Resource scheduling method and device based on elastic block storage service and storage medium

Abstract

Embodiments of the invention provide a resource scheduling method and device based on an elastic block storage service, and a storage medium. The method comprises the steps of generating a candidate cloud disk behavior tag set of a target tenant according to attribute information and operation event information of each cloud disk of the target tenant in a tenant set; according to cloud disk behavior changes of the target tenant and other tenants in the tenant set, determining a current weight of each candidate cloud disk behavior tag in the candidate cloud disk behavior tag set, and according to the current weight of each candidate cloud disk behavior tag, determining a target cloud disk behavior tag of the target tenant from the candidate cloud disk behavior tag set; generating tagged feature information of the target tenant according to the target cloud disk behavior tag; and scheduling the cloud disk of the target tenant according to the tagged feature information of the target tenant. And the multi-dimensional target cloud disk behavior tag of the target tenant is accurately determined to generate tenant tagged feature information for resource scheduling of the elastic block storage service, so that the resource scheduling capability is improved.

Description

Technical Field

[0001] Embodiments of the present disclosure relate to the technical field of computer and network communication, and particularly to a resource scheduling method, device, and storage medium based on an elastic block storage service. Background Art

[0002] In the scenario of the elastic block storage service provided by traditional cloud providers, resource scheduling pays more attention to the load model at the cluster level, and performs load management of individual resources based on the overall load of the cluster and the real-time load of resources to perform resource scheduling. However, the resource scheduling ability of traditional elastic block storage services for specific tenants needs to be further improved. Summary of the Invention

[0003] Embodiments of the present disclosure provide a resource scheduling method, device, and storage medium based on an elastic block storage service to improve the resource scheduling ability of the elastic block storage service.

[0004] In a first aspect, embodiments of the present disclosure provide a resource scheduling method based on an elastic block storage service, including:

[0005] Obtaining attribute information and operation event information of each cloud disk of a target tenant in a tenant set;

[0006] Generating a candidate cloud disk behavior label set of the target tenant according to the attribute information and operation event information of each cloud disk;

[0007] Determining the current weight of each candidate cloud disk behavior label in the candidate cloud disk behavior label set according to the cloud disk behavior changes of the target tenant and the other tenants in the tenant set, and determining the target cloud disk behavior label of the target tenant from the candidate cloud disk behavior label set according to the current weight of each candidate cloud disk behavior label; generating labeled feature information of the target tenant according to the target cloud disk behavior label;

[0008] Scheduling the cloud disks of the target tenant according to the labeled feature information of the target tenant.

[0009] In a second aspect, embodiments of the present disclosure provide a resource scheduling device based on an elastic block storage service, including:

[0010] An obtaining unit, configured to obtain attribute information and operation event information of each cloud disk of a target tenant in a tenant set;

[0011] A label generation unit, configured to generate a set of candidate cloud disk behavior labels for the target tenant according to the attribute information and operation event information of each cloud disk; determine the current weight of each candidate cloud disk behavior label in the set of candidate cloud disk behavior labels according to the cloud disk behavior changes of the target tenant and the remaining tenants in the tenant set, and determine the target cloud disk behavior label of the target tenant from the set of candidate cloud disk behavior labels according to the current weight of each candidate cloud disk behavior label;

[0012] A feature information generation unit, configured to generate labeled feature information for the target tenant according to the target cloud disk behavior label;

[0013] A scheduling unit, configured to schedule the cloud disks of the target tenant according to the labeled feature information of the target tenant.

[0014] In a third aspect, an embodiment of the present disclosure provides an electronic device, including: a processor and a memory;

[0015] The memory stores computer-executable instructions;

[0016] The processor executes the computer-executable instructions stored in the memory, so that the at least one processor executes the resource scheduling method based on the elastic block storage service described in the first aspect and various possible designs of the first aspect above.

[0017] In a fourth aspect, an embodiment of the present disclosure provides a computer-readable storage medium, in which computer-executable instructions are stored, and when the processor executes the computer-executable instructions, the resource scheduling method based on the elastic block storage service described in the first aspect and various possible designs of the first aspect above is implemented.

[0018] In a fifth aspect, an embodiment of the present disclosure provides a computer program product, including a computer program, and when the computer program is executed by a processor, the resource scheduling method based on the elastic block storage service described in the first aspect and various possible designs of the first aspect above is implemented.

[0019] The resource scheduling method, device, and storage medium based on elastic block storage service provided by the embodiments of the present disclosure obtain the attribute information and operation event information of each cloud disk of a target tenant in a tenant set; generate a candidate cloud disk behavior label set of the target tenant according to the attribute information and operation event information of each cloud disk; determine the current weight of each candidate cloud disk behavior label in the candidate cloud disk behavior label set according to the cloud disk behavior changes of the target tenant and the other tenants in the tenant set, and determine the target cloud disk behavior label of the target tenant from the candidate cloud disk behavior label set according to the current weight of each candidate cloud disk behavior label; generate labeled feature information of the target tenant according to the target cloud disk behavior label; and schedule the cloud disks of the target tenant according to the labeled feature information of the target tenant. In the embodiments of the present disclosure, through the statistics and mining of the attribute information and operation event information of each cloud disk of the target tenant, multi-dimensional target cloud disk behavior labels can be accurately marked for the target tenant, and the target cloud disk behavior labels of the target tenant can be dynamically adjusted in combination with the cloud disk behavior changes of the target tenant and the other tenants in the tenant set, so as to accurately generate tenant labeled feature information for resource scheduling use by the elastic block storage service and improve the resource scheduling ability. BRIEF DESCRIPTION OF THE DRAWINGS

[0020] In order to more clearly illustrate the technical solutions in the embodiments of the present disclosure or the prior art, the following will briefly introduce the drawings required for use in the description of the embodiments or the prior art. Obviously, the drawings in the following description are some embodiments of the present disclosure. For those of ordinary skill in the art, other drawings can be obtained based on these drawings without creative efforts.

[0021] Figure 1 FIG. is a schematic diagram of a resource scheduling scenario based on elastic block storage service provided by an embodiment of the present disclosure;

[0022] Figure 2 FIG. is a schematic flowchart of a resource scheduling method based on elastic block storage service provided by an embodiment of the present disclosure;

[0023] Figure 3 FIG. is a schematic flowchart of a resource scheduling method based on elastic block storage service provided by another embodiment of the present disclosure;

[0024] Figure 4 FIG. is a structural block diagram of a resource scheduling device based on elastic block storage service provided by an embodiment of the present disclosure;

[0025] Figure 5 FIG. is a schematic hardware structure diagram of a resource scheduling device based on elastic block storage service provided by an embodiment of the present disclosure. DETAILED DESCRIPTION OF THE EMBODIMENTS

[0026] To make the objectives, technical solutions, and advantages of the embodiments of the present disclosure clearer, the technical solutions in the embodiments of the present disclosure will be clearly and completely described below with reference to the accompanying drawings in the embodiments of the present disclosure. Apparently, the described embodiments are some, but not all, of the embodiments of the present disclosure. All other embodiments obtained by those of ordinary skill in the art based on the embodiments in the present disclosure without creative efforts shall fall within the scope of protection of the present disclosure.

[0027] A tenant profile is an abstract description of the characteristics of a specific tenant, usually based on data analysis and machine learning techniques. It can help service providers analyze the needs and behavior patterns of tenants, so as to provide more accurate service recommendations. At the same time, tenant behavior can be predicted through tenant profiles, providing decision-making suggestions for coordinating service resources among multi-tenants. Currently, tenant profiles are widely used in to c fields such as e-commerce and advertising. However, due to factors such as data complexity and discreteness, and the data volume of a single tenant, their application in the storage field is not common.

[0028] In the scenario of the Elastic Block Storage service provided by traditional cloud providers, resource scheduling pays more attention to the load model at the cluster level, and performs load management of individual resources based on the overall load of the cluster and the real-time load of resources for resource scheduling. However, tenant profile technology is not currently applied in traditional elastic block storage services, and the resource scheduling ability for specific tenants needs to be further improved.

[0029] To solve the above technical problems, the embodiments of the present disclosure provide a resource scheduling method based on elastic block storage services, which includes obtaining the attribute information and operation event information of each cloud disk of a target tenant in a tenant set; generating a candidate cloud disk behavior label set of the target tenant according to the attribute information and operation event information of each cloud disk; determining the current weight of each candidate cloud disk behavior label in the candidate cloud disk behavior label set according to the cloud disk behavior changes of the target tenant and the other tenants in the tenant set, and determining the target cloud disk behavior label of the target tenant from the candidate cloud disk behavior label set according to the current weight of each candidate cloud disk behavior label; generating labeled feature information of the target tenant according to the target cloud disk behavior label; and scheduling the cloud disks of the target tenant according to the labeled feature information of the target tenant. By statistically analyzing and mining the attribute information and operation event information of each cloud disk of the target tenant, multi-dimensional target cloud disk behavior labels can be accurately marked for the target tenant, and the target cloud disk behavior labels of the target tenant can be dynamically adjusted in combination with the cloud disk behavior changes of the target tenant and the other tenants in the tenant set, so as to accurately generate tenant labeled feature information for use in resource scheduling of elastic block storage services and improve the resource scheduling ability.

[0030] The application scenario of the resource scheduling method based on the elastic block storage service in the embodiments of the present disclosure is as follows Figure 1 As shown, obtain the attribute information and operation event information of each cloud disk of the target tenant; generate a candidate cloud disk behavior label set of the target tenant according to the attribute information and operation event information of each cloud disk; determine the current weight of each candidate cloud disk behavior label in the candidate cloud disk behavior label set according to the cloud disk behavior changes of the target tenant and the other tenants in the tenant set, and determine the target cloud disk behavior label of the target tenant from the candidate cloud disk behavior label set according to the current weight of each candidate cloud disk behavior label; generate the labeled feature information of the target tenant according to the target cloud disk behavior label; and schedule the cloud disks of the target tenant according to the labeled feature information of the target tenant.

[0031] It can be understood that the data involved in this technical solution (including but not limited to the data itself, the acquisition, use, storage or deletion of the data) should comply with the requirements of the corresponding laws, regulations and related provisions.

[0032] It can be understood that before using the technical solutions disclosed in the embodiments of the present disclosure, the type, scope of use, usage scenario, etc. of the information involved in the present disclosure should be informed to the relevant users in an appropriate manner and the authorization of the relevant users should be obtained according to the relevant laws and regulations. Among them, the relevant users may include any type of rights and entities, such as individuals, enterprises, and groups.

[0033] For example, in response to receiving an active request from a user, a prompt message is sent to the relevant user to clearly prompt the relevant user that the operation requested by it will require obtaining and using the information of the relevant user, so that the relevant user can autonomously choose whether to provide information to software or hardware such as an electronic device, an application program, a server, or a storage medium that executes the operation of the technical solution of the present disclosure according to the prompt message.

[0034] As an optional but non-limiting implementation manner, the manner of sending a prompt message to the relevant user in response to receiving an active request from the relevant user may be, for example, a pop-up window manner, and the prompt message may be presented in text in the pop-up window. In addition, the pop-up window may also carry a selection control for the user to select "agree" or "disagree" to provide information to the electronic device.

