Power business load balancing scheduling method and system based on hybrid cloud architecture

Abstract

The invention provides a hybrid cloud architecture-based power business load balancing scheduling method and system, and the method comprises the steps: classifying power business operation data collected in real time based on the importance of power businesses, and obtaining a target power business category; determining a current load state of the target power business category by combining resource state information of the private cloud resource pool and the public cloud resource pool based on the target power business category and the corresponding power business operation data; and determining a load scheduling strategy based on the current load state for the target power business category. According to the method, the problems that in an existing method, the resource capacity of a local physical server cluster is fixed, sudden peak loads are difficult to deal with, and service response delay and data processing interruption are likely to be caused are effectively solved, and meanwhile stable and efficient operation of a power system is guaranteed.

Description

Technical Field

[0001] This invention relates to the field of computer technology, and in particular to a power service load balancing scheduling method and system based on a hybrid cloud architecture. Background Technology

[0002] In the operation of power systems, load balancing and dispatching of power services is a crucial link in ensuring the stable and efficient operation of the power system. Existing load balancing and dispatching methods mainly rely on local physical server clusters for resource allocation and dispatching. On the one hand, since the resource capacity of local physical server clusters is fixed, when power services experience sudden peak loads (such as a surge in residential electricity consumption during extreme weather or temporary high-load production by large industrial enterprises), local resources cannot be rapidly expanded, easily leading to overload, causing problems such as service response delays and data processing interruptions, affecting the normal operation of power services. On the other hand, existing methods often adopt a "one-size-fits-all" resource allocation strategy to adapt to the differentiated resource capabilities of different types of power services (such as residential electricity services, industrial production electricity services, and power dispatching and management services). This easily leads to low resource utilization and also makes it difficult to take into account the security and elasticity requirements of the services. Summary of the Invention

[0003] This invention provides a power service load balancing scheduling method and system based on a hybrid cloud architecture to ensure the stable and efficient operation of the power system.

[0004] In a first aspect, the present invention provides a power service load balancing scheduling method based on a hybrid cloud architecture, comprising: Based on the importance of power services, the real-time collected power service operation data are classified to obtain the target power service category; the importance of power services is determined by the degree of impact on power system operation, data security requirements, and load fluctuation characteristics. Based on the target power service category and the corresponding power service operation data, and combined with the resource status information of the private cloud resource pool and the public cloud resource pool, the current load status of the target power service category is determined. For the target power service category, a load scheduling strategy is determined based on the current load status.

[0005] Secondly, the present invention also provides a power service load balancing scheduling system based on a hybrid cloud architecture, applied to the power service load balancing scheduling method based on a hybrid cloud architecture as described in the first aspect; the power service load balancing scheduling system based on a hybrid cloud architecture includes: The business classification module is used to classify the real-time collected power business operation data based on the importance of the power business to obtain the target power business category; the importance of the power business is determined by the degree of impact on the operation of the power system, the level of data security requirements, and the characteristics of load fluctuation. The load status determination module is used to determine the current load status of the target power service category based on the target power service category and the corresponding power service operation data, combined with the resource status information of the private cloud resource pool and the public cloud resource pool. The load scheduling module is used to determine a load scheduling strategy based on the current load status for the target power service category.

[0006] Thirdly, the present invention also provides an electronic device, comprising: a memory for storing computer software programs; and a processor for reading and executing the computer software programs, thereby realizing the power service load balancing scheduling method based on a hybrid cloud architecture as described above.

[0007] Fourthly, the present invention also provides a non-transitory computer-readable storage medium storing a computer software program, which, when executed by a processor, implements the power service load balancing scheduling method based on a hybrid cloud architecture as described above.

[0008] Fifthly, the present invention also provides a computer program product, including a computer program that, when executed by a processor, implements the power service load balancing scheduling method based on a hybrid cloud architecture as described above.

[0009] The power service load balancing scheduling method based on a hybrid cloud architecture provided in this invention classifies real-time power service operation data according to the importance of power services to the power system, clarifies the current power service category, and accurately determines the current load status of various power services by combining the resource status information of private cloud resource pools and public cloud resource pools. This breaks the limitation of relying solely on local physical server clusters to obtain resource information. Finally, based on the current load status and target service type, the corresponding load scheduling strategy is determined, achieving targeted and flexible resource allocation. In summary, this invention effectively solves the problems of fixed local physical server cluster resource capacity in existing methods, which makes it difficult to cope with sudden peak loads and easily leads to service response delays and data processing interruptions, ensuring the stable operation of power services under various load conditions. At the same time, it improves the overall resource utilization rate of the cloud resource pool, fully meets the differentiated security and elasticity requirements of different types of power services, and ensures the stable and efficient operation of the power system. Attached Figure Description

[0010] Figure 1This is a flowchart illustrating the power service load balancing scheduling method based on a hybrid cloud architecture provided in an embodiment of the present invention. Figure 2 This is a schematic diagram of the structure of a power service load balancing scheduling system based on a hybrid cloud architecture provided in an embodiment of the present invention; Figure 3 An embodiment diagram of the electronic device provided in this invention; Figure 4 An embodiment diagram of a computer-readable storage medium provided in accordance with the present invention. Detailed Implementation

[0011] The technical solutions of the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of the present invention, and not all embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of the present invention.

[0012] In the description of this invention, the terms "first" and "second" are used for descriptive purposes only and should not be construed as indicating or implying relative importance or implicitly specifying the number of indicated technical features. Thus, features defined as "first" or "second" may explicitly or implicitly include one or more features. In the description of this invention, "a plurality of" means two or more, unless otherwise explicitly specified.

