requested from the storage resource pool through the first cloud disk space planning model.

[0112] S303: Input the newly added space, the current space to be planned, and multiple cloud disk feature data into the second cloud disk space planning model. The second cloud disk space planning model is used to plan space for each user group and obtain the space to be planned for each user group.

[0113] In this embodiment of the application, the newly added space, the current space to be planned, and multiple cloud disk feature data are input into the second cloud disk space planning model. The second cloud disk space planning model is used to plan space for each user group to obtain the space to be planned for each user group.

[0114] S304: Input the unplanned space and cloud disk characteristic data of each user group into the third cloud disk space planning model, and use the third cloud disk space planning model to plan space for each user to obtain the unplanned space for each user.

[0115] In this embodiment of the application, the planned space and cloud disk feature data of each user group are input into the third cloud disk space planning model, and the space planning is performed for each user through the third cloud disk space planning model to obtain the planned space for each user.

[0116] In one example, the formula for calculating the expansion demand strategy is: (10) in, This indicates a strategy to meet capacity expansion needs. Indicates user i Priority weights; Indicates user i Storage space utilization; Indicates user i The user's most recent cloud drive access time; This indicates the threshold for triggering storage space expansion; This indicates the time when the most recent user accessed the cloud drive; i represents the user identifier; N represents the total number of users.

[0117] In this embodiment, by monitoring the cloud disk space usage status of users in real time, the timeliness of expansion response is ensured, avoiding service interruptions caused by insufficient space. Then, when the user's cloud disk space usage status meets the preset triggering conditions, the current space to be planned and multiple cloud disk feature data are input into the first cloud disk space planning model. The first cloud disk space planning model determines whether the current space to be planned meets the expansion requirements. If it is determined that it does not meet the requirements, the first cloud disk space planning model applies for new space from the storage resource pool, and the new space, the current space to be planned, and multiple cloud disk feature data are input into the second cloud disk space planning model. Then, the second cloud disk space planning model performs space planning for each user group to obtain the space to be planned for each user group. Finally, the space to be planned for each user group and the cloud disk feature data are input into the third cloud disk space planning model, and the third cloud disk space planning model performs space planning for each user to obtain the space to be planned for each user. By dynamically adjusting the space allocation through the three-layer model, the situation of insufficient space or over-allocation can be effectively avoided.

[0118] Figure 4 A flowchart illustrating a cloud disk space planning method according to another embodiment of the present invention is shown. Figure 4 As shown above, in the above Figure 1 Based on the illustrated embodiment, the method further includes the following step before step S101: S401: Real-time acquisition of the last access time of each user to the cloud disk, calculation of the time difference based on the last access time to the cloud disk and the current time, and determination of users whose time difference is greater than the first preset threshold as inactive users.

[0119] In this embodiment, by monitoring users' cloud drive characteristic data in real time, it is possible to determine whether a user is inactive or has logged out of the cloud drive service. In one example, the user's most recent cloud drive access time can be obtained in real time, and then the time difference between the most recent cloud drive access time and the current time can be calculated to determine whether the user is inactive.

[0120] In one example, the formula for calculating the time difference is: (11) in, Indicates the time difference; Indicates the current time; This indicates the time when user i last accessed the cloud drive.

[0121] when When a user is determined to be inactive or has logged out of the cloud storage service, the system will determine that the user is either not active or has logged out. This represents the first preset threshold. In one example, it's also possible to determine whether a user has logged out of the cloud storage service by obtaining the user's account status.

[0122] S402: Obtain the file access frequency corresponding to inactive users, and delete files whose access frequency is less than the second preset threshold.

[0123] In this embodiment, the file access frequency corresponding to inactive users is obtained, and files with a file access frequency less than a second preset threshold are deleted.

[0124] (12) in, This indicates the file access frequency corresponding to user i; This indicates the second preset threshold.

[0125] S403: The cloud disk space freed up by deleting files will be incorporated into the available cloud disk storage space for future space planning.