[0035] It can be understood that the above process of notifying and obtaining user authorization is only illustrative and does not constitute a limitation on the implementation manner of the present disclosure. Other manners that meet the relevant laws and regulations can also be applied to the implementation manner of the present disclosure.

[0036] For the enabling of relevant functions in the embodiments of the present disclosure, the data obtained, the processing and storage methods of the data, etc., prior authorization from the user and other rights holders associated with the user shall be obtained, and the agreements and rules between the rights holders and relevant laws and regulations shall be complied with.

[0037] The resource scheduling method based on elastic block storage service of the present disclosure will be introduced in detail below in combination with specific embodiments.

[0038] Refer to Figure 2 , Figure 2 which is a schematic flowchart of the resource scheduling method based on elastic block storage service provided by an embodiment of the present disclosure. The method of this embodiment can be applied to electronic devices such as terminal devices or servers. The resource scheduling method based on elastic block storage service includes:

[0039] S201. Obtain the attribute information and operation event information of each cloud disk of the target tenant in the tenant set.

[0040] In this embodiment, for any target tenant renting a cloud disk (BlockDevice) of the elastic block storage service, the attribute information and operation event information of each cloud disk of the target tenant can be collected.

[0041] The attribute information of any cloud disk includes, but is not limited to, configuration - type attributes and behavior - pattern - type attributes. Among them, the configuration - type attributes include, but are not limited to, the specifications, capacity, and life cycle of the cloud disk, etc., which can be obtained from databases such as business management and control databases that store configuration - type attributes; the behavior - pattern - type attributes include, but are not limited to, the traffic, qps, and data compression ratio of the cloud disk, etc., which can be obtained from databases such as cloud disk monitoring databases that store behavior - pattern - type attributes.

[0042] The operation event information of any cloud disk may include, but is not limited to, creation events, configuration - change events, etc., which can be obtained from log data. Based on the operation event information of any cloud disk, the operation situation of the target tenant on the cloud disk can be statistically analyzed, such as the number of batch creations of cloud disks, the number of configuration changes, etc.

[0043] Optionally, the attribute information and operation event information of each cloud disk of the target tenant can be stored in a data warehouse as the basic data support for subsequent processing, and then the data warehouse can be periodically polled to read the attribute information and operation event information of each cloud disk of the target tenant for subsequent processing.

[0044] Optionally, after obtaining the attribute information and operation event information of each cloud disk of the target tenant, data cleaning and filtering can be performed to screen out abnormal data.

[0045] S202. Generate a candidate cloud disk behavior label set of the target tenant according to the attribute information and operation event information of each cloud disk.

[0046] In this embodiment, based on the attribute information and operation event information of each cloud disk, various feasible algorithms can be used to perform operations such as statistics, mining, and prediction on the attribute information and operation event information of each cloud disk, generating candidate cloud disk behavior tags in multiple dimensions for the target tenant, forming a candidate cloud disk behavior tag set, which serves as the basis for constructing the labeled feature information of the target household.

[0047] Optionally, in this embodiment, tenant feature tags of the target tenant can be directly generated based on the attribute information and operation event information of each cloud disk; or, alternatively, single-cloud-disk behavior tags of each cloud disk can be directly obtained based on the attribute information and operation event information of each cloud disk, and by further summarizing and generalizing the single-cloud-disk behavior tags, multi-cloud-disk behavior tags of the target tenant can be obtained. Furthermore, the tenant feature tags and multi-cloud-disk behavior tags of the target tenant can be determined as the tenant tags of the target tenant.

[0048] In this embodiment, based on the attribute information and operation event information of each cloud disk, single-cloud-disk behavior tags of each cloud disk can be generated, such as the capacity, life cycle, compression ratio, usage rate, performance, average write size in the last 7 days, average read size in the last 7 days, throttling time in the last 7 days, average morning IOPS in the last 7 days, average evening IOPS in the last 7 days, average morning bandwidth in the last 7 days, average evening bandwidth in the last 7 days, traffic prediction for the next 24 hours, etc. By summarizing, multi-cloud-disk behavior tags can be obtained, such as the average capacity, average life cycle, average compression ratio, average usage rate, average performance, average write size in the last 7 days, average read size in the last 7 days, throttling time in the last 7 days, average morning IOPS in the last 7 days, average evening IOPS in the last 7 days, average morning bandwidth in the last 7 days, average evening bandwidth in the last 7 days, traffic prediction for the next 24 hours, etc. for all cloud disks of the target tenant; based on the attribute information and operation event information of each cloud disk, tenant feature tags of the target tenant can be generated. Tenant feature tags are tags that do not need to be summarized and generalized based on single-cloud-disk behavior tags, such as affinity, anti-affinity, cloud disk opening type, number of cloud disks, user level, number of cloud disks created in the last 7 days, number of cloud disks deleted in the last 7 days, number of times of batch creation of cloud disks in the last 7 days, etc.

[0049] Of course, in this embodiment, generating the candidate cloud disk behavior tag set of the target tenant based on the attribute information and operation event information of each cloud disk is not limited to the above method, and any other feasible method can also be used, such as directly generating the candidate cloud disk behavior tag set through a machine learning model, etc., which is not limited in this embodiment.

[0050] S203. Determine the current weight of each candidate cloud disk behavior label in the candidate cloud disk behavior label set according to the cloud disk behavior changes of the target tenant and the remaining tenants in the tenant set. According to the current weight of each candidate cloud disk behavior label, determine the target cloud disk behavior label of the target tenant from the candidate cloud disk behavior label set; generate the labeled feature information of the target tenant according to the target cloud disk behavior label.

[0051] In this embodiment, after obtaining the candidate cloud disk behavior label set of the target tenant, some or all of the candidate cloud disk behavior labels can be selected from the candidate cloud disk behavior label set and determined as the target cloud disk behavior label of the target tenant, and the labeled feature information of the target tenant is generated based on the target cloud disk behavior label, which is an abstract description of the features of the target tenant. The implementation method can be any feasible method. For example, the labeled feature information close to natural language can be obtained by describing the target cloud disk behavior label. For example, if the target cloud disk behavior labels of the target tenant include the large capacity label and the low bandwidth label, the label descriptions "large capacity, low bandwidth" of the large capacity label and the low bandwidth label can be obtained as the labeled feature information of the target tenant, and the way to obtain the label description can be implemented by a language model; or in this embodiment, the tenant label can also be directly used as the labeled feature information.

[0052] However, considering the cloud disk behavior changes of the target tenant and the remaining tenants in the tenant set, the representativeness of some candidate cloud disk behavior labels in the candidate cloud disk behavior label set will also change. For example, some candidate cloud disk behavior labels are generated for the target tenant in the past for a long time, and the behavior habits of the target user may change, and these candidate cloud disk behavior labels may not accurately reflect the recent cloud disk behavior characteristics of the target tenant; another example is that some candidate cloud disk behavior labels reflect some relative logics, such as high capacity, large bandwidth, etc. Whether the cloud disk of the target tenant is defined as high capacity depends not only on the absolute size of its cloud disk, but also on the cloud disk sizes of all tenants in the entire system. If the cloud disk sizes of the remaining tenants also become larger, the cloud disk of the target tenant may no longer be defined as high capacity. Therefore, in this embodiment, according to the cloud disk behavior changes of the target tenant and the remaining tenants in the tenant set, the current weight of each candidate cloud disk behavior label in the candidate cloud disk behavior label set is determined. Specifically, the current weight of one or more candidate cloud disk behavior labels can be increased or decreased according to different cloud disk behavior change trends. Furthermore, the target cloud disk behavior label of the target tenant can be dynamically determined from the candidate cloud disk behavior label set according to the current weight of each candidate cloud disk behavior label, and then the labeled feature information of the target tenant is generated according to the target cloud disk behavior label, so that the labeled feature information of the target tenant is more accurate and more timely, and can reflect the current characteristics of the target tenant.

[0053] The current weight for determining the candidate cloud disk behavior labels can be obtained by any feasible method. For example, the preset initial weight of any candidate cloud disk behavior label can be obtained, and then the preset initial weight can be updated according to the cloud disk behavior changes of the target tenant and the cloud disk behavior changes of the remaining tenants in the tenant set; alternatively, the current weight of each candidate cloud disk behavior label can also be directly set according to the cloud disk behavior changes of the target tenant and the cloud disk behavior changes of the remaining tenants in the tenant set; or other methods can also be used, which are not limited in this embodiment. In addition, according to the current weight of each candidate cloud disk behavior label, the target cloud disk behavior label of the target tenant is determined from the candidate cloud disk behavior label set. Multiple (TopN, where N is a preset integer) candidate cloud disk behavior labels with the largest current weight can be selected as the target cloud disk behavior label of the target tenant according to the magnitude of the current weight, or other determination rules can also be used.

[0054] S204. Schedule the cloud disk of the target tenant according to the labeled feature information of the target tenant.

[0055] In this embodiment, after obtaining the labeled feature information of the target tenant, the labeled feature information of the target tenant can be applied to subsequent cloud disk scheduling.

[0056] For example, the business traffic can be accurately predicted according to the labeled feature information of the target tenant, and the cluster risk can be perceived in advance. Specifically, through the large traffic-related data in the cloud disk prediction class labels, it can be perceived in advance that the cluster may have traffic risks in a certain period in the future, and some cloud disks can be migrated out in time to ensure the stability of the business.

[0057] The cloud disk specification can also be accurately recommended according to the labeled feature information of the target tenant. Specifically, by counting the labels such as the average capacity, average bandwidth, and daily traffic limiting times of the cloud disk of the target tenant, it can be reflected whether the current cloud disk specification matches the needs of the target tenant, and the cloud disk specification that matches the needs of the target tenant can be accurately recommended.

[0058] The usage status of the target tenant can also be perceived according to the labeled feature information of the target tenant, and the disk creation cluster can be reasonably scattered. Specifically, according to the labels such as the high or low capacity and high or low traffic of the target tenant, the storage pool resources can be reasonably allocated when creating a cloud disk, paying attention to the resource allocation of tenants with high capacity and / or high traffic labels, and relatively ignoring the resource allocation of tenants with low capacity and / or low traffic labels.

[0059] The change in the behavior mode of the target tenant can also be perceived according to the labeled feature information of the target tenant. Specifically, the time cooling algorithm pays more attention to the recent behavior of the target tenant, and when the behavior of the target tenant changes, it can be quickly perceived and the label can be modified.

[0060] The resource scheduling method based on the elastic block storage service provided in this embodiment obtains the attribute information and operation event information of each cloud disk of the target tenant in the tenant set; generates a candidate cloud disk behavior label set of the target tenant according to the attribute information and operation event information of each cloud disk; determines the current weight of each candidate cloud disk behavior label in the candidate cloud disk behavior label set according to the cloud disk behavior changes of the target tenant and the other tenants in the tenant set, and determines the target cloud disk behavior label of the target tenant from the candidate cloud disk behavior label set according to the current weight of each candidate cloud disk behavior label; generates the labeled feature information of the target tenant according to the target cloud disk behavior label; and schedules the cloud disks of the target tenant according to the labeled feature information of the target tenant. By statistically analyzing and mining the attribute information and operation event information of each cloud disk of the target tenant, multi-dimensional target cloud disk behavior labels can be accurately marked for the target tenant, and the target cloud disk behavior labels of the target tenant can be dynamically adjusted in combination with the cloud disk behavior changes of the target tenant and the other tenants in the tenant set. Furthermore, the labeled feature information of the tenant can be accurately generated for the resource scheduling of the elastic block storage service, improving the resource scheduling ability.