[0013] In the description of this invention, the term "for example" is used to mean "used as an example, illustration, or description." Any embodiment described as "for example" in this invention is not necessarily to be construed as being more preferred or advantageous than other embodiments. The following description is provided to enable any person skilled in the art to make and use the invention. Details are set forth in the following description for purposes of explanation. It should be understood that those skilled in the art will recognize that the invention can be made without using these specific details. In other instances, well-known structures and processes will not be described in detail to avoid obscuring the description of the invention with unnecessary detail. Therefore, the invention is not intended to be limited to the embodiments shown, but is consistent with the broadest scope of the principles and features disclosed herein.

[0014] See Figure 1 , Figure 1 This is a flowchart illustrating the power service load balancing scheduling method based on a hybrid cloud architecture provided by the present invention. The hybrid cloud architecture includes a private cloud resource pool and a public cloud resource pool. In this embodiment of the invention, the execution entity of the power service load balancing scheduling method based on a hybrid cloud architecture is the load balancing scheduling system. Therefore, the power service load balancing scheduling method based on a hybrid cloud architecture includes: Step 10: Classify the real-time collected power service operation data based on the importance of the power service to obtain the target power service category; the importance of the power service is determined by comprehensively considering the degree of impact on the operation of the power system, the level of data security requirements, and the characteristics of load fluctuation.

[0015] Optionally, the load balancing dispatch system acquires power service operation data through real-time data acquisition modules deployed at various nodes of the power system. This data includes information such as data transmission volume, data interaction frequency, scope of service interruption impact, data encryption requirement level, and historical load fluctuation amplitude. Subsequently, a power service importance assessment model is constructed using "power system operation impact level," "data security requirement level," and "load fluctuation characteristics" as core evaluation indicators. The collected power service operation data is then quantitatively analyzed, ultimately classifying power services into three target categories: core security services, regular elastic services, and non-core extended services. Based on these categories, pre-defined resource pool carrying rules are established: core security services are carried solely by private cloud resource pools; regular elastic services can be carried by either private or public cloud resource pools; and non-core extended services are carried solely by public cloud resource pools.

[0016] Furthermore, let the business importance assessment value be... The quantitative value of the impact on power system operation is (Value range 0-10), the data security requirement level is quantified as follows: (Level 1-Level 5 correspond to quantization values ​​1-5), the quantization value of load fluctuation characteristics is (Values ​​range from 0 to 1, with larger values ​​for smaller fluctuations), the formula for the power business importance assessment model is: ; in, , , The weights of the three evaluation indicators can be determined using the Analytic Hierarchy Process (AHP) to satisfy the following conditions: Furthermore, based on the characteristics of the power business, typically ,like , , And it performs analysis based on pre-set classification thresholds, when... At that time, it was determined to be a core security-related business; when When, it is determined to be a regular elastic business; when At that time, it was determined to be a non-core extended business.

[0017] In one embodiment, when the load balancing dispatch system collects power service operation data for a certain region, it finds that the operation data of the "Power Dispatch Instruction Transmission Service" shows that: service interruption will cause regional power grid dispatch failure, with an impact assessment score of 9.2 (out of 10) on the power system operation; the service data contains sensitive information such as power grid topology and generator start / stop instructions, and the data security requirement level is the highest (Level 5); the historical load fluctuation range is only ±3%, and the load fluctuation characteristics are stable. After analysis through the evaluation model, this service is classified as a core security service and allocated to a private cloud resource pool for carrying.

[0018] Meanwhile, operational data for the "Residential Daily Electricity Consumption Data Statistics Service" was collected: the service interruption only affected the efficiency of electricity bill generation, and the assessment score for the degree of impact on the operation of the power system was 4.5; the data contains general privacy information such as residents' electricity addresses and electricity consumption, and the data security requirement level is medium (Level 3); the historical load fluctuation range is ±15%, and the load will increase significantly during peak electricity consumption periods (such as summer evenings). The system classifies it as a regular elastic service, which can be flexibly allocated between private cloud and public cloud resource pools according to resource status.

[0019] In addition, the operational data collected from the "Power Science Popularization Knowledge Push Service" shows that: the service interruption has no substantial impact, and the assessment score for the degree of impact on the operation of the power system is 1.0; the data is publicly available science popularization content, and the data security requirement level is low (Level 1); the historical load fluctuation range is ±40%, and the load surges during holidays. The system classifies it as a non-core extended service and designates it to be carried by the public cloud resource pool.

[0020] Step 20: Based on the target power service category and the corresponding power service operation data, and combined with the resource status information of the private cloud resource pool and the public cloud resource pool, determine the current load status of the target power service category.

[0021] Optionally, after classifying the power services, the load balancing scheduling system retrieves the historical load baseline data for each target power service category (this data is generated in advance by the system based on the operation data of the same category of services over the past 12 months, including the normal load range threshold under different time periods and scenarios). It combines this data with the real-time power service operation data corresponding to the target power service category obtained in step 10 (such as real-time data processing volume, server CPU utilization, and memory usage). At the same time, it obtains the real-time resource status information of the private cloud resource pool and the public cloud resource pool through the cloud resource monitoring interface (including the number of remaining CPU cores, available memory capacity, and remaining storage space of the private cloud server cluster, and the number of callable computing nodes and bandwidth resource reserves of the public cloud resource pool). A load status judgment model is constructed, using "the deviation rate between the real-time load value and the historical load baseline threshold" and "the matching degree between the remaining resources in the resource pool and the real-time business demand" as input parameters to calculate the load status quantification value. When the quantification value exceeds the preset overload threshold, the current load status is determined to be overloaded; when the quantification value is lower than the preset low load threshold, the current load status is determined to be low load; and when it is between the two, it is a normal load status (since there is no need to perform load scheduling in the normal load status, this step only focuses on the overload and low load statuses and performs load scheduling based on these two statuses).

