Resource scheduling method and apparatus
By obtaining user identifiers and resource authorization credentials, and combining behavioral characteristics and node status information, scheduling priorities are determined and resource scheduling strategies are generated. This solves the problem of resource permission continuation when users access resources across nodes, and achieves dynamic adaptation and efficient allocation of resource supply and demand.
Patent Information
- Application Number
- CN202610980135.1
- Authority / Receiving Office
- CN · China
- Patent Type
- Applications(China)
- Current Assignee / Owner
- Filing Date
- 2026-07-02
- Publication Date
- 2026-11-17
AI Technical Summary
In distributed cloud computing service scenarios, resource permissions cannot be continued when users access resources across nodes, and resource supply and demand are prone to mismatch.
By obtaining user identifiers and bound resource authorization credentials, and combining the target user's behavioral characteristics with the resource status information of cloud service nodes, scheduling priorities are determined and resource scheduling policies are generated, enabling universal adaptation of resource permissions across multiple cloud service nodes.
It ensures the effective continuation of resource permissions when users access resources across nodes, adapts to real-time changes in resource supply and demand, and improves the accuracy of resource allocation and system response efficiency.
Smart Images

Figure CN123114723A_ABST
Abstract
Description
Technical Field
[0001] This disclosure relates to the field of communication technology, and in particular to a resource scheduling method and apparatus. Background Technology
[0002] In distributed cloud computing service scenarios, computing resources are typically provided to users by multiple geographically distributed service nodes. In related technologies, computing resources are mostly scheduled and managed independently by each service node; however, when users access resources across nodes, resource permissions cannot be continued, and resource supply and demand are prone to mismatch. Summary of the Invention
[0003] This disclosure provides a resource scheduling method and apparatus to address the problem in related technologies where computing resources are mostly scheduled and managed independently by each service node; however, when users access resources across nodes, resource permissions cannot be continued, and resource supply and demand are prone to mismatch.
[0004] According to a first aspect of this disclosure, a resource scheduling method is provided, comprising:
[0005] Obtain a user identifier and a resource authorization credential bound to the user identifier; wherein the resource authorization credential is used to indicate the available resources of the target user corresponding to the user identifier on each cloud service node; Based on the behavioral characteristics of the target user and the resource status information of each cloud service node, the scheduling priority of the target user for each cloud service node is determined. Based on the resource authorization credentials and the scheduling priority, a resource scheduling policy corresponding to the target user is generated.
[0006] In some embodiments of this disclosure, prior to obtaining the user identifier, the method further includes: Determine the resource metering parameters of the target device used by the target user; wherein the resource metering parameters are determined based on the performance parameters, usage duration and task type of the target device, the performance parameters are used to indicate the performance of the target device relative to the reference device, and the task type is used to indicate the degree of consumption of computing resources by the executed task; Generate the user identifier corresponding to the target user, and configure the resource authorization credential bound to the user identifier; wherein, the available resources in the resource authorization credential are determined based on the resource metering parameters.
[0007] In some embodiments of this disclosure, after obtaining the user identifier and the resource authorization credentials bound to the user identifier, the method further includes: When the target device is a single unit, the resource consumption corresponding to the task executed by the target user is determined based on the performance parameters, usage duration, task type, number of parallel tasks, maximum parallel capacity and load factor of the target device, and the resource consumption is removed from the resource authorization credentials. If it is determined that there are at least two target devices and the target user is performing at least two tasks, the total resource consumption corresponding to the tasks performed by the target user is determined based on the performance parameters of the target devices, the usage duration, the task type, and the resource conflict correction amount, and the total resource consumption is removed from the resource authorization credentials; wherein, the resource conflict correction amount is used to reflect the additional resource consumption caused by resource contention and communication loss during the operation of at least two target devices.
[0008] In some embodiments of this disclosure, before determining the scheduling priority of the target user for each cloud service node based on the behavioral characteristics of the target user and the resource status information of each cloud service node, the method further includes: Obtain the target user's historical and real-time behavior data; The historical behavior data and the real-time behavior data are weighted and fused to generate a comprehensive behavior feature vector; The comprehensive behavioral feature vector is input into a pre-trained classification and prediction network model to determine the behavioral label and behavioral correction coefficient of the target user; wherein, the behavioral features include the behavioral label and the behavioral correction coefficient, the behavioral label is used for the active time period characteristics of the target user, and the behavioral correction coefficient is used to adjust the scheduling priority parameter.
[0009] In some embodiments of this disclosure, determining the scheduling priority of the target user for each cloud service node based on the behavioral characteristics of the target user and the resource status information of each cloud service node includes: The scheduling priority parameter is determined based on the resource metering parameters of the target device, the resource occupancy rate of each cloud service node, the heat coefficient determined by the user's resource request heat information, and the behavior correction coefficient in the behavior characteristics; wherein, the resource status information includes the resource occupancy rate and the heat coefficient.
[0010] In some embodiments of this disclosure, generating a resource scheduling policy corresponding to the target user based on the resource authorization credentials and the scheduling priority includes: The resource authorization status of the target user is determined based on the available resources in the resource authorization credentials and the resource consumption required by the target user; Based on the resource authorization status and the scheduling priority, determine the resource scheduling strategy corresponding to the resource request of the target user.
[0011] In some embodiments of this disclosure, determining the resource scheduling strategy corresponding to the resource request of the target user based on the resource authorization status and the scheduling priority includes: Determine the resource consumption required for the task performed by the target user; When the resource authorization status is in the first state and the scheduling priority is greater than the first preset threshold, the resource request of the target user is allocated to the first cloud service node; When the resource authorization status is in the second state and / or the scheduling priority is less than the second preset threshold, the resource request of the target user is rate-limited or allocated to the second cloud service node; wherein, the first state is when the available resources in the resource authorization credentials are greater than or equal to the resource consumption required by the task performed by the target user, the second state is when the available resources in the resource authorization credentials are less than the resource consumption required by the task performed by the target user, the performance of the first cloud service node is higher than that of the second cloud service node, and the first preset threshold is greater than the second preset threshold.
[0012] In some embodiments of this disclosure, after generating the resource scheduling policy corresponding to the target user based on the resource authorization credentials and the scheduling priority, the method further includes: Execute the resource scheduling strategy and store user behavior data, resource allocation data, and scheduling result data during the execution process; Based on the stored user behavior data, resource allocation data, and scheduling result data, update the method for determining the scheduling priority parameter and / or the method for generating the resource scheduling strategy.
[0013] In some embodiments of this disclosure, the method further includes: In response to the target user switching between different cloud service nodes, the same resource configuration as the cloud service node before the switch is restored on the switched cloud service node according to the resource authorization credentials and the scheduling priority parameters.