[0126] In this embodiment of the application, the cloud disk space corresponding to the deleted file is reclaimed and included in the available cloud disk storage space for subsequent space planning.

[0127] In one example, the amount of cloud disk space reclaimed is calculated as follows: (13) in, Indicates the amount of space reclaimed by user i; This indicates the cloud disk space corresponding to the file currently deleted by user i; This indicates that storage space is reserved for user i under specific conditions for backup or recovery needs.

[0128] In this embodiment, by acquiring the most recent cloud disk access time of each user in real time, calculating the time difference based on the most recent cloud disk access time and the current time, and determining users whose time difference is greater than a first preset threshold as inactive users, dynamic monitoring of user activity status is achieved. Inactive users can be identified in a timely and accurate manner. Then, the file access frequency corresponding to inactive users is obtained, and files with access frequencies less than a second preset threshold are deleted. Redundant data is cleaned up in a targeted manner. After cleaning up the data, necessary backup cloud disk space is reserved for users. While reclaiming cloud disk space, the risk of data loss is reduced. Finally, the cloud disk space released by the deleted files is incorporated into the available cloud disk storage space for subsequent space planning. Therefore, the cloud disk space can maintain a high storage space utilization rate. Especially in large-scale user management scenarios, the automated reclamation mechanism can significantly reduce the idle rate of storage resources and ensure continuous optimization of resource allocation.

[0129] Figure 5 The diagram illustrates the overall process of a cloud disk space planning method according to this application, including: S501: data acquisition; S502: data preprocessing; S503: hierarchical reinforcement learning model initialization; S504: resource allocation and optimization; S505: dynamic adjustment strategy; S506: expansion mechanism triggering; S507: data migration; S508: user behavior detection; S509: resource reclamation and data cleaning.

[0130] Specifically, in step S501, by collecting cloud disk feature data corresponding to multiple users in real time and the current space to be planned, information on the dynamic changes of cloud disk space can be obtained in a timely manner, providing data for subsequent space planning.

[0131] Specifically, in step S502, by preprocessing the cloud disk feature data of multiple users and the current space to be planned, the data quality can be improved, making the subsequent model-based spatial planning judgment more accurate.

[0132] Specifically, in step S503, the hierarchical reinforcement learning model is initialized, that is, the first cloud disk space planning model is trained using the first training sample set, the second cloud disk space planning model is trained using the second training sample set, and the third cloud disk space planning model is trained using the third training sample set.

[0133] Specifically, in steps S504 and S505, the initialized hierarchical reinforcement learning model can allocate and optimize resources based on the collected data, thereby achieving a strategy of dynamically adjusting cloud disk space according to the user's real-time needs.

[0134] Specifically, in steps S506 and S507, the initialized hierarchical reinforcement learning model can determine whether to trigger the expansion mechanism based on the collected data, and after triggering the expansion mechanism, expand the user's cloud disk and then migrate the data to the expanded space.

[0135] Specifically, in steps S506 and S507, the initialized hierarchical reinforcement learning model can perform user behavior detection based on the collected data to determine whether the user is an inactive user or has exited the cloud disk service, and then reclaim or clean up the cloud disk data of inactive users.

[0136] Based on the cloud disk space planning method provided in the above embodiments, this application also provides a specific implementation of a cloud disk space device. Please refer to the following embodiments.

[0137] First see Figure 6 The cloud storage space planning device 600 provided in this application embodiment includes: The acquisition module 601 is used to acquire cloud disk feature data corresponding to multiple users and the current space to be planned that is shared by multiple users. The first processing module 602 is used to input the current space to be planned and the cloud disk feature data corresponding to multiple users into the first cloud disk space planning model, and to perform space planning for user groups through the first cloud disk space planning model to obtain the space to be planned for one or more user groups, wherein the user groups are classified based on the cloud disk feature data corresponding to multiple users. The second processing module 603 is used to input the space to be planned for each user group and the cloud disk feature data corresponding to each user in the user group into the second cloud disk space planning model, and to perform space planning for each user through the second cloud disk space planning model to obtain the space to be planned for each user. The third processing module 604 is used to obtain the user's cloud disk historical operation data for each user, input the user's cloud disk historical operation data, the user's space to be planned, and the user's corresponding cloud disk feature data into the third cloud disk space planning model, and generate the user's space allocation execution strategy through the third cloud disk space planning model. Planning module 605 is used to plan cloud disk space based on the space allocation strategy for each user.