[0022] Furthermore, in the process of constructing the load status judgment model, the load status quantification value is set as follows: Real-time load value The historical load baseline upper limit is The historical load baseline lower limit is The remaining resources in the resource pool can support a load value of The model formula for the load state judgment model is as follows: ; in, , These are the load deviation impact coefficient and the resource matching impact coefficient, respectively, satisfying... ,generally , ;when hour, Set to 0 (service paused); when hour, Set to 0 (resources fully meet demand). Additionally, a preset overload threshold is set. Low load threshold ,when When, it is determined to be an overload state; when When this occurs, it is determined to be a low-load state.

[0023] Continuing with the above embodiment, for the "Residential Daily Electricity Consumption Data Statistics Service" (a regular elastic service), the system retrieves its historical load baseline data: the normal load range is processing 1000-1500 electricity consumption data records per hour, the overload threshold is set to processing more than 2000 data records per hour, and the low load threshold is set to processing less than 500 data records per hour. Real-time data collection shows that the service is currently processing 1800 electricity consumption data records per hour. Simultaneously, the system obtains that the remaining CPU cores of the corresponding service node in the private cloud resource pool are 2 (only capable of supporting 1600 data records per hour), while the available computing nodes in the public cloud resource pool can support 2500 data records per hour. The system calculates the load status quantification value: real-time load deviation rate = (1800-1500) / 1500 = 20%, resource matching degree = (2 + number of CPU cores corresponding to the number of callable computing nodes) / number of CPU cores required for 1800 data entries = (2+3) / 4 = 125% (assuming that 1 CPU core is required to process 400 data entries per hour). The comprehensive quantification value exceeds the overload threshold, and the current load status of this regular elastic business is determined to be overloaded.

[0024] For the "Power Dispatch Command Transmission Service" (a core security service), the historical load baseline is 800-1000 commands per hour, the overload threshold is 1200 commands / hour, and the low load threshold is 500 commands / hour. Real-time data shows a current hourly transmission of 600 commands. The private cloud resource pool has 5 remaining CPU cores on the corresponding node (capable of supporting 1500 command transmissions per hour). The load status quantification is calculated as follows: Real-time load deviation rate = (600-800) / 800 = -25%, resource matching degree = 5 / (600 / 400) = 333%. The overall quantification value is lower than the low load threshold, therefore the current load status is determined to be low load.

[0025] Step 30: For the target power service category, determine the load scheduling strategy based on the current load status.

[0026] Optionally, the load balancing scheduling system determines the load scheduling strategy under the current load state based on the target power service category and the corresponding current load state determined in step 20, and performs load scheduling according to the load scheduling strategy to meet the service load demand, as described in steps 301-305.

[0027] This invention classifies real-time power service operation data based on the importance of power services to the power system, clearly defining the current power service category. By combining resource status information from private and public cloud resource pools, it accurately determines the current load status of various power services, breaking the limitation of relying solely on local physical server clusters for resource information. Finally, based on the current load status and target service type, it determines the corresponding load scheduling strategy, achieving targeted and flexible resource allocation. In summary, this invention effectively solves the problems of fixed local physical server cluster resource capacity in existing methods, which makes it difficult to cope with sudden peak loads and easily leads to service response delays and data processing interruptions, ensuring the stable operation of power services under various load conditions. Simultaneously, it improves the overall resource utilization rate of the cloud resource pool, fully meeting the differentiated security and elasticity requirements of different types of power services, and ensuring the stable and efficient operation of the power system.

[0028] In one embodiment, steps 301-305 are described as follows: Step 301: Based on the operational data of the target power business category and the resource status information of the private cloud resource pool and the public cloud resource pool, determine the resource demand matching coefficient corresponding to each business type.

[0029] Optionally, the load balancing scheduling system retrieves the determined target power service category and its operational data (data transmission volume, computing task volume), as well as the resource status information of the private cloud resource pool and public cloud resource pool (including remaining computing resources, remaining storage resources, and network bandwidth). Based on the service category, it calls the corresponding resource demand matching coefficient calculation formula to determine the resource demand matching coefficient for each service type.

[0030] Specifically, if the target power business category is a core security business, then only the matching degree with the private cloud resource pool is determined. The formula for calculating the matching coefficient of its resource requirements with the private cloud resource pool is as follows: ; in, This represents the matching coefficient between the resource requirements of core security-related services and the private cloud resource pool; This indicates the amount of data transmitted for core security-related services; This indicates the computational workload of core security-related tasks; This indicates the amount of remaining computing resources in the private cloud resource pool; This indicates the amount of remaining storage resources in the private cloud resource pool; This indicates the network bandwidth of the private cloud resource pool; If the target power service category is a regular elastic service, then the matching degree with the private cloud resource pool and the public cloud resource pool should be determined separately. The formula for calculating the resource demand matching coefficient with the private cloud resource pool is as follows: ; The formula for calculating the resource demand matching coefficient with the public cloud resource pool is as follows: ; in, This represents the matching coefficient between the resource requirements of regular elastic services and the private cloud resource pool; This represents the matching coefficient between the resource requirements of regular elastic services and the public cloud resource pool; This indicates the amount of data transmitted in a typical elastic service. This indicates the amount of computational tasks required for typical elastic business operations; This indicates the amount of remaining computing resources in the public cloud resource pool; This indicates the amount of remaining storage resources in the public cloud resource pool; This indicates the network bandwidth of the public cloud resource pool. If the target power business category is a non-core extended business, then only the matching degree with the public cloud resource pool is determined. The formula for calculating the matching coefficient of its resource demand with the public cloud resource pool is as follows:

[0031] in, This represents the matching coefficient between the resource requirements of non-core extended business functions and the public cloud resource pool: This indicates the amount of data transmitted for non-core extended business functions; This indicates the amount of computational tasks required for non-core extended business functions.