[0014] According to a second aspect of this disclosure, a resource scheduling apparatus is provided, comprising: An acquisition unit is used to acquire a user identifier and a resource authorization credential bound to the user identifier; wherein the resource authorization credential is used to indicate the available resources of the target user corresponding to the user identifier on each cloud service node; The determining unit is used to determine the scheduling priority of the target user for each cloud service node based on the behavioral characteristics of the target user and the resource status information of each cloud service node. The generation unit is used to generate a resource scheduling policy corresponding to the target user based on the resource authorization credentials and the scheduling priority.
[0015] In some embodiments of this disclosure, the apparatus is further configured to: Before the acquisition unit acquires the user identifier, the resource metering parameters of the target device used by the target user are determined; wherein, the resource metering parameters are determined based on the performance parameters, usage duration and task type of the target device, the performance parameters are used to indicate the performance of the target device relative to the reference device, and the task type is used to indicate the degree of consumption of computing resources by the executed task; Generate the user identifier corresponding to the target user, and configure the resource authorization credential bound to the user identifier; wherein, the available resources in the resource authorization credential are determined based on the resource metering parameters.
[0016] In some embodiments of this disclosure, the apparatus is further configured to: After the acquisition unit acquires the user identifier and the resource authorization credential bound to the user identifier, when the target device is a single unit, the resource consumption corresponding to the task executed by the target user is determined according to the performance parameters, usage duration, task type, number of parallel tasks, maximum parallel capacity and load coefficient of the target device, and the resource consumption is removed from the resource authorization credential. If it is determined that there are at least two target devices and the target user is performing at least two tasks, the total resource consumption corresponding to the tasks performed by the target user is determined based on the performance parameters of the target devices, the usage duration, the task type, and the resource conflict correction amount, and the total resource consumption is removed from the resource authorization credentials; wherein, the resource conflict correction amount is used to reflect the additional resource consumption caused by resource contention and communication loss during the operation of at least two target devices.
[0017] In some embodiments of this disclosure, the apparatus is further configured to: Before the determining unit determines the scheduling priority of the target user for each cloud service node based on the target user's behavioral characteristics and the resource status information of each cloud service node, the historical behavioral data and real-time behavioral data of the target user are obtained. The historical behavior data and the real-time behavior data are weighted and fused to generate a comprehensive behavior feature vector; The comprehensive behavioral feature vector is input into a pre-trained classification and prediction network model to determine the behavioral label and behavioral correction coefficient of the target user; wherein, the behavioral features include the behavioral label and the behavioral correction coefficient, the behavioral label is used for the active time period characteristics of the target user, and the behavioral correction coefficient is used to adjust the scheduling priority parameter.
[0018] In some embodiments of this disclosure, the determining unit is further configured to: The scheduling priority parameter is determined based on the resource metering parameters of the target device, the resource occupancy rate of each cloud service node, the heat coefficient determined by the user's resource request heat information, and the behavior correction coefficient in the behavior characteristics; wherein, the resource status information includes the resource occupancy rate and the heat coefficient.
[0019] In some embodiments of this disclosure, the generating unit is further configured to: The resource authorization status of the target user is determined based on the available resources in the resource authorization credentials and the resource consumption required by the target user; Based on the resource authorization status and the scheduling priority, determine the resource scheduling strategy corresponding to the resource request of the target user.
[0020] In some embodiments of this disclosure, the generating unit is further configured to: Determine the resource consumption required for the task performed by the target user; When the resource authorization status is in the first state and the scheduling priority is greater than the first preset threshold, the resource request of the target user is allocated to the first cloud service node; When the resource authorization status is in the second state and / or the scheduling priority is less than the second preset threshold, the resource request of the target user is rate-limited or allocated to the second cloud service node; wherein, the first state is when the available resources in the resource authorization credentials are greater than or equal to the resource consumption required by the task performed by the target user, the second state is when the available resources in the resource authorization credentials are less than the resource consumption required by the task performed by the target user, the performance of the first cloud service node is higher than that of the second cloud service node, and the first preset threshold is greater than the second preset threshold.
[0021] In some embodiments of this disclosure, the apparatus further includes: The execution unit is configured to execute the resource scheduling policy after the generation unit generates the resource scheduling policy corresponding to the target user based on the resource authorization credentials and the scheduling priority, and to store user behavior data, resource allocation data and scheduling result data during the execution process. The update unit is used to update the method for determining the scheduling priority parameter and / or the method for generating the resource scheduling strategy based on the stored user behavior data, resource allocation data, and scheduling result data.
[0022] In some embodiments of this disclosure, the apparatus is further configured to: In response to the target user switching between different cloud service nodes, the same resource configuration as the cloud service node before the switch is restored on the switched cloud service node according to the resource authorization credentials and the scheduling priority parameters.
[0023] According to a third aspect of this disclosure, an electronic device is provided, comprising: At least one processor; and A memory communicatively connected to the at least one processor; wherein, The memory stores instructions that can be executed by the at least one processor to enable the at least one processor to perform the method described in the first aspect embodiment.
[0024] According to a fourth aspect of this disclosure, a non-transitory computer-readable storage medium storing computer instructions is provided, wherein the computer instructions are configured to cause the computer to perform the method described in the first aspect of the present disclosure.
[0025] According to a fifth aspect of this disclosure, a computer program product is provided, including a computer program that, when executed by a processor, implements the method described in the first aspect of the preceding embodiments.
[0026] The resource scheduling method and apparatus disclosed herein include: obtaining a user identifier and obtaining a resource authorization credential bound to the user identifier; wherein the resource authorization credential is used to indicate the available resources of the target user corresponding to the user identifier on each cloud service node; determining the scheduling priority of the target user for each cloud service node based on the behavioral characteristics of the target user and the resource status information of each cloud service node; generating a resource scheduling policy corresponding to the target user based on the resource authorization credential and the scheduling priority; enabling universal adaptation of resource permissions across multiple cloud service nodes, ensuring the effective continuation of resource permissions when the user accesses across nodes, and dynamically generating a scheduling policy by combining user behavior and node resource status to adapt to real-time changes in resource supply and demand.
[0027] It should be understood that the description in this section is not intended to identify key or essential features of the embodiments of this application, nor is it intended to limit the scope of this application. Other features of this application will become readily apparent from the following description. Attached Figure Description
[0028] The accompanying drawings are provided to better understand this solution and do not constitute a limitation of this disclosure. Wherein: Figure 1 This is a flowchart illustrating a resource scheduling method provided in an embodiment of the present disclosure; Figure 2 This is a schematic diagram of the structure of a resource scheduling device provided in an embodiment of the present disclosure; Figure 3 This is a schematic diagram of another resource scheduling device provided in an embodiment of the present disclosure; Figure 4 A schematic block diagram of an example electronic device provided for embodiments of this disclosure. Detailed Implementation
[0029] The exemplary embodiments of this disclosure are described below with reference to the accompanying drawings, including various details of the embodiments to aid understanding, and should be considered merely exemplary. Therefore, those skilled in the art will recognize that various changes and modifications can be made to the embodiments described herein without departing from the scope and spirit of this disclosure. Similarly, for clarity and brevity, descriptions of well-known functions and structures are omitted in the following description.