[0061] In any of the above embodiments, when generating the candidate cloud disk behavior label set of the target tenant according to the attribute information and operation event information of each cloud disk, different means such as statistics, mining, and prediction can be used. Therefore, the candidate cloud disk behavior labels can be divided into statistical labels, mining labels, and prediction labels. Among them, the single cloud disk behavior labels and tenant feature labels can both be divided into statistical labels, mining labels, and prediction labels.

[0062] In the above embodiment, when generating the candidate cloud disk behavior label set of the target tenant according to the attribute information and operation event information of each cloud disk, different methods can be used for statistical labels, rule labels, and prediction labels, as follows:

[0063] Optionally, for statistical labels, according to the attribute information and / or operation event information of each cloud disk, obtain the cumulative amount of the preset cloud disk behavior indicators of each cloud disk within the target time window, and determine the single cloud disk behavior label corresponding to each cloud disk according to the cumulative amount of the preset cloud disk behavior indicators; and / or

[0064] According to the attribute information and / or operation event information of all the cloud disks of the target tenant, obtain the cumulative amount of the preset tenant characteristic indicators of the target tenant within the target time window, and determine the tenant feature label of the target tenant according to the cumulative amount of the preset tenant characteristic indicators.

[0065] In this embodiment, for the single disk behavior tags of the statistical type, the cumulative amount of the preset disk behavior metrics of each disk within the target time window (e.g., the most recent 7 days) can be obtained. Then, based on the cumulative amount of the preset disk behavior metrics, the single disk behavior tag of each disk can be determined. For example, the average write size in the most recent 7 days can be obtained by statistically averaging the cumulative amount of the write data of each disk in the most recent 7 days. In addition, relevant single disk behavior tags such as the resource life cycle, flow limiting situation, and time-sharing data of each disk can also be determined through statistical methods. Among them, the resource life cycle can be used to reasonably schedule disks with different resource life cycles during the disk scheduling process. The flow limiting situation can be used to allocate clusters that meet the bandwidth requirements during disk scheduling. The time-sharing data can be used to balance the resources between clusters according to the tenant tide during disk scheduling.

[0066] For the tenant feature tags of the statistical type, similarly, the cumulative amount of the preset tenant feature metrics of the target tenant within the target time window (e.g., the most recent 7 days) can be obtained. Then, based on the cumulative amount of the preset tenant feature metrics, the tenant feature tag of the target tenant can be determined. For example, the number of disks created in the most recent 7 days can be obtained by statistically counting the cumulative number of disks created by the target tenant in the most recent 7 days. The number of times of batch disk creation in the most recent 7 days can be obtained by determining whether there is batch disk creation based on the time of disk creation in the most recent 7 days and accumulating the number of times of batch disk creation.

[0067] Optionally, for the rule-based tags, it is determined whether the attribute information and / or operation event information of any disk satisfy the first preset rule corresponding to the single disk behavior tag. If satisfied, a single disk behavior tag is generated for the any disk; and / or

[0068] It is determined whether the attribute information and / or operation event information of all the disks of the target tenant satisfy the second preset rule corresponding to the tenant feature tag. If satisfied, a tenant feature tag is generated for the target tenant.

[0069] In this embodiment, discriminant rules corresponding to one or more single disk behavior tags, that is, the first preset rules, can be configured. For example, rules for judging the high and low levels of disk utilization rate, rules for judging the size levels of the average write size of the disk in the most recent 7 days, rules for judging the size levels of the average bandwidth in the early morning in the most recent 7 days, etc. Through the above rules, the levels of the above-mentioned rule metrics of the disk can be determined, and single disk behavior tags of corresponding levels can be obtained. For example, utilization rate level tags (which can be divided into high, medium, low levels, etc.), average write size level tags in the most recent 7 days (which can be divided into high, medium, low levels, etc.).

[0070] Similarly, corresponding discrimination rules, i.e., the second preset rules, can be configured for one or more tenant feature tags. For example, rules for judging the magnitude level of the number of cloud disks batch-created in the most recent 7 days, rules for judging the magnitude level of the frequency of cloud disks batch-created in the most recent 7 days, rules for judging the magnitude level of the utilization rate, etc. Through the above rules, the levels of the above rule indicators of the target tenant can be determined, and tenant feature tags corresponding to the levels can be obtained. For example, the number of cloud disks batch-created in the most recent 7 days level tag (which can be divided into high, medium, low levels, etc.), the magnitude level tag of the frequency of cloud disks batch-created in the most recent 7 days (which can be divided into high, medium, low levels, etc.).

[0071] Optionally, for prediction-type tags, according to the attribute information and / or operation event information of each cloud disk, historical data of the cloud disk behavior indicators to be predicted within the historical time window for each cloud disk is obtained. According to the time series historical data of the cloud disk behavior indicators to be predicted, a prediction model is called to obtain the predicted value of the cloud disk behavior indicators to be predicted. According to the predicted value of the cloud disk behavior indicators to be predicted, the single-cloud-disk behavior tag for each cloud disk is determined; and / or

[0072] According to the attribute information and / or operation event information of all the cloud disks of the target tenant, historical data of the tenant feature indicators to be predicted of the target tenant within the historical time window is obtained. According to the historical data of the tenant feature indicators to be predicted, a prediction model is called to obtain the predicted value of the tenant feature indicators to be predicted. According to the predicted value of the tenant feature indicators to be predicted, the tenant feature tag of the target tenant is obtained.

[0073] In this embodiment, for each cloud disk, the cloud disk behavior indicators to be predicted can be predicted. Historical data of the cloud disk behavior indicators to be predicted for each cloud disk within the target time window can be obtained. According to the historical data of the cloud disk behavior indicators to be predicted, a prediction model can be called, and the prediction model obtains the predicted value of the cloud disk behavior indicators to be predicted. The prediction model can be any machine learning model. For example, it can be a Long Short-Term Memory (LSTM) network, etc. Furthermore, the single-cloud-disk behavior tag for each cloud disk can be determined according to the predicted value of the cloud disk behavior indicators to be predicted. For example, to predict the traffic of each cloud disk in the next 24 hours, the time series of the traffic of each cloud disk in the past 7 days can be obtained, and the traffic in the next 24 hours can be predicted through the LSTM model as a single-cloud-disk behavior tag. As an optimization algorithm of the Recurrent Neural Network (RNN), LSTM can well capture long-term dependencies and predict future situations.

[0074] Similarly, for the target tenant, the tenant feature metrics to be predicted can also be predicted. Historical data of the tenant feature metrics to be predicted for the target tenant within the target time window can be obtained. The prediction model can be called based on the historical data of the tenant feature metrics to be predicted, and the prediction model can obtain the predicted values of the tenant feature metrics to be predicted. Furthermore, a tenant label can be determined based on the predicted values of the tenant feature metrics to be predicted. For example, to predict the number of times the target tenant calls the API in the next 24 hours, the time series of the number of times the target tenant calls the API in the past 7 days can be obtained, and the LSTM model can be used to predict the number of times the API will be called in the next 24 hours as a tenant label. Another example is to predict the time when the target tenant will batch-create cloud disks in the future. The time when the target tenant batch-created cloud disks in the past 7 days can be obtained, and the LSTM model can be used to predict the time when the target tenant will batch-create cloud disks in the future. The cloud disk behavior metrics to be predicted and the tenant feature metrics to be predicted can be of the same type as the cloud disk behavior metrics and tenant feature metrics mentioned in the above embodiments, except that they are the predicted values of these metrics in the future time to be predicted. Of course, they can also be of different types from the cloud disk behavior metrics and tenant feature metrics mentioned in the above embodiments.

[0075] Furthermore, on top of the prediction model, the actual data can be compared to judge the difference between the predicted data and the actual data, and the prediction parameter coefficients can be adjusted to adaptively optimize the prediction model.

[0076] In addition, considering the above-mentioned rule-based labels, the above first preset rule and / or second preset rule are rules for judging levels, that is, whether the relevant rule metrics exceed the corresponding preset metric thresholds. In order to improve the label accuracy and the adaptive optimization ability, these preset metric thresholds can be dynamically adjusted. Specifically, the adjustment amount of the preset metric thresholds can be determined according to the total resources, the used resources, and the predicted used resources of the cloud disk cluster to which the target user belongs, and the preset metric thresholds can be adjusted according to the adjustment amount.

[0077] In this embodiment, for rule-based labels, such as scheduling weight labels for tenant bandwidth, capacity usage, etc., in addition to comparing each rule metric related to all the tenant's resources with the corresponding rule metric thresholds, the water level changes of each resource in the cluster also need to be considered. Taking the capacity usage level as an example, it is assumed that a preset single-disk capacity of less than 100 GB is called low capacity, less than 400 GB is called medium capacity, and the rest is called high capacity. By combining the label thresholds with the resource water levels and dynamically adjusting the threshold sizes, the accuracy of the labels can be improved. For example, if the label threshold decreases as the resource water level increases, the label level of the corresponding resource can be raised. For example, a preset single-disk capacity of less than 80 GB can be called low capacity. Originally, a single-disk capacity of 90 GB was low capacity, but now it becomes medium capacity to increase the importance of the label.

[0078] Based on the adaptive optimization capability of label accuracy in the previous section, through the capacity and traffic forecast of tenants, the changes in the water level of each resource in each cluster in the next 24 hours can be counted. The rule-type label is also used to configure scheduling rules for a period of time in the future. Combining the current resource water level of the cluster with the adaptively optimized resource water level forecast for the next 24 hours, the preset indicator threshold can be adjusted more accurately, and future resource scheduling can be performed more accurately. The calculation formula for the preset indicator threshold is as follows:

[0079]

[0080] in:

[0081] T init : The initial threshold of the preset indicator threshold can be obtained by statistically analyzing historical data;

[0082] cluster used : The amount of resources that the cluster has used;

[0083] cluster total : The total resources of the cluster;

[0084] cluster futureUsed : The predicted value of the amount of resources that the cluster will use in the future;

[0085] cluster futureTotal : The total number of resources in the cluster in the future, generally equal to cluster total ;

[0086] α, β: weight factors;

[0087] γ: adjustment factor.

[0088] In the above formula for calculating the adjusted rule indicator threshold T, except for T init The part outside the threshold can be regarded as the adjustment amount of the preset indicator threshold.

[0089] Based on any of the above embodiments, according to the cloud disk behavior changes of the target tenant and the remaining tenants in the tenant set, the current weight of each candidate cloud disk behavior tag in the candidate cloud disk behavior tag set is determined, which can be specifically as follows: Figure 3 As shown, including:

[0090] S301, based on the target cloud disk behavior tags of all tenants in the tenant set, determining a real-time correlation coefficient between any candidate cloud disk behavior tag in the candidate cloud disk behavior tag set and the target tenant;

[0091] S302: Determine a time decay coefficient of any candidate cloud disk behavior label according to the cloud disk behavior change of the target tenant;

[0092] S303. Determine the current weight of any candidate cloud disk behavior label according to the real-time relevance coefficient and the time decay characteristic coefficient.