[0138] In one example, the cloud disk space planning device 600 also includes: The first acquisition module is used to acquire the first training sample set, which includes multiple first training samples. Each first training sample includes: the current space to be planned, the cloud disk feature data sample corresponding to each user, and the space to be planned for each user group. The first training module is used to train the first cloud disk space planning model to be trained based on the first training sample set, so as to obtain the trained first cloud disk space planning model. The second acquisition module is used for the second training sample set, which includes multiple second training samples. Each second training sample includes: a space sample to be planned for each user group, a cloud disk feature data sample corresponding to each user, and a space sample to be planned for each user. The second training module is used to train the second cloud disk space planning model to be trained based on the second training sample set, so as to obtain the trained second cloud disk space planning model. The third acquisition module is used to acquire the third training sample set, which includes multiple third training samples. Each third training sample includes: cloud disk historical operation data samples for each user, space samples to be planned for each user, cloud disk feature data samples corresponding to each user, and space allocation execution strategy samples for each user. The third training module is used to train the third cloud disk space planning model to be trained based on the third training sample set, so as to obtain the trained third cloud disk space planning model.

[0139] In one embodiment, the third processing module 604 includes: The first acquisition submodule is used to acquire each user's cloud disk historical operation data, each user's unplanned space, and each user's corresponding cloud disk feature data within each planning cycle; The first input submodule is used to input each user's cloud disk historical operation data, each user's space to be planned, and each user's corresponding cloud disk feature data into the third cloud disk space planning model to obtain each user's initial space allocation execution strategy. The first update submodule is used to update the cloud disk feature data corresponding to each user based on the initial space allocation execution strategy for each user. The first calculation submodule is used to calculate the first reward function value based on the updated cloud disk feature data corresponding to each user. The first reward function value is calculated by the third-level reward function and the cross-level reward function. The third-level reward function is calculated based on the updated cloud disk feature data corresponding to each user, the current available storage space, and the total storage space. The cross-level reward function calculates the overall resource utilization rate, space planning cost, and service level achievement rate based on the updated cloud disk feature data. The service level achievement rate is the service level achievement rate of each user and / or each user group. The first judgment submodule is used to determine whether the first preset training stopping condition is met based on the first reward function value. The first preset training stopping condition is that the first reward function value is the minimum. The first adjustment submodule is used to adjust the model parameters of the third cloud disk space planning model by adopting a first preset step size when the conditions are not met. It then uses each user's cloud disk historical operation data, each user's space to be planned, and each user's corresponding cloud disk feature data to continue training the adjusted third cloud disk space planning model until the first preset training stop condition is met, thereby obtaining the target third cloud disk space planning model and the space allocation execution strategy for each user.

[0140] In one embodiment, the second processing module 603 includes: The second acquisition submodule is used to acquire the planned space for each user group and the cloud disk feature data for each user in each planning period after multiple planning periods. The second input submodule is used to input the space to be planned for each user group in each planning period and the cloud disk feature data corresponding to each user into the second cloud disk space planning model to obtain the initial space to be planned for each user. The second calculation submodule is used to obtain the updated cloud disk feature data corresponding to each user and calculate the second reward function value. The second reward function value is calculated by the second-level reward function and the cross-level reward function. The second-level reward function is calculated based on the updated cloud disk feature data corresponding to each user. The second judgment submodule is used to determine whether the second preset training stopping condition is met based on the second reward function value. The second preset training stopping condition is that the second reward function value is the minimum. The second adjustment submodule is used to adjust the model parameters of the second cloud disk space planning model by adopting the second preset step size when the conditions are not met. It then uses the space to be planned for each user group and the cloud disk feature data corresponding to each user in each planning cycle to continue training the adjusted second cloud disk space planning model until the second preset training stop condition is met, thereby obtaining the target second cloud disk space planning model and the space to be planned for each user.