[0030] The resource scheduling method and apparatus of this disclosure are described below with reference to the accompanying drawings.
[0031] Figure 1 This is a flowchart illustrating a resource scheduling method provided in an embodiment of the present disclosure.
[0032] like Figure 1 As shown, the method includes the following steps: Step 101: Obtain the user identifier and the resource authorization credential bound to the user identifier; wherein the resource authorization credential is used to indicate the available resources of the target user corresponding to the user identifier on each cloud service node.
[0033] In some embodiments, the user identifier is a globally unique identifier used to uniformly identify the target user across multiple cloud service nodes. The target user can complete authentication at any cloud service node. The system generates a globally unique user identifier for the target user and generates a resource authorization credential bound to this user identifier in the cloud. The resource authorization credential is a data structure used to record the amount of resources the target user can use in each cloud service node, such as the available computing resource quota. This resource authorization credential corresponds one-to-one with the user identifier. When the target user accesses any cloud service node, the system obtains the corresponding resource authorization credential through the user identifier to determine the available resources for the target user on the current cloud service node. The resource authorization credential is universal across distributed cloud service nodes. Each cloud service node automatically completes service processes such as resource request initialization, current status reporting, behavior recording, and feedback feedback through an integrated software development kit (SDK), ensuring that the identification and use of the resource authorization credential remain consistent across different nodes. It should be noted that the above examples are merely illustrative and do not limit the specific content.
[0034] Through the above methods, target users can obtain consistent identity recognition and resource permission management across multiple cloud service nodes using a unified identity identifier and resource authorization credentials. This avoids cross-node identity incompatibility and resource permission interruption caused by independent user authentication and resource management mechanisms on each node. At the same time, the SDK's automated process enables standardized integration of resource requests and status feedback, improving the system's response efficiency across different nodes.
[0035] Step 102: Determine the scheduling priority of the target user for each cloud service node based on the behavioral characteristics of the target user and the resource status information of each cloud service node.
[0036] In some embodiments, the behavioral characteristics of the target user include feature information reflecting the target user's behavioral state during the service process, such as the target user's activity level, usage duration, operation frequency, and task type preference at different time periods. The resource status information of each cloud service node includes the load status of each node, such as the node's computing resource idleness, resource occupancy, or request popularity. By acquiring the target user's behavioral characteristics and the resource status information of each cloud service node, the system comprehensively assesses the urgency of the target user's resource needs and service preferences on each cloud service node, generating a scheduling priority that reflects the target user's resource usage priority on each cloud service node. Specifically, behavioral characteristics may include behavioral labels and behavioral correction coefficients. Behavioral labels indicate the target user's active time period characteristics, such as the target user being a daytime high-activity type ("Daytime High-Activity Pool") or a nighttime high-activity type ("Nighttime High-Activity Pool"). The behavioral correction coefficient has a value range of [0.3, 1.5] and is used to adjust the scheduling priority parameters. It should be noted that the above examples are merely illustrative and do not limit the specific content.
[0037] In this way, the system can combine the dynamic behavior of the target user with the real-time resource status of each cloud service node to sort the resource needs of the target user on each node, thereby providing a basis for the generation of subsequent resource scheduling strategies.
[0038] Step 103: Generate a resource scheduling policy corresponding to the target user based on the resource authorization credentials and the scheduling priority.
[0039] In some embodiments, the resource authorization credentials include the amount of available resources for the target user on each cloud service node, and the scheduling priority is used to indicate the priority order in which the target user obtains resources on each cloud service node. Based on the amount of available resources in the resource authorization credentials, the system determines whether the target user has the conditions to use the resources of a specific cloud service node, and then, in conjunction with the scheduling priority, determines the resource scheduling policy corresponding to the target user's resource request. The resource scheduling policy is used to indicate how to perform scheduling operations such as resource allocation, rate limiting, or degraded allocation for the target user. The scheduling priority can be determined as follows: based on the resource metering parameters of the target device, the resource occupancy rate of each cloud service node, the heat coefficient determined by the target user's resource request heat information, and the behavior correction coefficient in the behavioral characteristics, the scheduling priority parameter is calculated.
[0040] By using the above methods, the system can comprehensively consider the target user's available resource permissions and the priority of current resource needs, and generate a resource scheduling strategy that matches the target user's actual needs, thereby improving the accuracy of resource allocation.
[0041] In summary, the resource scheduling method provided in this disclosure includes: obtaining a user identifier and obtaining a resource authorization credential bound to the user identifier; wherein the resource authorization credential is used to indicate the available resources of the target user corresponding to the user identifier on each cloud service node; determining the scheduling priority of the target user for each cloud service node based on the behavioral characteristics of the target user and the resource status information of each cloud service node; generating a resource scheduling policy corresponding to the target user based on the resource authorization credential and the scheduling priority; enabling universal adaptation of resource permissions across multiple cloud service nodes, ensuring the effective continuation of resource permissions when users access across nodes, and dynamically generating scheduling policies by combining user behavior and node resource status to adapt to real-time changes in resource supply and demand.
[0042] Optionally, based on any of the above embodiments or optional methods, the present disclosure embodiments may also employ the following method, but not limited thereto: the method includes: determining resource metering parameters of the target device used by the target user; wherein the resource metering parameters are determined based on the performance parameters, usage duration, and task type of the target device, the performance parameters being used to indicate the performance of the target device relative to a reference device, and the task type being used to indicate the degree of computational resource consumption of the executed task; generating the user identifier corresponding to the target user, and configuring the resource authorization credentials bound to the user identifier; wherein the available resources in the resource authorization credentials are determined based on the resource metering parameters.
[0043] In some embodiments, resource measurement parameters are used to quantify the computing resource consumption of different target devices into comparable units of measurement. The performance parameter of the target device is represented by the ratio of its performance score to that of a benchmark device. The benchmark device can be an RTX 4080 graphics card with a performance score of 18000. The resource value generated by one hour of stable operation is defined as 1 cloud core, serving as the benchmark unit for the resource measurement parameter. The ratio of the target device's performance score to the benchmark device's performance score is the relative performance multiple of the target device. Task types include game tasks, graphics rendering tasks, and AI model training tasks. Different task types correspond to different correction coefficients; for example, the correction coefficient for game tasks is 1.0, for graphics rendering tasks it is 1.2, and for AI model training tasks it is 1.5. Resource Measurement Parameters It can be determined using the following formula: ,in, To score the performance of the benchmark equipment, To benchmark the performance of the target device. Resource usage duration (unit: hours). The adjustment factor is for the task type. The resource conversion standards for different target devices and task types are shown in Table 1 below: Table 1
[0044] After determining the resource metering parameters of the target user, the system generates a user identifier corresponding to the target user and configures a resource authorization credential for this user identifier. The available resources in the resource authorization credential are determined based on the aforementioned resource metering parameters; that is, the available resource quota of the target user recorded in the resource authorization credential matches the resource value corresponding to the resource metering parameters. It should be noted that Table 1 above is only an illustrative example and does not limit the specific content.