[0093] In this embodiment, for each candidate cloud disk behavior label in the candidate cloud disk behavior label set, the real-time relevance coefficient between the candidate cloud disk behavior label and the target tenant can be determined based on the target cloud disk behavior labels of all tenants in the tenant set. The real-time relevance coefficient can be used to represent the real-time relevance between the candidate cloud disk behavior label and the target tenant, or can also represent the representativeness size of the candidate cloud disk behavior label for the target tenant. For example, if many tenants in the tenant set have this candidate cloud disk behavior label, then the real-time relevance between the candidate cloud disk behavior label and the target tenant is relatively small, or the representativeness of the candidate cloud disk behavior label for the target tenant is relatively small. Therefore, the real-time relevance coefficient can be reduced.

[0094] In addition, the time decay coefficient of any candidate cloud disk behavior label can be determined according to the cloud disk behavior change of the target tenant. The time decay coefficient is used to reflect the representativeness size of the candidate cloud disk behavior label for the target tenant decaying over time, which can weaken the influence of the cloud disk behavior of the target tenant in the past for a long time on the candidate cloud disk behavior label, and at the same time strengthen the influence of the cloud disk behavior of the target tenant recently on the candidate cloud disk behavior label.

[0095] Furthermore, the current weight of the candidate cloud disk behavior label can be determined according to the real-time relevance coefficient and the time decay characteristic coefficient of the candidate cloud disk behavior label.

[0096] Specifically, the preset initial weight of the candidate cloud disk behavior label can be obtained, where the preset initial weight can be preset or the weight at a historical moment. Furthermore, the preset initial weight of the candidate cloud disk behavior label can be updated according to the real-time relevance coefficient and the time decay characteristic coefficient of the candidate cloud disk behavior label to obtain the current weight of the candidate cloud disk behavior label. When specifically implemented, the real-time relevance coefficient and the time decay characteristic coefficient can be multiplied on the basis of the preset weight to obtain the current weight of the candidate cloud disk behavior label. Further, the final target single cloud disk behavior label can be determined according to the current weights of all candidate cloud disk behavior labels. For example, one or more candidate cloud disk behavior labels with the highest current weights are selected as the multi-cloud disk behavior labels of the target tenant.

[0097] The real-time correlation coefficient between the target single cloud disk behavior label and the target tenant can represent the magnitude of the correlation between the target single cloud disk behavior label and the target tenant. The acquisition method can adopt any feasible method. For example, the TF-IDF (Term Frequency-Inverse Document Frequency) algorithm can be used. By multiplying the occurrence frequency of a certain candidate cloud disk behavior label of the target tenant by the inverse document frequency of this candidate cloud disk behavior label among all tenants, the weight ratio of the corresponding index of a single tenant in the overall tenants can be obtained, and the correlation coefficient between the target tenant and this candidate cloud disk behavior label can be obtained. Specifically, it is as follows:

[0098] Determine the occurrence frequency of any one of the candidate cloud disk behavior labels in the set of candidate cloud disk behavior labels as the term frequency of any one of the candidate cloud disk behavior labels;

[0099] Determine the reciprocal of the occurrence frequency of any one of the candidate cloud disk behavior labels in the target cloud disk behavior labels of all tenants in the tenant set as the inverse document frequency of any one of the candidate cloud disk behavior labels;

[0100] Adopt the term frequency-inverse document frequency method to determine the real-time correlation coefficient between any one of the candidate cloud disk behavior labels and the target tenant according to the term frequency and inverse document frequency of any one of the candidate cloud disk behavior labels.

[0101] Specifically, the TF-IDF real-time correlation coefficient w TF-IDF The calculation formula is as follows:

[0102]

[0103]

[0104] w TF-IDF =TF(P,T)*IDF(P,T)

[0105] Where:

[0106] TF(P,T): The term frequency of the candidate cloud disk behavior label, indicating the occurrence frequency of this candidate cloud disk behavior label in the set of candidate cloud disk behavior labels of the target tenant, that is, the ratio of the number of occurrences of this candidate cloud disk behavior label to the number of all candidate cloud disk behavior labels in the set of candidate cloud disk behavior labels. w(P,T) represents the number of times a candidate cloud disk behavior label T is used to label a certain tenant P;

[0107] IDF(P,T): The inverse document frequency of the candidate cloud disk behavior label, indicating the probability of this candidate cloud disk behavior label appearing in all labels tags of all tenants users, that is, the scarcity degree of label T.

[0108] For the behavior changes of single-tenant, a time cooling algorithm can be used to weaken the influence of tenant behavior in the past for a long time on the tenant label, while enhancing the influence of recent tenant behavior on the tenant label. Specifically, determining the time decay weight coefficient of any one of the candidate cloud disk behavior labels may include:

[0109] Determine the popularity of the candidate cloud disk behavior label at the current time according to the popularity of the candidate cloud disk behavior label at the previous historical time, a preset cooling coefficient, and the time interval between the previous historical time and the current time;

[0110] Determine the time decay weight coefficient of the candidate cloud disk behavior label at the current time according to the time decay weight coefficient of the candidate cloud disk behavior label at the previous historical time, the popularity at the current time, and the current quantity of the candidate cloud disk behavior label.

[0111] -c*t

[0112] t cur =t before *e

[0113] w t =w0*t cur *T

[0114] Wherein:

[0115] t cur : represents the popularity of any one of the candidate cloud disk behavior labels at the current time, which is calculated from the popularity t before of the candidate cloud disk behavior label at the previous historical time, the cooling coefficient c, and the interval time t;

[0116] w t : the time decay weight coefficient, which is obtained by calculating the time decay weight coefficient w0 of the candidate cloud disk behavior label at the previous historical time, the popularity t cur at the current time, and the current quantity T of the candidate cloud disk behavior label within the interval time t.

[0117] In summary, the current weight formula of the candidate cloud disk behavior label can be obtained as:

[0118] w=w k *w TF-IDF *w t

[0119] Wherein, w k is the preset weight of the candidate cloud disk behavior label.

[0120] It should be noted that in this embodiment, the execution order of S301 and S302 is not limited.

[0121] Based on any of the above embodiments, when determining the target cloud disk behavior label of the target tenant from the candidate cloud disk behavior label set, it may further include:

[0122] Screen out associated label groups from the candidate cloud disk behavior label set according to a preset associated label set; wherein the preset associated label set includes multiple groups of associated label groups composed of mutually associated cloud disk behavior labels;

[0123] For each group of screened associated label groups, only retain one of the cloud disk behavior labels for label deduplication.

[0124] In this embodiment, considering that there may be some highly correlated metrics, excessive correlation between labels may lead to duplicate label weights, thereby reducing the accuracy when using labels. Therefore, a preset associated label set can be obtained in advance, which includes multiple groups of associated label groups composed of mutually associated cloud disk behavior labels. Further, associated label groups can be screened from the candidate cloud disk behavior label set of the target tenant according to the preset associated label set. For example, in the preset associated label set, label A and label B are a group of associated label groups. If label A and label B are screened out from the candidate cloud disk behavior label set of the target tenant, then a deduplication operation can be performed, that is, for each group of screened associated label groups, only retain one of the labels for label deduplication, and either label A or label B can be retained.

[0125] Optionally, before screening out associated label groups from the candidate cloud disk behavior label set according to the preset associated label set, it further includes:

[0126] Obtain the candidate cloud disk behavior label sets of multiple tenants, and construct a co-occurrence matrix of candidate cloud disk behaviors according to the candidate cloud disk behavior label sets of multiple tenants;

[0127] Obtain the similarity between the co-occurring candidate cloud disk behavior labels in the co-occurrence matrix. If the similarity between any two candidate cloud disk behavior labels exceeds a preset similarity threshold, then determine the any two candidate cloud disk behavior labels as a group of associated label groups;

[0128] Construct the preset associated label set according to the determined associated label groups.

[0129] In this embodiment, to analyze the correlation between tags, a tag co-occurrence matrix is constructed to calculate the similarity between tags. Here, co-occurrence means that tags appear simultaneously, that is, a tenant is tagged with both tag A and tag B at the same time. If many tenants are tagged with both A and B, there may be a potential correlation between A and B tags. After constructing the tag co-occurrence matrix, the similarity between candidate cloud disk behavior tags co-occurring in the co-occurrence matrix is obtained. For example, the cosine similarity function is used to calculate the correlation between two tags. The cosine similarity function measures the size of the difference between two individuals through the cosine value of the angle between two vectors in space. The closer the cosine value is to 1, the greater the similarity between the two vectors. For example, the high bandwidth tag is applied to a tenants, and the burst tag is applied to b tenants, and among them, x tenants have both the high bandwidth tag and the burst tag at the same time. Then the similarity r between high bandwidth and burst is:

[0130]

[0131] Through tag correlation analysis, it can assist in making operation decisions, analyze the association between tags, and also assist in feature selection, remove duplicate features of tenants, and help with tenant clustering.

[0132] For example, when scheduling a block storage cluster, two dimensions of capacity and the number of segments need to be considered. Since a segment is the smallest unit that makes up a cloud disk, generally, the larger the capacity used by a cluster, the larger the number of segments. However, in the architecture design, there is a minimum capacity limit for a single segment. When there are a large number of cloud disks with a capacity smaller than the minimum capacity of a single segment in a certain cluster, there will be a deviation between the number of segments and the cluster capacity level. The tag association calculation can be used to determine whether these two dimensions need to be considered simultaneously during cluster scheduling.

[0133] Based on any of the above embodiments, for tags that distinguish high and low levels such as capacity and traffic, it is expected that the total number of tags for each level will not differ too much. Therefore, the preset index threshold for each level can be adjusted to balance the number of tags for each level. The specific process is as follows:

[0134] In the case of dividing tags into different levels according to the preset index threshold, determine the number of tags for each level;

[0135] Modify the preset index threshold according to the number of tags for each level to balance the number of tags for each level.

[0136] In this embodiment, the difference between the number of tags for each level and the mean value can be calculated through the following algorithm:

[0137]

[0138] Among them, x is the number of labels in any dimension of high, medium, and low, and N is the total number of such labels. When the difference d is greater than the critical value, the preset index threshold is triggered for correction. After correcting the preset index threshold, the corresponding labels can be obtained again, so as to balance the number of labels of each level.

[0139] Based on any of the above embodiments, the method can further determine whether the labeled feature information of the generated target tenant fluctuates. Specifically, the fluctuation situation of the labeled feature information of the target tenant can be determined first. If the fluctuation situation is abnormal, an alarm is issued.

[0140] Specifically, the target cloud disk behavior labels of the target tenant can be vectorized to obtain the label feature vector of the target tenant. Further, the label feature vectors of the target tenant at different times within the current time window can be obtained; the fluctuation situation of the labeled feature information of the target tenant is determined according to the label feature vectors of the target tenant at different times. If the fluctuation situation is abnormal, an alarm is issued.