[0141] In one embodiment, the first processing module 602 includes: The third acquisition submodule is used to acquire the current space to be planned and the cloud disk feature data of each user in each planning cycle before the preset trigger conditions are met when the cloud disk feature data of multiple users are detected to meet the preset trigger conditions. The third input submodule is used to input the current space to be planned in each planning cycle and the cloud disk feature data corresponding to each user into the first cloud disk space planning model to obtain the initial space to be planned for each user group. The third calculation submodule is used to obtain the updated cloud disk feature data corresponding to each user and calculate the third reward function value. The third reward function value is calculated by the first-level reward function and the cross-level reward function. The first-level reward function is calculated based on the updated cloud disk feature data corresponding to each user. The third judgment submodule is used to determine whether the third preset training stopping condition is met based on the value of the third reward function. The third preset training stopping condition is that the value of the third reward function is the maximum. The third adjustment submodule is used to adjust the model parameters of the first cloud disk space planning model by adopting a third preset step size when the conditions are not met. It then uses the current space to be planned and the cloud disk feature data corresponding to each user in each planning cycle to continue training the adjusted first cloud disk space planning model until the third preset training stop condition is met, thereby obtaining the target first cloud disk space planning model and the space to be planned for each user group.

[0142] In one example, the cloud disk space planning device 600 also includes: The monitoring module is used to monitor the cloud disk space usage status of users in real time. When the user's cloud disk space usage status meets the preset trigger conditions, the current space to be planned and multiple cloud disk feature data are input into the first cloud disk space planning model. The first cloud disk space planning model determines whether the current space to be planned meets the expansion requirements. The user's cloud disk space usage status meets the preset trigger conditions, which means that the cloud disk space usage status of any user meets the preset trigger conditions, and / or the cloud disk space usage status of any user group meets the preset trigger conditions. The judgment module is used to determine whether the conditions are met. If not, it applies for additional space from the storage resource pool through the first cloud disk space planning model.

[0143] The first input module is used to input the newly added space, the current space to be planned, and multiple cloud disk feature data into the second cloud disk space planning model. The second cloud disk space planning model performs space planning for each user group to obtain the space to be planned for each user group.

[0144] The second input module is used to input the space to be planned and cloud disk feature data of each user group into the third cloud disk space planning model. The third cloud disk space planning model performs space planning for each user to obtain the space to be planned for each user.

[0145] In one example, the cloud disk space planning device 600 also includes: The second acquisition module is used to acquire the time of each user's most recent access to the cloud disk in real time, calculate the time difference based on the time of the most recent access to the cloud disk and the current time, and determine users whose time difference is greater than the first preset threshold as inactive users. The third acquisition module is used to acquire the file access frequency of inactive users and delete files whose access frequency is less than the second preset threshold. The fourth processing module is used to incorporate the cloud disk space freed up by deleted files into the available cloud disk storage space for subsequent space planning.

[0146] The various modules of the cloud disk space planning device provided in this application embodiment can achieve Figure 1 This paper presents the functions of each step in a cloud disk space planning method, and demonstrates the corresponding technical effects. For the sake of brevity, these details will not be elaborated here.

[0147] Figure 7 This is a schematic diagram of the hardware structure of the electronic device provided in the embodiments of this application.

[0148] An electronic device may include a processor 701 and a memory 702 storing computer program instructions.

[0149] Specifically, the processor 701 may include a central processing unit (CPU), an application specific integrated circuit (ASIC), or one or more integrated circuits that can be configured to implement the embodiments of this application.