[0045] In this way, resource authorization credentials can accurately reflect the resource value consumption standards of target users on different target devices, providing a reliable basis for resource metering for subsequent resource scheduling.
[0046] Optionally, based on any of the above embodiments or optional methods, the present disclosure embodiments may also employ the following method, but not limited thereto: When the target device is a single unit, determine the resource consumption corresponding to the task executed by the target user based on the performance parameters, usage duration, task type, number of parallel tasks, maximum parallel capacity, and load coefficient of the target device, and remove the resource consumption from the resource authorization credentials; when it is determined that there are at least two target devices and the task executed by the target user is at least two tasks, determine the total resource consumption corresponding to the task executed by the target user based on the performance parameters, usage duration, task type, and resource conflict correction amount of the target devices, and remove the total resource consumption from the resource authorization credentials; wherein, the resource conflict correction amount is used to reflect the additional resource consumption caused by resource contention and communication loss during the operation of at least two target devices.
[0047] In some embodiments, when the target device is a single unit, the resource consumption is determined as follows: The basic resource consumption (i.e., the aforementioned resource metering parameters) is determined based on the target device's performance parameters, usage duration, and task type. ); Number of parallel tasks Indicates the number of tasks currently running concurrently on the target device, and the maximum parallel capacity. Indicates the maximum number of parallel tasks supported by the target device, and the load factor. Indicates the load status of the cloud service node where the target device resides (e.g., changes in resource scheduling efficiency under different load peaks and valleys). Resource consumption. It can be determined using the following formula: ,in, This represents the current number of parallel tasks. The maximum parallel capacity supported by the target device (e.g., RTX 4090). =3). When there are multiple target devices and at least two tasks are being executed (e.g., NVIDIA parallel large model rendering scene), the total resource consumption is... It can be determined using the following formula: ,in, This represents the base cloud core value of the k-th type of resource (e.g., the k-th graphics card), which is the theoretical resource consumption under conflict-free, full-load operation conditions. This is the resource conflict penalty coefficient, used to reflect the additional resource consumption caused by resource contention and communication loss when multiple target devices are running in parallel. To dynamically adjust parameters based on task volume, for example, light tasks (2 or fewer cloud cores) incur an additional consumption rate of 10%, medium tasks (2-5 cloud cores) 5%, and heavy tasks (5 or more cloud cores) 2%. Resource conflict penalty coefficient. The value can be determined based on the number of parallel processes and the degree of resource contention on the target device. For example, for an RTX 4090 graphics card, the system defaults to using... As a standard penalty benchmark, after determining the resource consumption or total resource consumption in the above manner, the system deducts the corresponding resource amount from the resource authorization certificate and updates the available resource balance in the resource authorization certificate.
[0048] Through the above methods, the system can accurately calculate the actual amount of resources consumed or the total amount of resources consumed by the target user in different scenarios, such as single device and multiple parallel devices, ensuring that the available resource balance in the resource authorization certificate is consistent with the actual consumption, and providing accurate resource usage basis for resource scheduling.
[0049] Optionally, based on any of the above embodiments or optional methods, the present disclosure embodiments may also employ the following method, but not limited thereto: the method includes: acquiring historical behavior data and real-time behavior data of the target user; performing weighted fusion processing on the historical behavior data and the real-time behavior data to generate a comprehensive behavior feature vector; inputting the comprehensive behavior feature vector into a pre-trained classification prediction network model to determine the behavior label and behavior correction coefficient of the target user; wherein, the behavior feature includes the behavior label and the behavior correction coefficient, the behavior label is used for the active time period characteristics of the target user, and the behavior correction coefficient is used to adjust the scheduling priority parameter.
[0050] In some embodiments, historical behavior data includes the target user's behavioral state information within a historical time window, while real-time behavior data includes the target user's behavioral state information at the current moment. The system continuously captures and identifies the target user's behavioral operations during service access through a service response engine and streaming computing module (e.g., a real-time computing framework based on Flink). The collected behavioral data covers game activity (e.g., KDA value, achievement completion progress), time characteristics (e.g., nighttime activity ratio, single session duration), and social authentication behavior (including real-name authentication status, activity participation, etc.). Under the premise of privacy compliance, the system performs anonymization and standardized encoding on the above-mentioned raw behavioral data to form a normalized behavioral feature vector. ,in, This represents the i-th behavioral dimension (such as nighttime activity rate, content completion rate, number of authentications, etc.). Historical and real-time behavioral data are weighted and fused to generate a comprehensive behavioral feature vector. Specifically: ,in, This represents the historical behavior vector of user u. This represents the user's real-time behavior vector at the current moment. and The weighting coefficients set for the model are used to balance the influence of historical and real-time behavior. (Comprehensive behavioral feature vector) The input is fed into a pre-trained classification prediction network model, which can be a multilayer perceptron (MLP) model, and uses a fully connected neural network structure to predict the current behavior label of the target user. and behavioral correction coefficient The behavior correction coefficient can be expressed as: ,in, For the output of the fully connected layer of the MLP, For the Sigmoid function, This is the bias term (default is -0.5). The learning weight vector is used (e.g., [0.3, 0.2, -0.1, ...]). Behavioral labels are used to indicate the activity level characteristics of the target user at different time periods, such as the target user being of the "Daytime Cloud High Activity Pool" type (high activity during the day) or the "Nighttime Cloud High Activity Pool" type (high activity at night); the behavioral correction coefficient is used to adjust the scheduling priority parameters to reflect the degree of impact of the target user's current behavioral state on resource demand.
[0051] In this way, the system can accurately obtain the dynamic behavioral characteristics of the target user and use these characteristics to determine the subsequent scheduling priority parameters, thereby improving the matching degree between the resource scheduling strategy and the user's actual needs.
[0052] In one feasible approach, the system can leverage a key-value storage service (such as Redis) to construct a dynamic behavior graph, meticulously recording the evolution of target user behavior event sequences and their status tags. Based on the dynamic changes in the target user's behavioral characteristics, the system considers the weight of behavior tags during the resource scheduling phase and flexibly adjusts the corresponding resource allocation strategy. For example, if the target user's behavior tag indicates a high-activity type during the day ("Daytime High-Activity Pool"), then its resource scheduling priority is increased during the daytime; if it indicates a high-activity type at night ("Nighttime High-Activity Pool"), then its excess resource usage needs are accommodated during the nighttime.