[0032] In one embodiment, the "Power Dispatch Command Transmission Service" (a core security service) is taken as an example. Its operational data includes: GB (data transfer volume) 10,000 records (computational workload). Private cloud resource status: Core (remaining computing resources) TB (Remaining storage resources) Gbps (network bandwidth). Substituting this into the formula yields the resource matching coefficient with the private cloud resource pool. .

[0033] Taking the "Residential Daily Electricity Consumption Data Statistics Service" (a regular flexible service) as an example, its operational data includes: GB (data transfer volume) 10,000 records (computational workload). Private cloud resource status: Core (remaining computing resources) TB (Remaining storage resources) Gbps (network bandwidth). Public cloud resource status: nuclear, TB, Gbps. Substituting into the formula, the resource requirement matching coefficient with the private cloud resource pool is calculated. Calculate the resource demand matching coefficient with the public cloud resource pool. .

[0034] Taking the "Power Science Popularization Knowledge Push Service" (a non-core extended service) as an example, its operational data includes: GB (data transfer volume) 10,000 records (computational workload). Public cloud resource status: nuclear, TB, Gbps. Substituting into the formula, the resource requirement matching coefficient with the public cloud resource pool is calculated. .

[0035] Step 302: Based on the current load status of the target power service category and the resource demand matching coefficient corresponding to each service category, determine the load scheduling priority coefficient of each service category; the current load status includes overload status and low load status.

[0036] Optionally, the load balancing dispatch system determines the load dispatch priority coefficient for each service category based on the determined current load status (including overload or low load) and the resource demand matching coefficients corresponding to each service category obtained in step 301. Specifically, when the target power service category is in an overload state, the formula for calculating the load dispatch priority coefficient is: ; in This indicates the load scheduling priority coefficient for a service category under overload conditions. Indicates the importance coefficient of the business category (core security business). Regular elastic business Non-core extended business ); This indicates the resource demand matching coefficient corresponding to this business category; This indicates the current load (under overload conditions) for this service category. The value is the portion of the actual load exceeding the rated load. When a service category is under low load, its load scheduling priority coefficient is calculated using the following formula: ; in, This indicates the load scheduling priority coefficient for a service category under low load conditions. Indicates the importance coefficient of the business category; This indicates the resource demand matching coefficient corresponding to this business category; This indicates the maximum load capacity for this business category; This indicates the current load of this business category.

[0037] Continuing with the above embodiments, let's take the residential daily electricity consumption data statistics service (i.e., regular flexible service) as an example. Taking the overload state of ) as an example. Among them, the resource demand matching coefficient (Resource demand matching coefficient with the private cloud resource pool), the portion of actual load exceeding the rated load. 10,000 lines / hour (rated load 15 million lines / hour, actual 18 million lines / hour). Substituting this into the load scheduling priority coefficient formula under overload conditions, we get... If the priority calculation for this service's subsequent matching with the public cloud (assuming the excess load remains unchanged) yields... .

[0038] For power dispatch instruction transmission services (i.e., core security services), Taking a low-load state as an example. Resource demand matching coefficient. Maximum service load Items / hour, current load Items / hour. Substituting this into the load scheduling priority coefficient formula under low load conditions yields... .

[0039] Step 303: Based on the load scheduling priority coefficient of each business category and the load carrying threshold of the private cloud resource pool and the public cloud resource pool, determine the business load that each resource pool can handle; the load carrying threshold includes the maximum computing load carrying threshold, the maximum storage load carrying threshold and the maximum network load carrying threshold.

[0040] Optionally, the load balancing scheduling system determines the service load that each resource pool can handle by using the minimum capacity threshold (i.e., the most restrictive threshold among computing, storage, and network) of each resource pool as a benchmark, combined with the load scheduling priority coefficient in step 302, through the coupling of resource capacity and priority.

[0041] Continuing with the above embodiments, let's take a private cloud resource pool undertaking regular elastic services as an example. Regarding the private cloud load capacity threshold: maximum computational load capacity threshold... Core, maximum storage load capacity threshold TB, maximum network load capacity threshold Gbps. Load scheduling priority coefficient for regular elastic services (such as residential electricity consumption data statistics). (Matching with private clouds). Based on this, the system calculates the service capacity corresponding to each threshold: Computing resources can support: 10,000 tasks per hour (assuming 0.002 cores are required for every 10,000 tasks).

[0042] Storage resources can support: 10,000 tasks per hour (assuming 0.002TB is required for each 10,000 tasks).

[0043] Network resources can support: 10,000 tasks per hour (assuming 0.0002Gbps is required for each 10,000 tasks).

[0044] Finally, taking the minimum value, we find that the private cloud can handle a service load of 137.3 million messages per hour.

[0045] Step 304: Based on the service load capacity that each resource pool can handle and the load migration cost coefficient for migrating each service category between different resource pools, determine the load migration allocation scheme for each service category.