[0141] In this embodiment, considering that the tenant behavior is relatively stable and there will be no large fluctuations, and the tenant labeled feature information will not change frequently, so in this embodiment, the fluctuation situation of the labeled feature information of the target tenant can be evaluated, and if an abnormality is found, an alarm is issued. Specifically, through the above embodiment, the label feature vector of the target tenant is constructed according to the target cloud disk behavior labels of the target tenant, and the label feature vectors of the target tenant at different times within the current time window (such as the most recent 15 days) can be obtained. The fluctuation situation of the labeled feature information of the target tenant is determined according to the label feature vectors of the target tenant at different times. For example, any feasible method such as variance calculation can be used to measure the fluctuation situation of the labeled feature information of the target tenant. If the fluctuation situation is abnormal, for example, exceeding the preset fluctuation situation threshold, it means that the labeled feature information changes frequently, and an alarm can be triggered. The abnormality can be manually eliminated. For example, the labeled feature information can be corrected, or the generation method of the labeled feature information or the generation method of the labels can be optimized, etc., to improve the accuracy of the labeled feature information and reduce the fluctuation.

[0142] Based on any of the above embodiments, considering that each tenant may have a large number of target cloud disk behavior labels, it is difficult to find the rules from the large number of target cloud disk behavior labels and accurately determine the labeled feature information of a single tenant. Therefore, in this embodiment, the tenant labels of multiple tenants can be clustered to find the commonalities among the tenants, and then the labeled feature information of similar tenants can be summarized, which is more reasonable and accurate than the labeled feature information of a single tenant. Specifically, in S203, the labeled feature information of the target tenant is generated according to the target cloud disk behavior labels, such asFigure 3 As shown, it may include:

[0143] Construct a label feature vector based on all target cloud disk behavior labels of the target tenant;

[0144] Cluster the label feature vectors of different tenants to obtain different tenant categories, determine the center point of each tenant category, and determine a preset number of target cloud disk behavior labels closest to the center point of each tenant category;

[0145] Respectively obtain the label descriptions of a preset number of target cloud disk behavior labels closest to the center point of each tenant category as the labeled feature information of each tenant category.

[0146] In this embodiment, for each tenant, all target cloud disk behavior labels of the tenant are constructed into a multi-dimensional vector. Therefore, the tenant can be regarded as a point in the multi-dimensional vector space. Furthermore, vectors of different tenants are clustered. The clustering method can adopt any clustering algorithm such as K-means. Through clustering, similar tenants can be grouped into a tenant category. The center point of each tenant category can be determined. Furthermore, a preset number of target cloud disk behavior labels closest to the center point of each tenant category can be determined (for example, determining which label dimension's coordinate axis is closer to the center point, etc.). It can be determined that these target cloud disk behavior labels have a greater correlation with the tenant category and are more in line with the tenant category. The label descriptions of these target cloud disk behavior labels can be obtained to get the labeled feature information of this type of tenant category.

[0147] Corresponding to the resource scheduling method based on the elastic block storage service in the above embodiment, Figure 4 This is the structural block diagram of the resource scheduling device based on the elastic block storage service provided by the embodiments of the present disclosure. For the sake of convenience of description, only parts related to the embodiments of the present disclosure are shown. Refer to Figure 4 , the resource scheduling device 400 based on the elastic block storage service includes: an acquisition unit 401, a label generation unit 402, a feature information generation unit 403, and a scheduling unit 404.

[0148] Among them, the acquisition unit 401 is used to acquire the attribute information and operation event information of each cloud disk of the target tenant in the tenant set;

[0149] The label generation unit 402 is used to generate a set of candidate cloud disk behavior labels of the target tenant according to the attribute information and operation event information of each cloud disk; determine the current weight of each candidate cloud disk behavior label in the set of candidate cloud disk behavior labels according to the target tenant and the cloud disk behavior changes of the remaining tenants in the tenant set, and determine the target cloud disk behavior labels of the target tenant from the set of candidate cloud disk behavior labels according to the current weight of each candidate cloud disk behavior label;

[0150] A feature information generation unit 403, configured to generate labeled feature information of the target tenant according to the target cloud disk behavior label;

[0151] A scheduling unit 404, configured to schedule the cloud disk of the target tenant according to the labeled feature information of the target tenant.

[0152] According to one or more embodiments of the present disclosure, when determining the current weight of each candidate cloud disk behavior label in the candidate cloud disk behavior label set according to the cloud disk behavior changes of the target tenant and the remaining tenants in the tenant set, the label generation unit 402 is configured to:

[0153] Based on the target cloud disk behavior labels of all tenants in the tenant set, determine the real-time correlation coefficient between any candidate cloud disk behavior label in the candidate cloud disk behavior label set and the target tenant;

[0154] According to the cloud disk behavior change of the target tenant, determine the time decay coefficient of any candidate cloud disk behavior label;

[0155] According to the real-time correlation coefficient and the time decay characteristic coefficient, determine the current weight of any candidate cloud disk behavior label.

[0156] According to one or more embodiments of the present disclosure, when determining the current weight of any candidate cloud disk behavior label according to the real-time correlation coefficient and the time decay characteristic coefficient, the label generation unit 402 is configured to:

[0157] Obtain the preset initial weight of any candidate cloud disk behavior label;

[0158] Update the preset initial weight according to the real-time correlation coefficient and the time decay characteristic coefficient to obtain the current weight of any candidate cloud disk behavior label.

[0159] According to one or more embodiments of the present disclosure, when determining the real-time correlation coefficient between any candidate cloud disk behavior label in the candidate cloud disk behavior label set and the target tenant based on the target cloud disk behavior labels of all tenants in the tenant set, the label generation unit 402 is configured to:

[0160] Determine the occurrence frequency of any candidate cloud disk behavior label in the candidate cloud disk behavior label set as the word frequency of any candidate cloud disk behavior label;

[0161] Determine the reciprocal of the occurrence frequency of any candidate cloud disk behavior label in the target cloud disk behavior labels of all tenants in the tenant set as the inverse document frequency of any candidate cloud disk behavior label;

[0162] Adopt the term frequency - inverse document frequency method to determine the real - time relevance coefficient between any candidate cloud disk behavior label and the target tenant according to the term frequency and inverse document frequency of any candidate cloud disk behavior label.

[0163] According to one or more embodiments of the present disclosure, when generating the set of candidate cloud disk behavior labels of the target tenant based on the attribute information and operation event information of each cloud disk, the label generation unit 402 is configured to:

[0164] Generate a single - cloud - disk behavior label corresponding to each cloud disk according to the attribute information and operation event information of each cloud disk; and generate a multi - cloud - disk behavior label of the target tenant according to the single - cloud - disk behavior label corresponding to each cloud disk; and / or

[0165] Generate a tenant feature label of the target tenant according to the attribute information and operation event information of each cloud disk;

[0166] Determine the tenant feature label and / or the multi - cloud - disk behavior label as the candidate cloud disk behavior labels of the target tenant, constituting the set of candidate cloud disk behavior labels.

[0167] According to one or more embodiments of the present disclosure, when generating a single - cloud - disk behavior label corresponding to each cloud disk based on the attribute information and operation event information of each cloud disk, the label generation unit 402 is configured to:

[0168] Obtain the cumulative amount of the preset cloud disk behavior metrics of each cloud disk within the target time window according to the attribute information and / or operation event information of each cloud disk, and determine the single - cloud - disk behavior label corresponding to each cloud disk according to the cumulative amount of the preset cloud disk behavior metrics; and / or

[0169] When generating a tenant feature label of the target tenant based on the attribute information and operation event information of each cloud disk, the label generation unit 402 is configured to:

[0170] Obtain the cumulative amount of the preset tenant characteristic metrics of the target tenant within the target time window according to the attribute information and / or operation event information of all the cloud disks of the target tenant, and determine the tenant feature label of the target tenant according to the cumulative amount of the preset tenant characteristic metrics.

[0171] According to one or more embodiments of the present disclosure, when generating a single - cloud - disk behavior label corresponding to each cloud disk based on the attribute information and operation event information of each cloud disk, the label generation unit 402 is configured to:

[0172] Determine whether the attribute information and / or operation event information of any cloud disk satisfies the first preset rule corresponding to the single cloud disk behavior label. If so, generate a single cloud disk behavior label for the any cloud disk; and / or

[0173] When generating the tenant feature label of the target tenant according to the attribute information and operation event information of each cloud disk, the label generation unit 402 is configured to:

[0174] Determine whether the attribute information and / or operation event information of all cloud disks of the target tenant satisfies the second preset rule corresponding to the tenant feature label. If so, generate a tenant feature label for the target tenant.

[0175] According to one or more embodiments of the present disclosure, the first preset rule and / or the second preset rule is a rule index for judgment to determine whether it exceeds the corresponding preset index threshold;

[0176] Correspondingly, the label generation unit 402 is further configured to:

[0177] Determine an adjustment amount of the preset index threshold according to the total resource amount, the used resource amount, and the predicted used resource amount of the cloud disk belonging to the cluster of the target user, and adjust the preset index threshold according to the adjustment amount.

[0178] According to one or more embodiments of the present disclosure, when generating a single cloud disk behavior label corresponding to each cloud disk according to the attribute information and operation event information of each cloud disk, the label generation unit 402 is configured to:

[0179] Obtain historical data of the cloud disk behavior index to be predicted for each cloud disk within the historical time window according to the attribute information and / or operation event information of each cloud disk, call a prediction model according to the time series historical data of the cloud disk behavior index to be predicted, obtain a predicted value of the cloud disk behavior index to be predicted, and determine a single cloud disk behavior label for each cloud disk according to the predicted value of the cloud disk behavior index to be predicted; and / or

[0180] When generating the tenant feature label of the target tenant according to the attribute information and operation event information of each cloud disk, the label generation unit 402 is configured to:

[0181] Obtain historical data of the tenant characteristic index to be predicted of the target tenant within the historical time window according to the attribute information and / or operation event information of all cloud disks of the target tenant, call a prediction model according to the historical data of the tenant characteristic index to be predicted, obtain a predicted value of the tenant characteristic index to be predicted, and obtain the tenant feature label of the target tenant according to the predicted value of the tenant characteristic index to be predicted.

[0182] According to one or more embodiments of the present disclosure, when determining the target cloud disk behavior label of the target tenant from the candidate cloud disk behavior label set, the label generation unit 402 is configured to:

[0183] Filter out associated label groups from the candidate cloud disk behavior label set according to a preset associated label set; wherein the preset associated label set includes multiple groups of associated label groups composed of mutually associated cloud disk behavior labels;

[0184] For each group of filtered associated label groups, only one cloud disk behavior label is retained for label deduplication.

[0185] According to one or more embodiments of the present disclosure, before filtering out associated label groups from the candidate cloud disk behavior label set according to the preset associated label set, the label generation unit 402 is further configured to:

[0186] Obtain the candidate cloud disk behavior label sets of multiple tenants, and construct a co-occurrence matrix of candidate cloud disk behaviors according to the candidate cloud disk behavior label sets of multiple tenants;

[0187] Obtain the similarity between the co-occurring candidate cloud disk behavior labels in the co-occurrence matrix. If the similarity between any two candidate cloud disk behavior labels exceeds a preset similarity threshold, then determine the any two candidate cloud disk behavior labels as a group of associated label groups;

[0188] Construct the preset associated label set according to the determined associated label groups.