[0053] Optionally, based on any of the above embodiments or optional methods, the present disclosure embodiments may also employ the following method, but not limited thereto: the method includes: determining the scheduling priority parameter based on the resource metering parameters of the target device, the resource occupancy rate of each cloud service node, the heat coefficient determined by the user's resource request heat information, and the behavior correction coefficient in the behavior characteristics; wherein, the resource status information includes the resource occupancy rate and the heat coefficient.
[0054] In some embodiments, the scheduling priority parameter (i.e., Factor) is determined according to the following formula: ,in, The resource metering parameters (i.e., basic cloud core values) for the target device. This refers to resource utilization (Idle represents GPU / CPU idle rate). The heat coefficient, This is a behavior correction coefficient. Heat coefficient. This is determined based on the resource request intensity information of the target users, for example, by statistically analyzing access density and time period distribution through log streams (e.g., 20:00 to 23:00 is the peak period, and the intensity coefficient for this period can be increased by 10%-20%), reflecting the temporal density and concentration of resource requests. Resource utilization rate (i.e. This is obtained by real-time monitoring of the idle rate of computing resources in cloud service nodes. For example, if the utilization rate of a node is 80%, then the idle rate is 0.2, and the resource utilization rate is 0.8. Behavior correction coefficient. The output of the aforementioned classification prediction network model is limited to the range [0.3, 1.5]. The system uses... As a behavioral scoring guarantee mechanism, it avoids over-compression due to the sparsity of behavioral data.
[0055] By using the above methods, the system can comprehensively determine the scheduling priority parameters of the target user by integrating the target user's resource metering capabilities, the real-time load status of cloud service nodes, the target user's request intensity, and the target user's behavioral characteristics, thereby improving the scientific nature and accuracy of resource allocation decisions.
[0056] In one feasible approach, the system can identify resource demand trends in advance based on the behavioral tags of target users, predict the distribution curve of scheduling priority parameters, trigger resource pre-allocation strategies, and dynamically adjust scheduling weights. For example, in high-traffic scenarios (such as the release phase of popular games), the system can combine behavioral tags (such as "Night Cloud High Activity Pool" or "Day Cloud Light Interaction Pool") to predict peak resource demand of users within a specific time period, and pre-allocate computing resources before the peak arrives to cope with sudden traffic surges, ensuring resource scheduling quality and system elasticity.
[0057] Optionally, based on any of the above embodiments or optional methods, the present disclosure embodiments may also employ the following method, but not limited thereto: the method includes: determining the resource authorization status of the target user based on the available resources in the resource authorization credentials and the resource consumption required by the target user; and determining the resource scheduling strategy corresponding to the resource request of the target user based on the resource authorization status and the scheduling priority.
[0058] In some embodiments, the resource authorization status of the target user includes a first status and a second status. The system compares the available resources in the resource authorization credentials with the resource consumption required for the target user to execute the current task. If the available resources are greater than or equal to the required resource consumption, the resource authorization status is in the first status, indicating that the target user has the resource conditions to execute the current task; if the available resources are less than the required resource consumption, the resource authorization status is in the second status, indicating that the target user's available resources are insufficient to support the execution of the current task. Based on the resource authorization status and scheduling priority, the system comprehensively determines the resource scheduling policy corresponding to the target user's resource request. The resource scheduling policy is used to instruct operations such as resource allocation, rate limiting, or degraded allocation for the target user's resource request.
[0059] In this way, the system can simultaneously consider the resource authorization status and scheduling priority of the target user in the resource scheduling decision, ensuring that the resource scheduling strategy not only meets the resource permission conditions of the target user, but also reflects its priority order.
[0060] Optionally, based on any of the above embodiments or optional methods, the present disclosure embodiments may also employ the following method, but not limited thereto: the method includes: determining the resource consumption required for the task performed by the target user; when the resource authorization state is a first state and the scheduling priority is greater than a first preset threshold, allocating the resource request of the target user to a first cloud service node; when the resource authorization state is a second state and / or the scheduling priority is less than a second preset threshold, performing rate limiting on the resource request of the target user, or allocating it to a second cloud service node; wherein, the first state is a state in which the available resources in the resource authorization credentials are greater than or equal to the resource consumption required for the task performed by the target user, the second state is a state in which the available resources in the resource authorization credentials are less than the resource consumption required for the task performed by the target user, the performance of the first cloud service node is higher than that of the second cloud service node, and the first preset threshold is greater than the second preset threshold.
[0061] In some embodiments, when a target user performs a task, the system first determines the resource consumption required for that task. When the resource authorization status is in the first state (i.e., available resources are sufficient) and the scheduling priority is greater than a first preset threshold (e.g., 0.8), it indicates that the target user has the resource conditions and their resource request has a high priority. In this case, the system allocates the target user's resource request to a first cloud service node, which is a high-performance cloud service node. When the resource authorization status is in the second state (i.e., available resources are insufficient) and / or the scheduling priority is less than a second preset threshold (e.g., 0.3), it indicates that the target user has insufficient resources or their resource request has a low priority. In this case, the system performs rate limiting on the target user's resource request or allocates it to a second cloud service node, whose performance is lower than that of the first cloud service node.
[0062] In this way, the system can flexibly determine different scheduling methods such as resource allocation, rate limiting, or degradation based on the resource authorization status and scheduling priority of the target user, so as to realize differentiated resource scheduling for target users with different resource conditions and priority order, and improve the overall resource utilization efficiency of the system.
[0063] Optionally, based on any of the above embodiments or optional methods, the present disclosure embodiments may also employ the following methods, but not limited thereto: the method includes: executing the resource scheduling strategy and storing user behavior data, resource allocation data and scheduling result data during the execution process; updating the determination method of the scheduling priority parameter and / or the generation method of the resource scheduling strategy according to the stored user behavior data, resource allocation data and scheduling result data.
[0064] In some embodiments, after executing a resource scheduling strategy, the system constructs a storage service (e.g., a game memory bank) through a distributed file system and object storage module to store user behavior data, resource allocation data, and scheduling result data during the execution of the resource scheduling strategy. User behavior data includes information such as target user access events, behavioral state change trajectories, and system resource usage logs; resource allocation data includes information such as the values of scheduling priority parameters, node weight allocation, and resource authorization credential consumption; scheduling result data includes performance metrics such as task completion rate and resource scheduling hit rate. Based on the stored data, the system updates the method for determining scheduling priority parameters through an online learning mechanism. Specifically, the system uses a transfer learning mechanism to migrate a maturely trained user behavior network to scenarios with scarce data, such as new users or new devices, to address the cold start problem. Simultaneously, the system introduces a reinforcement learning strategy based on factor combination optimization, using task completion rate and resource scheduling hit rate as incentive metrics to dynamically adjust the weight configuration and allocation threshold of scheduling priority parameters, thereby optimizing the overall hit rate and task throughput efficiency of the scheduling strategy.