[0046] Optionally, the load balancing scheduling system retrieves the load migration cost coefficients for each business category when migrating between different resource pools, and combines this with the business load capacity that each resource pool can handle in step 303 to finally determine the load migration allocation scheme for each business category, as described in steps 3041-3045.

[0047] Step 305: Determine the load scheduling strategy based on the load migration and allocation schemes for each business category and the real-time resource change rates of the private cloud resource pool and the public cloud resource pool.

[0048] Optionally, the load balancing scheduling system obtains the real-time resource change rate (i.e., the rate of resource consumption or release, such as the rate of CPU utilization change and storage capacity change) of the private cloud resource pool and public cloud resource pool. Combined with the load migration and allocation scheme in step 304, the load scheduling strategy is finally determined through dynamic adaptation of the resource change rate and migration amount. This load scheduling strategy includes load migration operation instructions between resource pools, resource expansion / contraction instructions, etc.

[0049] Continuing with the above embodiments, let's take the migration scheme for residential daily electricity consumption data statistics as an example. Among them, the real-time change rate of public cloud resources... Cores / second (slow resource consumption rate). Load to be migrated. 10,000 messages per hour, which translates to approximately 8,333,300 messages per second. Based on this calculation, the dynamic adaptation value is: The final scheduling strategy was determined to be: to migrate the business load from the private cloud to the public cloud at a rate of 8.3333 million messages per second, while monitoring the change rate of public cloud resources. If the change rate exceeds 0.05 cores / second, the migration rate will be adjusted to 6 million messages per second.

[0050] This invention achieves precise quantitative decision-making throughout the entire process from business classification to resource scheduling by multi-dimensional quantitative calculation of resource demand matching coefficient and load scheduling priority coefficient, as well as comprehensive consideration of resource pool capacity, migration cost, and real-time resource change rate. Furthermore, by leveraging the load migration allocation scheme and dynamic adaptation of real-time resource change rate, load balancing between resource pools is achieved, which avoids overloading of a single resource pool and makes full use of the remaining resources of each resource pool.

[0051] In one embodiment, steps 3041-3045 are described as follows: Step 3041: Based on the service load capacity that each resource pool can handle and the demand of each service category for different types of load, construct an initial matching matrix for the service resource pool.

[0052] Optionally, the load balancing scheduling system may adjust the load based on the service load capacity of each resource pool determined in step 303 (including the load capacity of the private cloud resource pool). The public cloud resource pool can handle a certain amount of load. ), combined with the different types of load requirements for each business category (including computing load requirements) Storage load requirements Network load requirements Based on this, an initial matching matrix for the three-dimensional business resource pool is constructed. The initial matching matrix is ​​defined as follows: ,in, Business categories (1-Core security category, 2-Regular elastic category, 3-Non-core extension category). Resource pool type (1-Private cloud, 2-Public cloud). For load type requirements (1-Compute load, 2-Storage load, 3-Network load), the matrix element values ​​are business categories. With resource pool In load type The initial matching degree. Its calculation formula is: ; in, For resource pool For load type The capacity to handle; Business Category For load type The demand.

[0053] In one embodiment, the business and resource pool data in step 303 are taken as an example. The business category is 2 (regular elastic type, residential daily electricity consumption data statistics business); the resource pools are 1 (private cloud) and 2 (public cloud); the load type requirement is calculated as: computational load. 10,000 records (corresponding to computing resource requirements), storage load GB, network load Gbps. The resource pool can handle: private cloud computing loads. 10,000 items, storage load (100,000 GB), network load Gbps, public cloud computing load 10,000 items, storage load TB (500,000 GB), network load Gbps. Therefore, when calculating the elements of the initial matching matrix, For private cloud computing load: .

[0054] For private cloud storage load: ; For private cloud - network load is: ; For public cloud computing load: ; For public cloud storage load: ; For public cloud - network load is: ; The final initial matching matrix is ​​as follows: .

[0055] Step 3042: Based on the initial matching matrix of the business resource pool and the load migration cost coefficients for each business category migrating between different resource pools, construct the migration cost correction matching matrix.

[0056] Optionally, the load balancing scheduling system can migrate loads between different resource pools based on pre-set load migration cost coefficients for each service category. (i represents the business category, j represents the target resource pool, and k represents the load type, with values ​​ranging from 0 to 1; the larger the value, the higher the cost), combined with the initial matching matrix determined in step 3041. Construct a migration cost-corrected matching matrix The calculation formula is as follows: ; This allows for the correction of the initial matching degree using a cost coefficient, thereby reducing the matching degree weight corresponding to migration costs.

[0057] Continuing with the above embodiments, for conventional elastic services Its migration cost coefficient is preset as follows: For private cloud → private cloud (same pool scheduling): (Calculate the load) (Storage load) (Network load).

[0058] For the transition from private cloud to public cloud: (Calculate the load) (Storage load) (Network load).

[0059] Based on this, the elements of the corrected matching matrix are calculated as follows: For private cloud computing load: ; For private cloud storage load: ; For private cloud - network load is: ; For public cloud computing load: ; For public cloud storage load: ; For public cloud - network load is: ; The final corrected matching matrix is ​​constructed as follows: .

[0060] Step 3043: Based on the migration cost correction matching matrix and the load capacity threshold of each resource pool, determine the upper limit of the service load of each service category that the resource pool can accept.