[0189] According to one or more embodiments of the present disclosure, the feature information generation unit 404 is further configured to:

[0190] Determine the fluctuation of the labeled feature information of the target tenant. If the fluctuation is abnormal, an alarm is generated.

[0191] The device provided in this embodiment can be used to execute the technical solutions of the above method embodiments, and its implementation principles and technical effects are similar, which will not be elaborated here in this embodiment.

[0192] To implement the above embodiments, an electronic device is further provided in the embodiments of the present disclosure.

[0193] Refer to Figure 5, which shows a schematic structural diagram of an electronic device 500 suitable for implementing the embodiments of the present disclosure. The electronic device 500 may be a terminal device or a server. Among them, the terminal device may include, but is not limited to, mobile terminals such as mobile phones, laptop computers, digital broadcast receivers, personal digital assistants (PDAs), tablet computers, portable media players (PMPs), in-vehicle terminals (such as in-vehicle navigation terminals), etc., and fixed terminals such as digital TVs, desktop computers, etc. Figure 5 The electronic device shown is merely an example and should not impose any limitations on the functions and usage scope of the embodiments of the present disclosure.

[0194] As Figure 5 shown, the electronic device 500 may include a processing device (such as a central processing unit, a graphics processing unit, etc.) 501, which may perform various appropriate actions and processes according to the program stored in the read-only memory (ROM) 502 or the program loaded from the storage device 508 into the random access memory (RAM) 503. In the RAM 503, various programs and data required for the operation of the electronic device 500 are also stored. The processing device 501, the ROM 502, and the RAM 503 are connected to each other through a bus 504. The input / output (I / O) interface 505 is also connected to the bus 504.

[0195] Generally, the following devices may be connected to the I / O interface 505: an input device 506 including, for example, a touch screen, a touchpad, a keyboard, a mouse, a camera, a microphone, an accelerometer, a gyroscope, etc.; an output device 507 including, for example, a liquid crystal display (LCD), a speaker, a vibrator, etc.; a storage device 508 including, for example, a magnetic tape, a hard disk, etc.; and a communication device 509. The communication device 509 may allow the electronic device 500 to communicate with other devices wirelessly or wiredly to exchange data. Although Figure 5 the electronic device 500 with various devices is shown, it should be understood that it is not required to implement or have all the shown devices. Instead, more or fewer devices may be implemented or had.

[0196] In particular, according to an embodiment of the present disclosure, the processes described above with reference to the flowcharts can be implemented as computer software programs. For example, an embodiment of the present disclosure includes a computer program product that includes a computer program carried on a computer-readable medium, and the computer program includes program code for performing the method shown in the flowchart. In such an embodiment, the computer program can be downloaded and installed from the network via the communication device 509, or installed from the storage device 508, or installed from the ROM 502. When the computer program is executed by the processing device 501, the above-described functions defined in the method of the embodiment of the present disclosure are performed.

[0197] It should be noted that the above-mentioned computer-readable medium in the present disclosure can be a computer-readable signal medium or a computer-readable storage medium or any combination of the two. A computer-readable storage medium can be, for example, but not limited to, an electrical, magnetic, optical, electromagnetic, infrared, or semiconductor system, apparatus, or device, or any combination of the above. More specific examples of the computer-readable storage medium can include, but are not limited to: an electrical connection having one or more wires, a portable computer disk, a hard disk, a random access memory (RAM), a read-only memory (ROM), an erasable programmable read-only memory (EPROM or flash memory), an optical fiber, a portable compact disk read-only memory (CD-ROM), an optical storage device, a magnetic storage device, or any suitable combination of the above. In the present disclosure, a computer-readable storage medium can be any tangible medium that contains or stores a program that can be used by or in conjunction with an instruction execution system, apparatus, or device. In the present disclosure, a computer-readable signal medium can include a data signal propagated in a baseband or as part of a carrier wave, which carries computer-readable program code. Such a propagated data signal can take various forms, including but not limited to electromagnetic signals, optical signals, or any suitable combination of the above. A computer-readable signal medium can also be any computer-readable medium other than a computer-readable storage medium, which can send, propagate, or transmit a program for use by or in conjunction with an instruction execution system, apparatus, or device. The program code contained on the computer-readable medium can be transmitted by any appropriate medium, including but not limited to: wires, optical cables, RF (radio frequency), etc., or any suitable combination of the above.

[0198] The above-mentioned computer-readable medium can be included in the above-mentioned electronic device; or it can exist separately without being assembled into the electronic device.

[0199] The above-mentioned computer-readable medium carries one or more programs, and when the one or more programs are executed by the electronic device, the electronic device is caused to perform the method shown in the above embodiment.

[0200] Computer program code for performing the operations of the present disclosure may be written in one or more programming languages or combinations thereof. The above-mentioned programming languages include object-oriented programming languages such as Java, Smalltalk, C++, and also include conventional procedural programming languages such as the "C" language or similar programming languages. The program code may be executed entirely on the user's computer, partially on the user's computer, executed as a stand-alone software package, partially on the user's computer and partially on a remote computer, or entirely on a remote computer or server. In the case of a remote computer, the remote computer may be connected to the user's computer through any type of network, including a local area network (LAN) or a wide area network (WAN), or may be connected to an external computer (e.g., through the Internet using an Internet service provider).

[0201] The flowcharts and block diagrams in the accompanying drawings illustrate the possible architectures, functions, and operations of systems, methods, and computer program products according to various embodiments of the present disclosure. In this regard, each block in the flowchart or block diagram may represent a module, a segment of a program, or a part of code that contains one or more executable instructions for implementing a specified logical function. It should also be noted that in some alternative implementations, the functions noted in the blocks may occur in an order different from that noted in the accompanying drawings. For example, two consecutive blocks shown may actually be executed substantially in parallel, or they may sometimes be executed in the reverse order, depending on the functions involved. It should also be noted that each block in the block diagram and / or flowchart, and combinations of blocks in the block diagram and / or flowchart, may be implemented by a dedicated hardware-based system that performs the specified functions or operations, or may be implemented by a combination of dedicated hardware and computer instructions.

[0202] The units involved in the embodiments described in the present disclosure may be implemented in software or in hardware. Among them, the name of the unit does not constitute a limitation on the unit itself in some cases. For example, the first acquisition unit may also be described as "the unit for acquiring at least two Internet protocol addresses".

[0203] The functions described above herein may be performed at least in part by one or more hardware logic components. For example, without limitation, exemplary types of hardware logic components that may be used include: field programmable gate arrays (FPGA), application specific integrated circuits (ASIC), application specific standard products (ASSP), system on a chip (SOC), complex programmable logic devices (CPLD), and so on.

[0204] In a first aspect, according to one or more embodiments of the present disclosure, a resource scheduling method based on an elastic block storage service is provided, including:

[0205] Obtain the attribute information and operation event information of each cloud disk of a target tenant in a tenant set;

[0206] Generate a set of candidate cloud disk behavior labels for the target tenant according to the attribute information and operation event information of each cloud disk;

[0207] According to the cloud disk behavior changes of the target tenant and the remaining tenants in the tenant set, determine the current weight of each candidate cloud disk behavior label in the set of candidate cloud disk behavior labels, and according to the current weight of each candidate cloud disk behavior label, determine the target cloud disk behavior label of the target tenant from the set of candidate cloud disk behavior labels; generate tagged feature information of the target tenant according to the target cloud disk behavior label;

[0208] Schedule the cloud disks of the target tenant according to the tagged feature information of the target tenant.

[0209] According to one or more embodiments of the present disclosure, the determining the current weight of each candidate cloud disk behavior label in the set of candidate cloud disk behavior labels according to the cloud disk behavior changes of the target tenant and the remaining tenants in the tenant set includes:

[0210] Based on the target cloud disk behavior labels of all tenants in the tenant set, determine the real-time correlation coefficient between any candidate cloud disk behavior label in the set of candidate cloud disk behavior labels and the target tenant;

[0211] According to the cloud disk behavior change of the target tenant, determine the time decay coefficient of the any candidate cloud disk behavior label;

[0212] Determine the current weight of the any candidate cloud disk behavior label according to the real-time correlation coefficient and the time decay characteristic coefficient.

[0213] According to one or more embodiments of the present disclosure, the determining the current weight of the any candidate cloud disk behavior label according to the real-time correlation coefficient and the time decay characteristic coefficient includes:

[0214] Obtain the preset initial weight of the any candidate cloud disk behavior label;

[0215] Update the preset initial weight according to the real-time correlation coefficient and the time decay characteristic coefficient to obtain the current weight of the any candidate cloud disk behavior label.

[0216] According to one or more embodiments of the present disclosure, determining the real-time relevance coefficient between any candidate cloud disk behavior label in the candidate cloud disk behavior label set and the target tenant based on the target cloud disk behavior labels of all tenants in the tenant set includes:

[0217] Determine the occurrence frequency of any candidate cloud disk behavior label in the candidate cloud disk behavior label set as the word frequency of the any candidate cloud disk behavior label;

[0218] Determine the reciprocal of the occurrence frequency of any candidate cloud disk behavior label in the target cloud disk behavior labels of all tenants in the tenant set as the inverse document frequency of the any candidate cloud disk behavior label;

[0219] Adopt the term frequency-inverse document frequency method to determine the real-time relevance coefficient between any candidate cloud disk behavior label and the target tenant according to the word frequency and inverse document frequency of the any candidate cloud disk behavior label.

[0220] According to one or more embodiments of the present disclosure, generating the candidate cloud disk behavior label set of the target tenant according to the attribute information and operation event information of each cloud disk includes:

[0221] Generate a single cloud disk behavior label corresponding to each cloud disk according to the attribute information and operation event information of each cloud disk; and generate a multi-cloud disk behavior label of the target tenant according to the single cloud disk behavior label corresponding to each cloud disk; and / or

[0222] Generate a tenant feature label of the target tenant according to the attribute information and operation event information of each cloud disk;

[0223] Determine the tenant feature label and / or the multi-cloud disk behavior label as the candidate cloud disk behavior label of the target tenant to form the candidate cloud disk behavior label set.

[0224] According to one or more embodiments of the present disclosure, generating a single cloud disk behavior label corresponding to each cloud disk according to the attribute information and operation event information of each cloud disk includes:

[0225] Obtain the cumulative amount of the preset cloud disk behavior metrics of each cloud disk within the target time window according to the attribute information and / or operation event information of each cloud disk, and determine the single cloud disk behavior label corresponding to each cloud disk according to the cumulative amount of the preset cloud disk behavior metrics; and / or

[0226] The generating the tenant feature label of the target tenant according to the attribute information and operation event information of each cloud disk includes:

[0227] Obtain the cumulative amount of preset tenant characteristic metrics of the target tenant within a target time window according to the attribute information and / or operation event information of all the cloud disks of the target tenant, and determine the tenant characteristic label of the target tenant according to the cumulative amount of the preset tenant characteristic metrics.