[0065] Through the above methods, the system can continuously use historical scheduling result data to iteratively optimize the scheduling strategy, forming a closed-loop feedback mechanism of "scheduling execution - behavior feedback - strategy adjustment", thereby gradually improving the accuracy of resource scheduling and the efficiency of resource utilization.
[0066] Optionally, based on any of the above embodiments or optional methods, the present disclosure embodiments may also employ the following method, but not limited thereto: the method includes: in response to the target user switching between different cloud service nodes, restoring the same resource configuration on the switched cloud service node as the cloud service node before the switch, according to the resource authorization credentials and the scheduling priority parameters.
[0067] In some embodiments, when a target user switches between different cloud service nodes, the system obtains the target user's user identifier, corresponding resource authorization credentials, and scheduling priority parameters. The target user's behavior trajectory and resource usage information on the cloud service node before the switch are synchronously uploaded to a storage service (e.g., a game memory bank) via edge nodes and pushed to the cloud service node the target user has switched to via a content delivery network (CDN). Based on the resource authorization credentials and scheduling priority parameters, the switched cloud service node restores the same resource configuration for the target user as before the switch, including computing resources, storage resources, and bandwidth resources, ensuring that the target user's resource access permissions and configurations remain consistent across different cloud service nodes.
[0068] By using the above methods, target users can maintain a consistent resource configuration status when switching between different cloud service nodes, avoiding resource permission interruptions or configuration inconsistencies caused by node switching, and improving the continuity and experience of users when using cloud services across nodes.
[0069] Corresponding to the resource scheduling method described above, this invention also proposes a resource scheduling device. Since the device embodiments of this invention correspond to the method embodiments described above, details not disclosed in the device embodiments can be referred to in the method embodiments described above, and will not be repeated here.
[0070] Figure 2 This is a schematic diagram of the structure of a resource scheduling device provided in an embodiment of the present disclosure, as shown below. Figure 2 As shown, it includes: an acquisition unit 21, a determination unit 22, and a generation unit 23.
[0071] The acquisition unit 21 is used to acquire a user identifier and acquire a resource authorization credential bound to the user identifier; wherein the resource authorization credential is used to indicate the available resources of the target user corresponding to the user identifier on each cloud service node; The determining unit 22 is used to determine the scheduling priority of the target user for each cloud service node based on the behavioral characteristics of the target user and the resource status information of each cloud service node. The generation unit 23 is used to generate a resource scheduling policy corresponding to the target user based on the resource authorization credentials and the scheduling priority.
[0072] In summary, the resource scheduling device provided in this disclosure includes: obtaining a user identifier and obtaining a resource authorization credential bound to the user identifier; wherein the resource authorization credential is used to indicate the available resources of the target user corresponding to the user identifier on each cloud service node; determining the scheduling priority of the target user for each cloud service node based on the behavioral characteristics of the target user and the resource status information of each cloud service node; generating a resource scheduling policy corresponding to the target user based on the resource authorization credential and the scheduling priority; enabling universal adaptation of resource permissions across multiple cloud service nodes, ensuring the effective continuation of resource permissions when the user accesses across nodes, and dynamically generating scheduling policies by combining user behavior and node resource status to adapt to real-time changes in resource supply and demand.
[0073] Furthermore, in one possible implementation of the embodiments of this disclosure, such as Figure 3 As shown, the device is also used for: Before the acquisition unit 21 acquires the user identifier, the resource metering parameters of the target device used by the target user are determined; wherein, the resource metering parameters are determined based on the performance parameters, usage duration and task type of the target device, the performance parameters are used to indicate the performance of the target device relative to the reference device, and the task type is used to indicate the degree of consumption of computing resources by the executed task; Generate the user identifier corresponding to the target user, and configure the resource authorization credential bound to the user identifier; wherein, the available resources in the resource authorization credential are determined based on the resource metering parameters.
[0074] Furthermore, in one possible implementation of the embodiments of this disclosure, such as Figure 3 As shown, the device is also used for: After the acquisition unit 21 acquires the user identifier and the resource authorization credential bound to the user identifier, when the target device is a single unit, the resource consumption corresponding to the task executed by the target user is determined according to the performance parameters, usage duration, task type, number of parallel tasks, maximum parallel capacity and load coefficient of the target device, and the resource consumption is removed from the resource authorization credential. If it is determined that there are at least two target devices and the target user is performing at least two tasks, the total resource consumption corresponding to the tasks performed by the target user is determined based on the performance parameters of the target devices, the usage duration, the task type, and the resource conflict correction amount, and the total resource consumption is removed from the resource authorization credentials; wherein, the resource conflict correction amount is used to reflect the additional resource consumption caused by resource contention and communication loss during the operation of at least two target devices.
[0075] Furthermore, in one possible implementation of the embodiments of this disclosure, such as Figure 3 As shown, the device is also used for: Before the determining unit 22 determines the scheduling priority of the target user for each cloud service node based on the target user's behavioral characteristics and the resource status information of each cloud service node, it acquires the target user's historical behavioral data and real-time behavioral data. The historical behavior data and the real-time behavior data are weighted and fused to generate a comprehensive behavior feature vector; The comprehensive behavioral feature vector is input into a pre-trained classification and prediction network model to determine the behavioral label and behavioral correction coefficient of the target user; wherein, the behavioral features include the behavioral label and the behavioral correction coefficient, the behavioral label is used for the active time period characteristics of the target user, and the behavioral correction coefficient is used to adjust the scheduling priority parameter.
[0076] Furthermore, in one possible implementation of the embodiments of this disclosure, such as Figure 3 As shown, the determining unit 22 is further configured to: The scheduling priority parameter is determined based on the resource metering parameters of the target device, the resource occupancy rate of each cloud service node, the heat coefficient determined by the user's resource request heat information, and the behavior correction coefficient in the behavior characteristics; wherein, the resource status information includes the resource occupancy rate and the heat coefficient.
[0077] Furthermore, in one possible implementation of the embodiments of this disclosure, such as Figure 3 As shown, the generation unit 23 is further configured to: The resource authorization status of the target user is determined based on the available resources in the resource authorization credentials and the resource consumption required by the target user; Based on the resource authorization status and the scheduling priority, determine the resource scheduling strategy corresponding to the resource request of the target user.
[0078] Furthermore, in one possible implementation of the embodiments of this disclosure, such as Figure 3 As shown, the generation unit 23 is further configured to: Determine the resource consumption required for the task performed by the target user; When the resource authorization status is in the first state and the scheduling priority is greater than the first preset threshold, the resource request of the target user is allocated to the first cloud service node; When the resource authorization status is in the second state and / or the scheduling priority is less than the second preset threshold, the resource request of the target user is rate-limited or allocated to the second cloud service node; wherein, the first state is when the available resources in the resource authorization credentials are greater than or equal to the resource consumption required by the task performed by the target user, the second state is when the available resources in the resource authorization credentials are less than the resource consumption required by the task performed by the target user, the performance of the first cloud service node is higher than that of the second cloud service node, and the first preset threshold is greater than the second preset threshold.