[0061] Optionally, the load balancing scheduling system corrects the matching matrix based on the migration cost constructed in step 3042. Combined with the load capacity thresholds of each resource pool in step 303 (including the maximum computational load threshold) Maximum storage load threshold Maximum network load threshold The upper limit of service load that can be accepted by the resource pool for each service category. The calculation formula is as follows: ; In this formula, the denominator is the resource pool. In load type The sum of the corrected matching degree for all business categories is used to ensure that the load limit allocation is proportional to the matching degree.

[0062] Continuing with the above embodiments, the load capacity threshold of the private cloud resource pool is as follows: Core (corresponding to the maximum computing load) TB (Storage Load Cap) Gbps (Network load limit). Assuming only regular elastic services are currently in use. If scheduling is required, and the sum of the matching degrees for other business categories is 0, then: Private Cloud - Compute Load Limit: nuclear; Private Cloud - Storage Load Cap: TB; Private Cloud - Network Load Cap: Gbps; Public cloud resource pool load capacity threshold: nuclear, TB Gbps.

[0063] Similarly, the calculation yields: Public Cloud - Compute Load Limit: nuclear; Public Cloud - Storage Load Cap: TB; Public Cloud - Network Load Cap: Gbps.

[0064] Step 3044: Based on the business load limit and the current load status of each business category, determine the initial allocation value of the migration amount of the load type for each business category to migrate to different resource pools.

[0065] Optionally, the load balancing scheduling system determines the service load limit based on step 3043. Combined with the current load status of each service category in step 20 (excess load under overload conditions) Redundant load under low load conditions Finally, the initial allocation values ​​for the migration volume of each business category to different resource pools were determined. In overload conditions: Under low load conditions: .

[0066] Continuing with the above embodiments, for regular elastic services that are in an overloaded state, the excess load is calculated as follows: Calculate the load. 10,000 records (corresponding to 30 cores of computing resources), storage load GB, network load Gbps.

[0067] The initial migration allocation value is then calculated for the following when the load is migrated to the private cloud: nuclear; When migrating workloads to the public cloud: nuclear; When migrating network load to a private cloud: Gbps; When migrating network load to the public cloud: Gbps.

[0068] Among them, the storage load has an initial allocation value of approximately 0 due to its small excess amount and low matching degree.

[0069] Step 3045: Minimize the total migration cost based on the initial migration amount allocation value and the load migration cost coefficient to obtain the load migration allocation scheme for each business category.

[0070] Optionally, the load balancing scheduling system allocates initial values ​​based on the migration amount. Based on this, and combined with the load migration cost coefficient, a model for minimizing the total migration cost is constructed, with the objective function being: The load balancing scheduling system initially allocates the migration amount as specified in step 3044. Based on this, a model for minimizing total migration cost is constructed, with the objective function being: ; Among them, the constraints are: (Overload condition) or (Low load state), and Therefore, by solving this model, the load migration allocation scheme for each business category can be obtained. .

[0071] Continuing with the above embodiments, the objective function for optimizing the computational load migration of conventional elastic services is as follows: ; Constraints: , , Solving for the problem, we get: nuclear, (Due to the lower cost of migrating to the private cloud, more cores are prioritized for migration, while still meeting the constraint of a total migration volume of 20+10=30 cores). Similarly, network load migration is optimized, and the final allocation scheme is determined as follows: 10 cores of computing load and 0.256Gbps of network load are migrated from the private cloud to the public cloud, and 20 cores of computing load and 0.744Gbps of network load are scheduled within the private cloud.

[0072] This invention achieves precise matching between services and resource pools by constructing an initial matching matrix, balances matching degree and migration cost by combining a migration cost correction matrix, and finally optimizes the allocation scheme by minimizing the total migration cost model. Compared with traditional experience-based migration decisions, this scheme quantifies and integrates multiple dimensions such as business needs, resource status, and migration costs, making load migration decisions more scientific and interpretable. Under the premise of ensuring that the load of the resource pool does not exceed the carrying capacity threshold, it achieves dual optimization of resource utilization and migration cost through dynamic allocation of the upper limit of business load and cost optimization of migration volume. Finally, through matrix management and refined allocation of migration volume by load type (computing, storage, network), it achieves differentiated migration scheduling for different types of loads, avoiding business instability caused by excessive resource consumption by a single load type.

[0073] Furthermore, the power service load balancing scheduling system based on a hybrid cloud architecture provided by the present invention will be described below. The power service load balancing scheduling system based on a hybrid cloud architecture described below can be referred to in correspondence with the power service load balancing scheduling method based on a hybrid cloud architecture described above.

[0074] Optional, refer to Figure 2 , Figure 2 This is a schematic diagram of the power service load balancing scheduling system based on a hybrid cloud architecture provided by the present invention. The power service load balancing scheduling system based on a hybrid cloud architecture includes: The business classification module 210 is used to classify the real-time collected power business operation data based on the importance of the power business to obtain the target power business category; the importance of the power business is determined by the degree of impact on the operation of the power system, the level of data security requirements, and the characteristics of load fluctuation. The load status determination module 220 is used to determine the current load status of the target power service category based on the target power service category and the corresponding power service operation data, combined with the resource status information of the private cloud resource pool and the public cloud resource pool. The load scheduling module 230 is used to determine the load scheduling strategy based on the current load status for the target power service category.

[0075] This invention classifies real-time power service operation data based on the importance of power services to the power system, clearly defining the current power service category. By combining resource status information from private and public cloud resource pools, it accurately determines the current load status of various power services, breaking the limitation of relying solely on local physical server clusters for resource information. Finally, based on the current load status and target service type, it determines the corresponding load scheduling strategy, achieving targeted and flexible resource allocation. In summary, this invention effectively solves the problems of fixed local physical server cluster resource capacity in existing methods, which makes it difficult to cope with sudden peak loads and easily leads to service response delays and data processing interruptions, ensuring the stable operation of power services under various load conditions. Simultaneously, it improves the overall resource utilization rate of the cloud resource pool, fully meeting the differentiated security and elasticity requirements of different types of power services, and ensuring the stable and efficient operation of the power system.