[0150] Memory 702 may include mass storage for data or instructions. For example, and not limitingly, memory 702 may include a hard disk drive (HDD), a floppy disk drive, flash memory, optical disk, magneto-optical disk, magnetic tape, or a Universal Serial Bus (USB) drive, or a combination of two or more of these. In one instance, memory 702 may include removable or non-removable (or fixed) media, or memory 702 may be a non-volatile solid-state memory.

[0151] In one instance, memory 702 may be read-only memory (ROM). In one instance, the ROM may be a mask-programmed ROM, a programmable ROM (PROM), an erasable PROM (EPROM), an electrically erasable PROM (EEPROM), an electrically rewritable ROM (EAROM), or flash memory, or a combination of two or more of these.

[0152] Memory 702 may include read-only memory (ROM), random access memory (RAM), disk storage media device, optical storage media device, flash memory device, electrical, optical, or other physical / tangible memory storage device. Therefore, generally, memory includes one or more tangible (non-transitory) computer-readable storage media (e.g., memory devices) encoded with software including computer-executable instructions, and when the software is executed (e.g., by one or more processors), it is operable to perform the operations described with reference to the method according to one aspect of this disclosure.

[0153] The processor 701 reads and executes computer program instructions stored in the memory 702 to implement a computer room temperature control method in the above embodiment.

[0154] In one example, the electronic device may also include a communication interface 703 and a bus 704. Wherein, as... Figure 7 As shown, the processor 701, memory 702, and communication interface 703 are connected through bus 704 and complete communication with each other.

[0155] The communication interface 703 is mainly used to realize communication between various modules, devices, units and / or equipment in the embodiments of this application.

[0156] Bus 704 includes hardware, software, or both, that couples components of an online data traffic metering device together. For example, and not limitingly, the bus may include an Accelerated Graphics Port (AGP) or other graphics bus, an Extended Industry Standard Architecture (EISA) bus, a Front Side Bus (FSB), a Hyper Transport (HT) interconnect, an Industry Standard Architecture (ISA) bus, an Infinite Bandwidth Interconnect, a Low Pin Count (LPC) bus, a memory bus, a Microchannel Architecture (MCA) bus, a Peripheral Component Interconnect (PCI) bus, a PCI-Express (PCI-X) bus, a Serial Advanced Technology Attachment (SATA) bus, a Video Electronics Standards Association Local (VLB) bus, or other suitable buses, or combinations of two or more of these. Where appropriate, bus 704 may include one or more buses. Although specific buses are described and illustrated in embodiments of this application, this application contemplates any suitable bus or interconnect.

[0157] In addition, in conjunction with the cloud disk space planning method in the above embodiments, this application embodiment can provide a computer storage medium for implementation. The computer storage medium stores computer program instructions; when these computer program instructions are executed by a processor, they implement any of the cloud disk space planning methods in the above embodiments.

[0158] This application also provides a computer program product, including a computer program, which, when executed, implements any of the cloud disk space planning methods described in the above embodiments.

[0159] It should be clarified that this application is not limited to the specific configurations and processes described above and shown in the figures. For the sake of brevity, detailed descriptions of known methods are omitted here. In the above embodiments, several specific steps are described and shown as examples. However, the method process of this application is not limited to the specific steps described and shown. Those skilled in the art can make various changes, modifications, and additions, or change the order of steps, after understanding the spirit of this application.

[0160] The functional blocks shown in the above block diagram can be implemented as hardware, software, firmware, or a combination thereof. When implemented in hardware, they can be, for example, electronic circuits, application-specific integrated circuits (ASICs), appropriate firmware, plug-ins, function cards, etc. When implemented in software, the elements of this application are programs or code segments used to perform the required tasks. Programs or code segments can be stored on a machine-readable medium or transmitted over a transmission medium or communication link via data signals carried on a carrier wave. "Machine-readable medium" can include any medium capable of storing or transmitting information. Examples of machine-readable media include electronic circuits, semiconductor memory devices, read-only memory (ROM), flash memory, erasable read-only memory (EROM), floppy disks, compact disc read-only memory (CD-ROM), optical disks, hard disks, fiber optic media, radio frequency (RF) links, etc. Code segments can be downloaded via computer networks such as the Internet, intranets, etc.