[0079] Furthermore, in one possible implementation of the embodiments of this disclosure, such as Figure 3 As shown, the device further includes: The execution unit 24 is configured to execute the resource scheduling policy after the generation unit generates the resource scheduling policy corresponding to the target user based on the resource authorization credentials and the scheduling priority, and to store user behavior data, resource allocation data and scheduling result data during the execution process. The update unit 25 is used to update the method for determining the scheduling priority parameter and / or the method for generating the resource scheduling strategy based on the stored user behavior data, resource allocation data, and scheduling result data.
[0080] Furthermore, in one possible implementation of the embodiments of this disclosure, such as Figure 3 As shown, the device is also used for: In response to the target user switching between different cloud service nodes, the same resource configuration as the cloud service node before the switch is restored on the switched cloud service node according to the resource authorization credentials and the scheduling priority parameters.
[0081] It should be noted that the foregoing explanation of the method embodiments also applies to the apparatus of the embodiments of this disclosure, and the principle is the same. Therefore, the embodiments of this disclosure are not limited thereto.
[0082] According to embodiments of this disclosure, this disclosure also provides an electronic device, a readable storage medium, and a computer program product.
[0083] Figure 4 A schematic block diagram of an example electronic device 1200 that can be used to implement embodiments of the present disclosure is shown. The electronic device is intended to represent various forms of digital computers, such as laptop computers, desktop computers, workstations, personal digital assistants, servers, blade servers, mainframe computers, and other suitable computers. The components shown herein, their connections and relationships, and their functions are merely illustrative and are not intended to limit the implementation of the present disclosure described and / or claimed herein.
[0084] like Figure 4 As shown, the electronic device 1200 includes a computing unit 1201, which can perform various appropriate actions and processes according to a computer program stored in a read-only memory 1202 or a computer program loaded from a storage unit 1208 into a random access memory 1203. The random access memory 1203 may also store various programs and data required for the operation of the electronic device 1200. The computing unit 1201, the read-only memory 1202, and the random access memory 1203 are interconnected via a bus 1204. An input / output interface 1205 is also connected to the bus 1204.
[0085] Multiple components in electronic device 1200 are connected to input / output interface 1205, including: input unit 1206, such as keyboard, mouse, etc.; output unit 1207, such as various types of monitors, speakers, etc.; storage unit 1208, such as disk, optical disk, etc.; and communication unit 1209, such as network card, modem, wireless transceiver, etc. Communication unit 1209 allows electronic device 1200 to exchange information / data with other devices through computer networks such as the Internet and / or various telecommunications networks.
[0086] The computing unit 1201 can be a variety of general-purpose and / or special-purpose processing components with processing and computing capabilities. Some examples of the computing unit 1201 include, but are not limited to, CPUs (Central Processing Units), GPUs (Graphics Processing Units), various special-purpose AI (Artificial Intelligence) computing chips, various computing units running machine learning model algorithms, DSPs (Digital Signal Processors), and any suitable processor, controller, microcontroller, etc. The computing unit 1201 performs the various methods and processes described above, such as resource scheduling methods. For example, in some embodiments, the resource scheduling method may be implemented as a computer software program tangibly contained in a machine-readable medium, such as storage unit 1208. In some embodiments, part or all of the computer program may be loaded and / or installed on the electronic device 1200 via read-only memory 1202 and / or communication unit 1209. When the computer program is loaded into random access memory 1203 and executed by the computing unit 1201, one or more steps of the methods described above may be performed. Alternatively, in other embodiments, the computing unit 1201 may be configured to perform the aforementioned resource scheduling method by any other suitable means (e.g., by means of firmware).
[0087] Various implementations of the systems and techniques described above herein can be implemented in digital electronic circuit systems, integrated circuit systems, FPGAs (Field Programmable Gate Arrays), ASICs (Application-Specific Integrated Circuits), ASSPs (Application-Specific Standard Products), SOCs (System-on-Chips), CPLDs (Complex Programmable Logic Devices), computer hardware, firmware, software, and / or combinations thereof. These various implementations may include implementations in one or more computer programs that can be executed and / or interpreted on a programmable system including at least one programmable processor, which may be a dedicated or general-purpose programmable processor, capable of receiving data and instructions from a storage system, at least one input device, and at least one output device, and transmitting data and instructions to the storage system, the at least one input device, and the at least one output device.
[0088] The program code used to implement the methods of this disclosure may be written in any combination of one or more programming languages. This program code may be provided to a processor or controller of a general-purpose computer, special-purpose computer, or other programmable data processing apparatus, such that when executed by the processor or controller, the program code causes the functions / operations specified in the flowcharts and / or block diagrams to be implemented. The program code may be executed entirely on a machine, partially on a machine, as a standalone software package partially on a machine and partially on a remote machine, or entirely on a remote machine or server.
[0089] In the context of this disclosure, a machine-readable medium can be a tangible medium that may contain or store a program for use by or in conjunction with an instruction execution system, apparatus, or device. A machine-readable medium can be a machine-readable signal medium or a machine-readable storage medium. A machine-readable medium can be, but is not limited to, electronic, magnetic, optical, electromagnetic, infrared, or semiconductor systems, apparatus, or devices, or any suitable combination of the foregoing. More specific examples of machine-readable storage media include electrical connections based on one or more wires, portable computer disks, hard disks, RAM, ROM, EPROM (Electrically Programmable Read-Only Memory) or flash memory, optical fiber, CD-ROM (Compact Disc Read-Only Memory), optical storage devices, magnetic storage devices, or any suitable combination of the foregoing.
[0090] To provide interaction with a user, the systems and techniques described herein can be implemented on a computer having: a display device for displaying information to the user (e.g., a CRT (Cathode-Ray Tube) or LCD (Liquid Crystal Display) monitor); and a keyboard and pointing device (e.g., a mouse or trackball) through which the user provides input to the computer. Other types of devices can also be used to provide interaction with the user; for example, feedback provided to the user can be any form of sensory feedback (e.g., visual feedback, auditory feedback, or tactile feedback); and input from the user can be received in any form (including sound input, voice input, or tactile input).
[0091] The systems and technologies described herein can be implemented in computing systems that include backend components (e.g., as data servers), or middleware components (e.g., application servers), or frontend components (e.g., user computers with graphical user interfaces or web browsers through which users can interact with implementations of the systems and technologies described herein), or any combination of such backend, middleware, or frontend components. The components of the system can be interconnected via digital data communication of any form or medium (e.g., communication networks). Examples of communication networks include LANs (Local Area Networks), WANs (Wide Area Networks), the Internet, and blockchain networks.