[0228] According to one or more embodiments of the present disclosure, generating a single-cloud-disk behavior label corresponding to each cloud disk according to the attribute information and operation event information of each cloud disk includes:

[0229] Judge whether the attribute information and / or operation event information of any cloud disk satisfy a first preset rule corresponding to the single-cloud-disk behavior label. If so, generate a single-cloud-disk behavior label for the any cloud disk; and / or

[0230] Generating the tenant characteristic label of the target tenant according to the attribute information and operation event information of each cloud disk includes:

[0231] Judge whether the attribute information and / or operation event information of all the cloud disks of the target tenant satisfy a second preset rule corresponding to the tenant characteristic label. If so, generate a tenant characteristic label for the target tenant.

[0232] According to one or more embodiments of the present disclosure, the first preset rule and / or the second preset rule is a rule index for judging whether it exceeds a corresponding preset index threshold;

[0233] Correspondingly, the method further includes:

[0234] Determine an adjustment amount of the preset index threshold according to the total resource amount, the used resource amount, and the predicted used resource amount of the cluster to which the cloud disk of the target user belongs, and adjust the preset index threshold according to the adjustment amount.

[0235] According to one or more embodiments of the present disclosure, generating a single-cloud-disk behavior label corresponding to each cloud disk according to the attribute information and operation event information of each cloud disk includes:

[0236] Obtain historical data of the cloud disk behavior metrics to be predicted for each cloud disk within a historical time window according to the attribute information and / or operation event information of each cloud disk. Call a prediction model according to the time series historical data of the cloud disk behavior metrics to be predicted, obtain the predicted value of the cloud disk behavior metrics to be predicted, and determine the single-cloud-disk behavior label of each cloud disk according to the predicted value of the cloud disk behavior metrics to be predicted; and / or

[0237] Generating the tenant characteristic label of the target tenant according to the attribute information and operation event information of each cloud disk includes:

[0238] Obtain historical data of the tenant characteristic metrics to be predicted for the target tenant within a historical time window according to the attribute information and / or operation event information of all the cloud disks of the target tenant. Invoke a prediction model based on the historical data of the tenant characteristic metrics to be predicted to obtain a predicted value of the tenant characteristic metrics to be predicted. Determine the tenant characteristic label of the target tenant according to the predicted value of the tenant characteristic metrics to be predicted.

[0239] According to one or more embodiments of the present disclosure, determining the target cloud disk behavior label of the target tenant from the candidate cloud disk behavior label set includes:

[0240] Screen out associated label groups from the candidate cloud disk behavior label set according to a preset associated label set; wherein the preset associated label set includes multiple groups of associated label groups composed of mutually associated cloud disk behavior labels;

[0241] For each group of screened associated label groups, only retain one of the cloud disk behavior labels for label de-duplication.

[0242] According to one or more embodiments of the present disclosure, before screening out the associated label groups from the candidate cloud disk behavior label set according to the preset associated label set, it further includes:

[0243] Obtain the candidate cloud disk behavior label sets of multiple tenants, and construct a co-occurrence matrix of the candidate cloud disk behaviors according to the candidate cloud disk behavior label sets of the multiple tenants;

[0244] Obtain the similarity between the co-occurring candidate cloud disk behavior labels in the co-occurrence matrix. If the similarity between any two candidate cloud disk behavior labels exceeds a preset similarity threshold, determine the any two candidate cloud disk behavior labels as a group of associated label groups;

[0245] Construct the preset associated label set according to the determined associated label groups.

[0246] According to one or more embodiments of the present disclosure, the method further includes:

[0247] Determine the fluctuation condition of the labeled characteristic information of the target tenant. If the fluctuation condition is abnormal, issue an alarm.

[0248] In a second aspect, according to one or more embodiments of the present disclosure, a resource scheduling device based on an elastic block storage service is provided, including:

[0249] An obtaining unit, configured to obtain the attribute information and operation event information of each cloud disk of a target tenant in a tenant set;

[0250] A label generation unit, configured to generate a set of candidate cloud disk behavior labels for the target tenant according to the attribute information and operation event information of each cloud disk; determine the current weight of each candidate cloud disk behavior label in the set of candidate cloud disk behavior labels according to the cloud disk behavior changes of the target tenant and the remaining tenants in the tenant set, and determine the target cloud disk behavior label of the target tenant from the set of candidate cloud disk behavior labels according to the current weight of each candidate cloud disk behavior label;

[0251] A feature information generation unit, configured to generate labeled feature information of the target tenant according to the target cloud disk behavior label;

[0252] A scheduling unit, configured to schedule the cloud disk of the target tenant according to the labeled feature information of the target tenant.

[0253] According to one or more embodiments of the present disclosure, when determining the current weight of each candidate cloud disk behavior label in the set of candidate cloud disk behavior labels according to the cloud disk behavior changes of the target tenant and the remaining tenants in the tenant set, the label generation unit is configured to:

[0254] Based on the target cloud disk behavior labels of all tenants in the tenant set, determine the real-time relevance coefficient between any candidate cloud disk behavior label in the set of candidate cloud disk behavior labels and the target tenant;

[0255] According to the cloud disk behavior change of the target tenant, determine the time decay coefficient of the any candidate cloud disk behavior label;

[0256] Determine the current weight of the any candidate cloud disk behavior label according to the real-time relevance coefficient and the time decay characteristic coefficient.

[0257] According to one or more embodiments of the present disclosure, when determining the current weight of any candidate cloud disk behavior label according to the real-time relevance coefficient and the time decay characteristic coefficient, the label generation unit is configured to:

[0258] Obtain the preset initial weight of the any candidate cloud disk behavior label;

[0259] Update the preset initial weight according to the real-time relevance coefficient and the time decay characteristic coefficient to obtain the current weight of the any candidate cloud disk behavior label.

[0260] According to one or more embodiments of the present disclosure, when determining the real-time relevance coefficient between any candidate cloud disk behavior label in the set of candidate cloud disk behavior labels and the target tenant based on the target cloud disk behavior labels of all tenants in the tenant set, the label generation unit is configured to:

[0261] Determine the occurrence frequency of any one of the candidate cloud disk behavior tags in the candidate cloud disk behavior tag set as the word frequency of the any one of the candidate cloud disk behavior tags;

[0262] Determine the reciprocal of the occurrence frequency of any one of the candidate cloud disk behavior tags in the target cloud disk behavior tags of all tenants in the tenant set as the inverse document frequency of the any one of the candidate cloud disk behavior tags;

[0263] Adopt the term frequency - inverse document frequency method to determine the real - time relevance coefficient between any one of the candidate cloud disk behavior tags and the target tenant according to the word frequency and inverse document frequency of the any one of the candidate cloud disk behavior tags.

[0264] According to one or more embodiments of the present disclosure, when generating the candidate cloud disk behavior tag set of the target tenant according to the attribute information and operation event information of each cloud disk, the tag generation unit is configured to:

[0265] Generate a single cloud disk behavior tag corresponding to each cloud disk according to the attribute information and operation event information of each cloud disk; and generate a multi - cloud disk behavior tag of the target tenant according to the single cloud disk behavior tag corresponding to each cloud disk; and / or

[0266] Generate a tenant feature tag of the target tenant according to the attribute information and operation event information of each cloud disk;

[0267] Determine the tenant feature tag and / or the multi - cloud disk behavior tag as the candidate cloud disk behavior tag of the target tenant, and form the candidate cloud disk behavior tag set.

[0268] According to one or more embodiments of the present disclosure, when generating a single cloud disk behavior tag corresponding to each cloud disk according to the attribute information and operation event information of each cloud disk, the tag generation unit is configured to:

[0269] Obtain the cumulative amount of the preset cloud disk behavior metrics of each cloud disk within the target time window according to the attribute information and / or operation event information of each cloud disk, and determine the single cloud disk behavior tag corresponding to each cloud disk according to the cumulative amount of the preset cloud disk behavior metrics; and / or

[0270] When generating the tenant feature tag of the target tenant according to the attribute information and operation event information of each cloud disk, the tag generation unit is configured to:

[0271] Obtain the cumulative amount of the preset tenant characteristic metrics of the target tenant within the target time window according to the attribute information and / or operation event information of all cloud disks of the target tenant, and determine the tenant feature tag of the target tenant according to the cumulative amount of the preset tenant characteristic metrics.

[0272] According to one or more embodiments of the present disclosure, when generating a single disk behavior label corresponding to each disk based on the attribute information and operation event information of each disk, the label generation unit is configured to:

[0273] Determine whether the attribute information and / or operation event information of any disk satisfies a first preset rule corresponding to the single disk behavior label. If so, generate a single disk behavior label for the any disk; and / or

[0274] When generating a tenant feature label of the target tenant based on the attribute information and operation event information of each disk, the label generation unit is configured to:

[0275] Determine whether the attribute information and / or operation event information of all disks of the target tenant satisfies a second preset rule corresponding to the tenant feature label. If so, generate a tenant feature label for the target tenant.

[0276] According to one or more embodiments of the present disclosure, the first preset rule and / or the second preset rule is a rule for determining whether a relevant rule index exceeds a corresponding preset index threshold;

[0277] Correspondingly, the label generation unit is further configured to:

[0278] Determine an adjustment amount of the preset index threshold according to the total resource amount, the used resource amount, and the predicted used resource amount of the cluster to which the disks of the target user belong, and adjust the preset index threshold according to the adjustment amount.

[0279] According to one or more embodiments of the present disclosure, when generating a single disk behavior label corresponding to each disk based on the attribute information and operation event information of each disk, the label generation unit is configured to:

[0280] Obtain historical data of the to-be-predicted disk behavior index of each disk within a historical time window according to the attribute information and / or operation event information of each disk. Call a prediction model according to the time series historical data of the to-be-predicted disk behavior index, obtain a predicted value of the to-be-predicted disk behavior index, and determine a single disk behavior label for each disk according to the predicted value of the to-be-predicted disk behavior index; and / or

[0281] When generating a tenant feature label of the target tenant based on the attribute information and operation event information of each disk, the label generation unit is configured to:

[0282] Obtain historical data of the tenant feature metrics to be predicted for the target tenant within a historical time window according to the attribute information and / or operation event information of all the cloud disks of the target tenant. Invoke a prediction model based on the historical data of the tenant feature metrics to be predicted to obtain a predicted value of the tenant feature metrics to be predicted. Determine the tenant feature label of the target tenant according to the predicted value of the tenant feature metrics to be predicted.

[0283] According to one or more embodiments of the present disclosure, when determining the target cloud disk behavior label of the target tenant from the candidate cloud disk behavior label set, the label generation unit is configured to:

[0284] Screen an associated label group from the candidate cloud disk behavior label set according to a preset associated label set; wherein the preset associated label set includes multiple associated label groups composed of mutually associated cloud disk behavior labels;

[0285] For each screened associated label group, only retain one of the cloud disk behavior labels for label deduplication.

[0286] According to one or more embodiments of the present disclosure, before screening the associated label group from the candidate cloud disk behavior label set according to the preset associated label set, the label generation unit is further configured to:

[0287] Obtain candidate cloud disk behavior label sets of multiple tenants, and construct a co-occurrence matrix of candidate cloud disk behaviors according to the candidate cloud disk behavior label sets of multiple tenants;

[0288] Obtain the similarity between the co-occurring candidate cloud disk behavior labels in the co-occurrence matrix. If the similarity between any two candidate cloud disk behavior labels exceeds a preset similarity threshold, determine the any two candidate cloud disk behavior labels as a group of associated label groups;

[0289] Construct the preset associated label set according to the determined associated label groups.