[0161] It should also be noted that the exemplary embodiments mentioned in this application describe methods or systems based on a series of steps or apparatus. However, this application is not limited to the order of the above steps; that is, the steps can be performed in the order mentioned in the embodiments, or in a different order, or several steps can be performed simultaneously.

[0162] The aspects of this disclosure have been described above with reference to flowchart illustrations and / or block diagrams of methods, apparatus (systems), and computer program products according to embodiments of this disclosure. It should be understood that each block in the flowchart illustrations and / or block diagrams, and combinations of blocks in the flowchart illustrations and / or block diagrams, can be implemented by computer program instructions. These computer program instructions can be provided to a processor of a general-purpose computer, a special-purpose computer, or other programmable data processing apparatus to produce a machine such that these instructions, executable via the processor of the computer or other programmable data processing apparatus, enable the implementation of the functions / actions specified in one or more blocks of the flowchart illustrations and / or block diagrams. Such a processor can be, but is not limited to, a general-purpose processor, a special-purpose processor, a special application processor, or a field-programmable logic circuit. It is also understood that each block in the block diagrams and / or flowcharts, and combinations of blocks in the block diagrams and / or flowcharts, can also be implemented by special-purpose hardware performing the specified functions or actions, or can be implemented by a combination of special-purpose hardware and computer instructions.

[0163] The above are merely specific embodiments of this application. Those skilled in the art will clearly understand that, for the sake of convenience and brevity, the specific working processes of the systems, modules, and units described above can be referred to the corresponding processes in the foregoing method embodiments, and will not be repeated here. It should be understood that the protection scope of this application is not limited thereto. Any person skilled in the art can easily conceive of various equivalent modifications or substitutions within the technical scope disclosed in this application, and these modifications or substitutions should all be covered within the protection scope of this application.

Claims

1. A cloud disk space planning method, characterized in that, The method comprises the following steps: obtaining cloud disk feature data corresponding to a plurality of users respectively and current space to be planned corresponding to the plurality of users collectively; inputting the current space to be planned and the cloud disk feature data corresponding to the plurality of users respectively into a first cloud disk space planning model, performing space planning for a user group through the first cloud disk space planning model, and obtaining space to be planned of one or more user groups, wherein the user group is obtained based on classification of the cloud disk feature data corresponding to the plurality of users respectively; for each user group, inputting the space to be planned of the user group and the cloud disk feature data corresponding to each user in the user group into a second cloud disk space planning model, performing space planning for each user through the second cloud disk space planning model, and obtaining space to be planned of each user; for each user, obtaining cloud disk historical operation data of the user, inputting the cloud disk historical operation data of the user, the space to be planned of the user and the cloud disk feature data corresponding to the user into a third cloud disk space planning model, and generating a space allocation execution strategy of the user through the third cloud disk space planning model; performing cloud disk space planning based on the space allocation execution strategy of each user.

2. The method of claim 1, wherein, Before the step of inputting the current space to be planned and the cloud disk feature data corresponding to the plurality of users respectively into a first cloud disk space planning model, performing space planning for a user group through the first cloud disk space planning model, and obtaining space to be planned of one or more user groups, the method further comprises the following steps: obtaining a first training sample set, wherein the first training sample set comprises a plurality of first training samples, and each first training sample comprises a current space to be planned sample, cloud disk feature data samples corresponding to each user, and space to be planned samples of each user group; based on the first training sample set, performing model training on a first cloud disk space planning model to be trained to obtain a trained first cloud disk space planning model; obtaining a second training sample set, wherein the second training sample set comprises a plurality of second training samples, and each second training sample comprises space to be planned samples of each user group, cloud disk feature data samples corresponding to each user, and space to be planned samples of each user; based on the second training sample set, performing model training on a second cloud disk space planning model to be trained to obtain a trained second cloud disk space planning model; obtaining a third training sample set, wherein the third training sample set comprises a plurality of third training samples, and each third training sample comprises cloud disk historical operation data samples of each user, space to be planned samples of each user, cloud disk feature data samples corresponding to each user, and space allocation execution strategy samples of each user; based on the third training sample set, performing model training on a third cloud disk space planning model to be trained to obtain a trained third cloud disk space planning model.