[0092] Computer systems can include clients and servers. Clients and servers are generally geographically separated and typically interact via communication networks. The client-server relationship is created by computer programs running on the respective computers and having a client-server relationship with each other. A server can be a cloud server, also known as a cloud computing server or cloud host, a hosting product within the cloud computing service system that addresses the shortcomings of traditional physical hosts and VPS (Virtual Private Server) services, such as high management difficulty and weak business scalability. Servers can also be servers for distributed systems or servers incorporating blockchain technology.
[0093] It's important to note that artificial intelligence (AI) is the study of enabling computers to simulate certain human thought processes and intelligent behaviors (such as learning, reasoning, thinking, and planning). It encompasses both hardware and software technologies. AI hardware technologies generally include sensors, dedicated AI chips, cloud computing, distributed storage, and big data processing. AI software technologies primarily include computer vision, speech recognition, natural language processing, machine learning / deep learning, big data processing, and knowledge graph technologies.
[0094] It should be understood that the various forms of processes shown above can be used to rearrange, add, or delete steps. For example, the steps described in this disclosure can be executed in parallel, sequentially, or in different orders, as long as the desired result of the technical solution disclosed in this application can be achieved, and this is not limited herein.
[0095] The specific embodiments described above do not constitute a limitation on the scope of protection of this disclosure. Those skilled in the art should understand that various modifications, combinations, sub-combinations, and substitutions can be made according to design requirements and other factors. Any modifications, equivalent substitutions, and improvements made within the spirit and principles of this disclosure should be included within the scope of protection of this disclosure.
Claims
1. A resource scheduling method, characterized in that, include: Obtain a user identifier and a resource authorization credential bound to the user identifier; wherein the resource authorization credential is used to indicate the available resources of the target user corresponding to the user identifier on each cloud service node; Based on the behavioral characteristics of the target user and the resource status information of each cloud service node, the scheduling priority of the target user for each cloud service node is determined. Based on the resource authorization credentials and the scheduling priority, a resource scheduling policy corresponding to the target user is generated.
2. The method according to claim 1, characterized in that, Prior to obtaining the user identifier, the method further includes: Determine the resource metering parameters of the target device used by the target user; wherein the resource metering parameters are determined based on the performance parameters, usage duration and task type of the target device, the performance parameters are used to indicate the performance of the target device relative to the reference device, and the task type is used to indicate the degree of consumption of computing resources by the executed task; Generate the user identifier corresponding to the target user, and configure the resource authorization credential bound to the user identifier; wherein, the available resources in the resource authorization credential are determined based on the resource metering parameters.
3. The method according to claim 2, characterized in that, After obtaining the user identifier and the resource authorization credentials bound to the user identifier, the method further includes: When the target device is a single unit, the resource consumption corresponding to the task executed by the target user is determined based on the performance parameters, usage duration, task type, number of parallel tasks, maximum parallel capacity and load factor of the target device, and the resource consumption is removed from the resource authorization credentials. If it is determined that there are at least two target devices and the target user is performing at least two tasks, the total resource consumption corresponding to the tasks performed by the target user is determined based on the performance parameters of the target devices, the usage duration, the task type, and the resource conflict correction amount, and the total resource consumption is removed from the resource authorization credentials; wherein, the resource conflict correction amount is used to reflect the additional resource consumption caused by resource contention and communication loss during the operation of at least two target devices.
4. The method according to claim 1, characterized in that, Before determining the scheduling priority of the target user for each cloud service node based on the target user's behavioral characteristics and the resource status information of each cloud service node, the method further includes: Obtain the target user's historical and real-time behavior data; The historical behavior data and the real-time behavior data are weighted and fused to generate a comprehensive behavior feature vector; The comprehensive behavioral feature vector is input into a pre-trained classification and prediction network model to determine the behavioral label and behavioral correction coefficient of the target user; wherein, the behavioral features include the behavioral label and the behavioral correction coefficient, the behavioral label is used for the active time period characteristics of the target user, and the behavioral correction coefficient is used to adjust the scheduling priority parameter.
5. The method according to claim 4, characterized in that, The step of determining the scheduling priority of the target user for each cloud service node based on the target user's behavioral characteristics and the resource status information of each cloud service node includes: The scheduling priority parameter is determined based on the resource metering parameters of the target device, the resource occupancy rate of each cloud service node, the heat coefficient determined by the user's resource request heat information, and the behavior correction coefficient in the behavior characteristics; wherein, the resource status information includes the resource occupancy rate and the heat coefficient.
6. The method according to claim 1, characterized in that, The step of generating a resource scheduling policy corresponding to the target user based on the resource authorization credential and the scheduling priority includes: The resource authorization status of the target user is determined based on the available resources in the resource authorization credentials and the resource consumption required by the target user; Based on the resource authorization status and the scheduling priority, determine the resource scheduling strategy corresponding to the resource request of the target user.
7. The method according to claim 6, characterized in that, The step of determining the resource scheduling strategy corresponding to the target user's resource request based on the resource authorization status and the scheduling priority includes: Determine the resource consumption required for the task performed by the target user; When the resource authorization status is in the first state and the scheduling priority is greater than the first preset threshold, the resource request of the target user is allocated to the first cloud service node; When the resource authorization status is in the second state and / or the scheduling priority is less than the second preset threshold, the resource request of the target user is rate-limited or allocated to the second cloud service node; wherein, the first state is when the available resources in the resource authorization credentials are greater than or equal to the resource consumption required by the task performed by the target user, the second state is when the available resources in the resource authorization credentials are less than the resource consumption required by the task performed by the target user, the performance of the first cloud service node is higher than that of the second cloud service node, and the first preset threshold is greater than the second preset threshold.
8. The method according to claim 1, characterized in that, After generating the resource scheduling policy corresponding to the target user based on the resource authorization credentials and the scheduling priority, the method further includes: Execute the resource scheduling strategy and store user behavior data, resource allocation data, and scheduling result data during the execution process; Based on the stored user behavior data, resource allocation data, and scheduling result data, update the method for determining the scheduling priority parameter and / or the method for generating the resource scheduling strategy.
9. The method according to claim 1, characterized in that, The method further includes: In response to the target user switching between different cloud service nodes, the same resource configuration as the cloud service node before the switch is restored on the switched cloud service node according to the resource authorization credentials and the scheduling priority parameters.
10. A resource scheduling device, characterized in that, include: An acquisition unit is used to acquire a user identifier and a resource authorization credential bound to the user identifier; wherein the resource authorization credential is used to indicate the available resources of the target user corresponding to the user identifier on each cloud service node; The determining unit is used to determine the scheduling priority of the target user for each cloud service node based on the behavioral characteristics of the target user and the resource status information of each cloud service node. The generation unit is used to generate a resource scheduling policy corresponding to the target user based on the resource authorization credentials and the scheduling priority.