[0076] Please see Figure 3 , Figure 3 An embodiment diagram of an electronic device provided in accordance with the present invention. For example... Figure 3 As shown, this embodiment of the invention provides an electronic device 300, including a memory 310, a processor 320, and a computer program 311 stored in the memory 310 and executable on the processor 320. When the processor 320 executes the computer program 311, it performs the following steps: Based on the importance of power services, the real-time collected power service operation data are classified to obtain the target power service category; the importance of power services is determined by the degree of impact on power system operation, data security requirements, and load fluctuation characteristics. Based on the target power service category and the corresponding power service operation data, combined with the resource status information of the private cloud resource pool and the public cloud resource pool, the current load status of the target power service category is determined. For the target power service category, determine the load scheduling strategy based on the current load status.

[0077] Please see Figure 4 , Figure 4 An embodiment diagram of a computer-readable storage medium provided in accordance with an embodiment of the present invention is shown. Figure 4 As shown, this embodiment provides a computer-readable storage medium 400 on which a computer program 311 is stored. When the computer program 311 is executed by a processor, it performs the following steps: Based on the importance of power services, the real-time collected power service operation data are classified to obtain the target power service category; the importance of power services is determined by the degree of impact on power system operation, data security requirements, and load fluctuation characteristics. Based on the target power service category and the corresponding power service operation data, combined with the resource status information of the private cloud resource pool and the public cloud resource pool, the current load status of the target power service category is determined. For the target power service category, determine the load scheduling strategy based on the current load status.

[0078] On the other hand, the present invention also provides a computer program product, which includes a computer program that can be stored on a non-transitory computer-readable storage medium. When the computer program is executed by a processor, the computer is able to execute the power service load balancing scheduling method based on a hybrid cloud architecture provided by the above methods. The method includes: Based on the importance of power services, the real-time collected power service operation data are classified to obtain the target power service category; the importance of power services is determined by the degree of impact on power system operation, data security requirements, and load fluctuation characteristics. Based on the target power service category and the corresponding power service operation data, combined with the resource status information of the private cloud resource pool and the public cloud resource pool, the current load status of the target power service category is determined. For the target power service category, determine the load scheduling strategy based on the current load status.

[0079] The system embodiments described above are merely illustrative. The units described as separate components may or may not be physically separate, and the components shown as units may or may not be physical units; that is, they may be located in one place or distributed across multiple network units. Some or all of the modules can be selected to achieve the purpose of this embodiment according to actual needs. Those skilled in the art can understand and implement this without any creative effort.

[0080] Through the above description of the embodiments, those skilled in the art can clearly understand that each embodiment can be implemented by means of software plus necessary general-purpose hardware platforms, and of course, it can also be implemented by hardware. Based on this understanding, the above technical solutions, in essence or the part that contributes to the prior art, can be embodied in the form of a software product. This computer software product can be stored in a computer-readable storage medium, such as ROM / RAM, magnetic disk, optical disk, etc., including several instructions to cause a computer device (which may be a personal computer, server, or network device, etc.) to execute the methods of various embodiments or some parts of embodiments.

[0081] Finally, it should be noted that the above embodiments are only used to illustrate the technical solutions of the present invention, and not to limit them; although the present invention has been described in detail with reference to the foregoing embodiments, those skilled in the art should understand that modifications can still be made to the technical solutions described in the foregoing embodiments, or equivalent substitutions can be made to some of the technical features; and these modifications or substitutions do not cause the essence of the corresponding technical solutions to deviate from the spirit and scope of the technical solutions of the embodiments of the present invention.

Claims

1. A power service load balancing scheduling method based on a hybrid cloud architecture, characterized in that, The hybrid cloud architecture includes a private cloud resource pool and a public cloud resource pool, including: Based on the importance of power services, the real-time collected power service operation data are classified to obtain the target power service category; the importance of power services is determined by the degree of impact on power system operation, data security requirements, and load fluctuation characteristics. Based on the target power service category and the corresponding power service operation data, and combined with the resource status information of the private cloud resource pool and the public cloud resource pool, the current load status of the target power service category is determined. For the target power service category, a load scheduling strategy is determined based on the current load status.

2. The power service load balancing scheduling method based on a hybrid cloud architecture according to claim 1, characterized in that, The target power business categories include core security businesses, conventional flexible businesses, and non-core extended businesses; The core security services are carried only by the private cloud resource pool, the regular elastic services are carried by the private cloud resource pool / public cloud resource pool, and the non-core extended services are carried only by the public cloud resource pool.

3. The power service load balancing scheduling method based on a hybrid cloud architecture according to claim 2, characterized in that, The determination of a load scheduling strategy based on the current load status for the target power service category includes: Based on the operational data of the target power service category and the resource status information of the private cloud resource pool and public cloud resource pool, the resource demand matching coefficient corresponding to each service type is determined. Based on the current load status of the target power service category and the resource demand matching coefficient corresponding to each service category, the load scheduling priority coefficient of each service category is determined; the current load status includes overload status and low load status. Based on the load scheduling priority coefficients of each business category and the load carrying thresholds of private cloud resource pools and public cloud resource pools, the business load that each resource pool can handle is determined; the load carrying thresholds include the maximum computing load carrying threshold, the maximum storage load carrying threshold, and the maximum network load carrying threshold. Based on the service load capacity that each resource pool can handle and the load migration cost coefficient for each service category to migrate between different resource pools, determine the load migration allocation scheme for each service category. The load scheduling strategy is determined based on the load migration and allocation schemes for each business category and the real-time resource change rates of the private cloud resource pool and the public cloud resource pool.