[0290] According to one or more embodiments of the present disclosure, the feature information generation unit is further configured to:

[0291] Determine the fluctuation condition of the labeled feature information of the target tenant. If the fluctuation condition is abnormal, issue an alarm.

[0292] In a third aspect, according to one or more embodiments of the present disclosure, there is provided an electronic device, including: at least one processor and a memory;

[0293] The memory stores computer execution instructions;

[0294] The at least one processor executes the computer-executable instructions stored in the memory, such that the at least one processor executes the resource scheduling method based on the elastic block storage service as described in the first aspect above and various possible designs of the first aspect.

[0295] In a fourth aspect, according to one or more embodiments of the present disclosure, there is provided a computer-readable storage medium storing computer-executable instructions, which, when executed by a processor, implement the resource scheduling method based on the elastic block storage service as described in the first aspect above and various possible designs of the first aspect.

[0296] In a fifth aspect, according to one or more embodiments of the present disclosure, there is provided a computer program product including a computer program, which, when executed by a processor, implements the resource scheduling method based on the elastic block storage service as described in the first aspect above and various possible designs of the first aspect.

[0297] The above description is only a preferred embodiment of the present disclosure and an explanation of the applied technical principles. Those skilled in the art should understand that the scope of the disclosure involved in the present disclosure is not limited to the technical solution formed by the specific combination of the above technical features, and should also cover other technical solutions formed by any combination of the above technical features or their equivalent features without departing from the above disclosure concept. For example, the technical solutions formed by mutually replacing the above features with the technical features (but not limited to) having similar functions disclosed in the present disclosure.

[0298] In addition, although the operations are depicted in a particular order, this should not be construed as requiring that the operations be performed in the particular order shown or in sequential order. In certain environments, multitasking and parallel processing may be advantageous. Similarly, although a number of specific implementation details are included in the above discussion, these should not be construed as limiting the scope of the present disclosure. Certain features described in the context of separate embodiments may also be implemented combinatorially in a single embodiment. Conversely, the various features described in the context of a single embodiment may also be implemented separately or in any suitable sub-combination in multiple embodiments.

[0299] Although the subject matter has been described in language specific to structural features and / or methodological act logic, it should be understood that the subject matter defined in the appended claims is not necessarily limited to the specific features or acts described above. Rather, the specific features and acts described above are merely example forms for implementing the claims.

Claims

1. A resource scheduling method based on an elastic block storage service, characterized in that, Including: Obtaining the attribute information and operation event information of each cloud disk of a target tenant in a tenant set; Generating a candidate cloud disk behavior label set of the target tenant according to the attribute information and operation event information of each cloud disk; Determining the current weight of each candidate cloud disk behavior label in the candidate cloud disk behavior label set according to the cloud disk behavior changes of the target tenant and the remaining tenants in the tenant set, and determining the target cloud disk behavior label of the target tenant from the candidate cloud disk behavior label set according to the current weight of each candidate cloud disk behavior label; Generating labeled feature information of the target tenant according to the target cloud disk behavior label; Scheduling the cloud disks of the target tenant according to the labeled feature information of the target tenant.

2. The method according to claim 1, characterized in that The determining the current weight of each candidate cloud disk behavior label in the candidate cloud disk behavior label set according to the cloud disk behavior changes of the target tenant and the remaining tenants in the tenant set includes: Based on the target cloud disk behavior labels of all tenants in the tenant set, determining the real-time relevance coefficient between any candidate cloud disk behavior label in the candidate cloud disk behavior label set and the target tenant; Determining the time decay coefficient of the any candidate cloud disk behavior label according to the cloud disk behavior change of the target tenant; Determining the current weight of the any candidate cloud disk behavior label according to the real-time relevance coefficient and the time decay characteristic coefficient.

3. The method according to claim 2, wherein The determining the current weight of the any candidate cloud disk behavior label according to the real-time relevance coefficient and the time decay characteristic coefficient includes: Obtaining the preset initial weight of the any candidate cloud disk behavior label; Updating the preset initial weight according to the real-time relevance coefficient and the time decay characteristic coefficient to obtain the current weight of the any candidate cloud disk behavior label.

4. The method according to claim 2, wherein The determining the real-time relevance coefficient between any candidate cloud disk behavior label in the candidate cloud disk behavior label set and the target tenant based on the target cloud disk behavior labels of all tenants in the tenant set includes: Determining the occurrence frequency of the any candidate cloud disk behavior label in the candidate cloud disk behavior label set as the word frequency of the any candidate cloud disk behavior label; Determining the reciprocal of the occurrence frequency of the any candidate cloud disk behavior label in the target cloud disk behavior labels of all tenants in the tenant set as the inverse document frequency of the any candidate cloud disk behavior label; Using the term frequency-inverse document frequency method, determining the real-time relevance coefficient between the any candidate cloud disk behavior label and the target tenant according to the word frequency and inverse document frequency of the any candidate cloud disk behavior label.

5. The method according to any one of claims 1-4, characterized in that, The generating a candidate cloud disk behavior label set of the target tenant according to the attribute information and operation event information of each cloud disk includes: Generating a single cloud disk behavior label corresponding to each cloud disk according to the attribute information and operation event information of each cloud disk; and generating a multi-cloud disk behavior label of the target tenant according to the single cloud disk behavior label corresponding to each cloud disk; and / or Generating a tenant feature label of the target tenant according to the attribute information and operation event information of each cloud disk. Determine the tenant feature label and / or the multi-cloud disk behavior label as the candidate cloud disk behavior label of the target tenant, and form a set of candidate cloud disk behavior labels.

6. The method according to claim 5, wherein Generating a single cloud disk behavior label corresponding to each cloud disk according to the attribute information and operation event information of each cloud disk includes: Obtaining the cumulative amount of the preset cloud disk behavior indicators of each cloud disk within the target time window according to the attribute information and / or operation event information of each cloud disk, and determining the single cloud disk behavior label corresponding to each cloud disk according to the cumulative amount of the preset cloud disk behavior indicators; and / or Generating a tenant feature label of the target tenant according to the attribute information and operation event information of each cloud disk includes: Obtaining the cumulative amount of the preset tenant characteristic indicators of the target tenant within the target time window according to the attribute information and / or operation event information of all the cloud disks of the target tenant, and determining the tenant feature label of the target tenant according to the cumulative amount of the preset tenant characteristic indicators.

7. The method according to claim 5, characterized in that, Generating a single cloud disk behavior label corresponding to each cloud disk according to the attribute information and operation event information of each cloud disk includes: Judging whether the attribute information and / or operation event information of any cloud disk meet the first preset rule corresponding to the single cloud disk behavior label. If so, generating a single cloud disk behavior label for the any cloud disk; and / or Generating a tenant feature label of the target tenant according to the attribute information and operation event information of each cloud disk includes: Judging whether the attribute information and / or operation event information of all the cloud disks of the target tenant meet the second preset rule corresponding to the tenant feature label. If so, generating a tenant feature label for the target tenant.

8. The method according to claim 7, characterized in that, The first preset rule and / or the second preset rule is a rule for judging whether the relevant rule indicators exceed the corresponding preset indicator thresholds; Correspondingly, the method further includes: Determining an adjustment amount of the preset indicator threshold according to the total resource amount, the used resource amount, and the predicted used resource amount of the cloud disk cluster to which the cloud disk of the target user belongs, and adjusting the preset indicator threshold according to the adjustment amount.

9. The method according to claim 2, characterized in that, Generating a single cloud disk behavior label corresponding to each cloud disk according to the attribute information and operation event information of each cloud disk includes: Obtaining the historical data of the cloud disk behavior indicators to be predicted for each cloud disk within the historical time window according to the attribute information and / or operation event information of each cloud disk, calling a prediction model according to the time series historical data of the cloud disk behavior indicators to be predicted, obtaining the predicted value of the cloud disk behavior indicators to be predicted, and determining the single cloud disk behavior label of each cloud disk according to the predicted value of the cloud disk behavior indicators to be predicted; and / or Generating a tenant feature label of the target tenant according to the attribute information and operation event information of each cloud disk includes: Obtaining the historical data of the tenant characteristic indicators to be predicted for the target tenant within the historical time window according to the attribute information and / or operation event information of all the cloud disks of the target tenant, calling a prediction model according to the historical data of the tenant characteristic indicators to be predicted, obtaining the predicted value of the tenant characteristic indicators to be predicted, and using the predicted value of the tenant characteristic indicators to be predicted as the tenant feature label of the target tenant.

10. The method according to any one of claims 1 to 4, characterized in that Determining the target cloud disk behavior label of the target tenant from the candidate cloud disk behavior label set includes: Filtering associated label groups from the candidate cloud disk behavior label set according to a preset associated label set; wherein the preset associated label set includes multiple groups of associated label groups composed of mutually associated cloud disk behavior labels; For each group of filtered associated label groups, only one cloud disk behavior label is retained for label deduplication.

11. The method according to claim 10, wherein Before filtering the associated label groups from the candidate cloud disk behavior label set according to the preset associated label set, it further includes: Obtaining the candidate cloud disk behavior label sets of multiple tenants, and constructing a co-occurrence matrix of the candidate cloud disk behavior labels according to the candidate cloud disk behavior label sets of the multiple tenants; Obtaining the similarity between the co-occurring candidate cloud disk behavior labels in the co-occurrence matrix, and if the similarity between any two candidate cloud disk behavior labels exceeds a preset similarity threshold, determining the any two candidate cloud disk behavior labels as a group of associated label groups; Constructing the preset associated label set according to the determined associated label groups.

12. The method according to claim 1, wherein The method further includes: Determining the fluctuation condition of the labeled feature information of the target tenant, and if the fluctuation condition is abnormal, issuing an alarm.

13. A resource scheduling device based on an elastic block storage service, characterized in that It includes: An acquisition unit for acquiring the attribute information and operation event information of each cloud disk of the target tenant in the tenant set; A label generation unit for generating the candidate cloud disk behavior label set of the target tenant according to the attribute information and operation event information of each cloud disk; determining the current weight of each candidate cloud disk behavior label in the candidate cloud disk behavior label set according to the cloud disk behavior changes of the target tenant and the other tenants in the tenant set, and determining the target cloud disk behavior label of the target tenant from the candidate cloud disk behavior label set according to the current weight of each candidate cloud disk behavior label; A feature information generation unit for generating the labeled feature information of the target tenant according to the target cloud disk behavior label; A scheduling unit for scheduling the cloud disk of the target tenant according to the labeled feature information of the target tenant.

14. An electronic device, characterized in that, It includes: A processor and a memory; The memory stores computer execution instructions; The processor executes the computer execution instructions stored in the memory, so that the processor executes the method according to any one of claims 1-12.

15. A computer-readable storage medium, characterized in that, The computer-readable storage medium stores computer execution instructions, and when the processor executes the computer execution instructions, the method according to any one of claims 1-12 is implemented.

16. A computer program product comprising a computer program, characterized in that, The computer program, when executed by the processor, implements the method according to any one of claims 1-12.

Images