3. The method of claim 1, wherein, The cloud disk space planning method comprises the following steps: In each planning period, the cloud disk historical operation data of each user, the space to be planned of each user, and the cloud disk characteristic data corresponding to each user are obtained; The cloud disk historical operation data of each user, the space to be planned of each user, and the cloud disk characteristic data corresponding to each user are input into the third cloud disk space planning model to obtain an initial space allocation execution strategy of each user; Based on the initial space allocation execution strategy of each user, the cloud disk characteristic data corresponding to each user is updated; Based on the updated cloud disk characteristic data corresponding to each user, a first reward function value is calculated, wherein the first reward function value is calculated by a third layer reward function and a cross-layer reward function, the third layer reward function is calculated based on the updated cloud disk characteristic data corresponding to each user, the current available storage space, and the total storage space, and the cross-layer reward function is used to calculate the overall resource utilization, the space planning cost, and the service level achievement rate based on the updated cloud disk characteristic data, wherein the service level achievement rate is the service level achievement rate of each user and / or each user group; It is determined whether the first preset training stop condition is met according to the first reward function value, wherein the first preset training stop condition is that the first reward function value is minimum; If not, the model parameters of the third cloud disk space planning model are adjusted by using a first preset step, and the cloud disk historical operation data of each user, the space to be planned of each user, and the cloud disk characteristic data corresponding to each user are used to continue training the adjusted third cloud disk space planning model until the first preset training stop condition is met, so as to obtain a target third cloud disk space planning model and the space allocation execution strategy of each user.

4. The method of claim 3, wherein, The cloud disk space planning method comprises the following steps: After a plurality of planning periods, the space to be planned of each user group in each planning period and the cloud disk characteristic data corresponding to each user are obtained; The space to be planned of each user group in each planning period and the cloud disk characteristic data corresponding to each user are input into the second cloud disk space planning model to obtain the initial space to be planned of each user; The updated cloud disk characteristic data corresponding to each user is obtained, and a second reward function value is calculated, wherein the second reward function value is calculated by a second layer reward function and the cross-layer reward function, and the second layer reward function is calculated based on the updated cloud disk characteristic data corresponding to each user; determining whether a second preset training stop condition is met according to the second reward function value, the second preset training stop condition being that the second reward function value is minimum; if the second preset training stop condition is not met, adjusting model parameters of the second cloud disk space planning model by using a second preset step size, and continuing to train the adjusted second cloud disk space planning model by using the to-be-planned space of each user group in each planning period and the cloud disk feature data corresponding to each user, until the second preset training stop condition is met, to obtain a target second cloud disk space planning model and the to-be-planned space of each user.

5. The method of claim 4, wherein, the current to-be-planned space and the cloud disk feature data corresponding to the plurality of users are input into the first cloud disk space planning model, and space planning is performed for the user groups by using the first cloud disk space planning model to obtain the to-be-planned space of one or more user groups, including: when it is detected that the cloud disk feature data of the plurality of users meets a preset triggering condition, obtaining the current to-be-planned space in each planning period and the cloud disk feature data corresponding to each user before the preset triggering condition is met; the current to-be-planned space in each planning period and the cloud disk feature data corresponding to each user are input into the first cloud disk space planning model to obtain the initial to-be-planned space of each user group; the updated cloud disk feature data corresponding to each user is obtained, and a third reward function value is calculated, the third reward function value being calculated by a first layer reward function and the cross-layer reward function, wherein the first layer reward function is calculated based on the updated cloud disk feature data corresponding to each user; determining whether a third preset training stop condition is met according to the third reward function value, the third preset training stop condition being that the third reward function value is maximum; if the third preset training stop condition is not met, adjusting model parameters of the first cloud disk space planning model by using a third preset step size, and continuing to train the adjusted first cloud disk space planning model by using the current to-be-planned space in each planning period and the cloud disk feature data corresponding to each user, until the third preset training stop condition is met, to obtain a target first cloud disk space planning model and the to-be-planned space of the one or more user groups.