4. The power service load balancing scheduling method based on a hybrid cloud architecture according to claim 3, characterized in that, The resource status information includes the remaining computing resources, the remaining storage resources, and the network bandwidth; The process of determining the resource demand matching coefficients corresponding to each business type includes: If the target power business category is a core security business, then only the matching degree with the private cloud resource pool is determined. The formula for calculating the matching coefficient of its resource requirements with the private cloud resource pool is as follows: ； in, This represents the matching coefficient between the resource requirements of core security-related services and the private cloud resource pool; This indicates the amount of data transmitted for core security-related services; This indicates the computational workload of core security-related tasks; This indicates the amount of remaining computing resources in the private cloud resource pool; This indicates the amount of remaining storage resources in the private cloud resource pool; This indicates the network bandwidth of the private cloud resource pool; If the target power service category is a regular elastic service, then the matching degree with the private cloud resource pool and the public cloud resource pool should be determined separately. The formula for calculating the resource demand matching coefficient with the private cloud resource pool is as follows: ； The formula for calculating the resource demand matching coefficient with the public cloud resource pool is as follows: ； in, This represents the matching coefficient between the resource requirements of regular elastic services and the private cloud resource pool; This represents the matching coefficient between the resource requirements of regular elastic services and the public cloud resource pool; This indicates the amount of data transmitted in a typical elastic service. This indicates the amount of computational tasks required for typical elastic business operations; This indicates the amount of remaining computing resources in the public cloud resource pool; This indicates the amount of remaining storage resources in the public cloud resource pool; This indicates the network bandwidth of the public cloud resource pool. If the target power business category is a non-core extended business, then only the matching degree with the public cloud resource pool is determined. The formula for calculating the matching coefficient of its resource demand with the public cloud resource pool is as follows: ； in, This represents the matching coefficient between the resource requirements of non-core extended business functions and the public cloud resource pool: This indicates the amount of data transmitted for non-core extended business functions; This indicates the amount of computational tasks required for non-core extended business functions.

5. The power service load balancing scheduling method based on a hybrid cloud architecture according to claim 3, characterized in that, When the target power service category is in an overload state, the formula for calculating its load scheduling priority coefficient is as follows: ； in This indicates the load scheduling priority coefficient for service categories under overload conditions. Indicates the importance coefficient of the business category (core security business). Regular elastic business Non-core extended business ); This indicates the resource requirement matching coefficient corresponding to this business category; This indicates the current load (under overload conditions) for this service category. The value is the portion of the actual load that exceeds the rated load.

6. The power service load balancing scheduling method based on a hybrid cloud architecture according to claim 3, characterized in that, When the aforementioned service category is under low load, its load scheduling priority coefficient is calculated using the following formula: ； in, This indicates the load scheduling priority coefficient for a service category under low load conditions. Indicates the importance coefficient of the business category; This indicates the resource requirement matching coefficient corresponding to this business category; This indicates the maximum load capacity for this business category; This indicates the current load of this business category.

7. The power service load balancing scheduling method based on a hybrid cloud architecture according to claim 3, characterized in that, The determination of the load migration allocation scheme for each service category, based on the service load capacity that each resource pool can handle and the load migration cost coefficient for migrating each service category between different resource pools, includes: Based on the business load capacity that each resource pool can handle and the different types of load requirements of each business category, an initial matching matrix for business resource pools is constructed. Based on the initial matching matrix of the business resource pool and the load migration cost coefficient of each business category migrating between different resource pools, a migration cost correction matching matrix is ​​constructed. Based on the migration cost correction matching matrix and the load capacity threshold of each resource pool, the upper limit of the service load of each service category that the resource pool can accept is determined. Based on the service load limit and the current load status of each service category, determine the initial allocation value of the migration amount of each service category to different resource pools. Based on the initial migration allocation value and the load migration cost coefficient, the total migration cost is minimized to obtain the load migration allocation scheme for each business category.

8. A power service load balancing and dispatching system based on a hybrid cloud architecture, characterized in that, Applied to the power service load balancing scheduling method based on a hybrid cloud architecture as described in any one of claims 1 to 7; The power service load balancing and scheduling system based on a hybrid cloud architecture includes: The business classification module is used to classify the real-time collected power business operation data based on the importance of the power business to obtain the target power business category; the importance of the power business is determined by the degree of impact on the operation of the power system, the level of data security requirements, and the characteristics of load fluctuation. The load status determination module is used to determine the current load status of the target power service category based on the target power service category and the corresponding power service operation data, combined with the resource status information of the private cloud resource pool and the public cloud resource pool. The load scheduling module is used to determine a load scheduling strategy based on the current load status for the target power service category.

9. An electronic device, characterized in that, include: Memory, used to store computer software programs; A processor is configured to read and execute the computer software program, wherein when the processor executes the computer software program, it implements the power service load balancing scheduling method based on a hybrid cloud architecture as described in any one of claims 1 to 7.

10. A non-transitory computer-readable storage medium, characterized in that, The storage medium stores a computer software program, which, when executed by a processor, implements the power service load balancing scheduling method based on a hybrid cloud architecture as described in any one of claims 1 to 7.

Images