6. The method of claim 1, wherein, after the plurality of users respectively corresponding to the cloud disk feature data and the current to-be-planned space corresponding to the plurality of users are obtained, the method further includes: real-time monitoring of a cloud disk space usage state of a user, when the cloud disk space usage state of the user meets a preset triggering condition, inputting the current to-be-planned space and the plurality of cloud disk feature data into the first cloud disk space planning model, and determining whether the current to-be-planned space meets an expansion requirement by using the first cloud disk space planning model, wherein the cloud disk space usage state of the user meeting the preset triggering condition means that the cloud disk space usage state of any user meets the preset triggering condition, and / or the cloud disk space usage state of any user group meets the preset triggering condition; if it is determined that the current to-be-planned space does not meet the expansion requirement, applying for new space from a storage resource pool by using the first cloud disk space planning model; input the new space, the current space to be planned, and the plurality of cloud disk feature data into the second cloud disk space planning model, perform space planning for each user group through the second cloud disk space planning model, and obtain the space to be planned for each user group; input the space to be planned for each user group and the cloud disk feature data into a third cloud disk space planning model, perform space planning for each user through the third cloud disk space planning model, and obtain the space to be planned for each user.

7. The method of claim 1, wherein, Before the cloud disk feature data corresponding to a plurality of users respectively and the current space to be planned corresponding to the plurality of users are obtained, the method further includes: obtain the time difference value based on the latest cloud disk access time and the current time, determine a user whose time difference value is greater than a first preset threshold as an inactive user, and obtain the file access frequency corresponding to the inactive user; delete a file with an access frequency less than a second preset threshold, and release the cloud disk space of the deleted file into the available cloud disk storage space for subsequent space planning. The cloud disk space planning device includes:

8. A cloud disk space planning apparatus, characterized by, a obtaining module configured to obtain cloud disk feature data corresponding to a plurality of users respectively and a current space to be planned corresponding to the plurality of users; a first processing module configured to input the current space to be planned and the cloud disk feature data corresponding to the plurality of users respectively into a first cloud disk space planning model, perform space planning for a user group through the first cloud disk space planning model, and obtain the space to be planned for one or more user groups, wherein the user group is obtained based on classification of the cloud disk feature data corresponding to the plurality of users respectively; a second processing module configured to, for each user group, input the space to be planned for the user group and the cloud disk feature data corresponding to each user in the user group into a second cloud disk space planning model, perform space planning for each user through the second cloud disk space planning model, and obtain the space to be planned for each user; a third processing module configured to, for each user, obtain cloud disk historical operation data of the user, input the cloud disk historical operation data of the user, the space to be planned for the user, and the cloud disk feature data corresponding to the user into a third cloud disk space planning model, and generate a space allocation execution strategy for the user through the third cloud disk space planning model; a planning module configured to perform cloud disk space planning based on the space allocation execution strategy for each user. The device includes:

9. An electronic device, comprising: a processor and a memory storing computer program instructions; the processor executes the computer program instructions to implement the cloud disk space planning method of any one of claims 1-7. The computer program instructions are stored on the computer readable storage medium, and the computer program instructions are executed by the processor to implement the cloud disk space planning method of any one of claims 1-7.

10. A computer-readable storage medium, characterized in that, The instructions in the computer program product are executed by the processor of the electronic device, so that the electronic device can perform the cloud disk space planning method of any one of claims 1-7.

11. A computer program product, characterised in that, ​

Images