A multi-cloud heterogeneous computing power scheduling system and scheduling method

Abstract

The present invention relates to the technical field of computing power resource allocation data processing, and discloses a multi-cloud heterogeneous computing power scheduling system and scheduling method, wherein the system comprises a computing power capacity analysis module, a cloud computing power platform object evaluation module, and a computing power scheduling execution module; the target cloud computing power platform type is efficiently analyzed and processed by standard storage of cloud platform objects corresponding to different computing power capacities combined with an intelligent search algorithm and computing power capacity parameters of tasks to be processed, so as to realize accurate search of target cloud computing power platform objects based on multi-cloud heterogeneity; the service prices of multiple target cloud computing power platform objects are digitally analyzed, and the cloud computing power platform with the minimum service price is scientifically analyzed, so as to realize accurate screening of the best cloud computing power platform object that meets the tasks to be processed, thereby improving the accuracy and reliability of scheduling operations of the computing power scheduling system, and improving the scheduling operation quality of the computing power scheduling system.

Description

Technical Field

[0001] The present invention relates to the technical field of computing resource allocation data processing, and specifically to a multi-cloud heterogeneous computing resource scheduling system and scheduling method. Background Art

[0002] Computing power scheduling refers to the process of rationally allocating and utilizing computer resources in a certain area or system. Its main purpose is to improve the utilization rate of computer resources, reduce resource waste, and ensure efficient execution of tasks. That is, based on the brain of the computing network, intelligent arrangement and flexible scheduling of computing power resources in the entire network are carried out. Computing power scheduling is similar to power scheduling. The power system scheduling is a complex system in which many power plants provide electricity and supply electricity to a large number of users through transmission, substation, distribution, and power supply networks. Its production, supply, and sales processes are completed and balanced at the same time in an instant. The computing power center can be compared to a power plant, the communication network to the power network, and the users of computing power are large models, applications, and other manufacturers. Computing power scheduling is to balance the demand and supply of computing power through the scheduling of computing power. At present, computing power scheduling is still limited to manual scheduling and local computing power scheduling. With the development of big data, artificial intelligence, and 5G communications, the construction of a scientific and intelligent computing power scheduling platform has become a very urgent issue; however, the existing computing power scheduling platform cannot accurately evaluate the computing power capacity required for the tasks to be processed, nor can it select the best cloud computing power platform for the tasks to be processed.

[0003] The Chinese invention patent with announcement number CN117785482B discloses a computing power scheduling system and method for a computing power network, which performs predictive analysis on data requests of a target data source within a preset time period in the future to obtain a corresponding trend prediction index; receives data requests from the target data source, and sends the data requests to a computing power resource scheduling module; performs computing power scheduling allocation calculations based on the trend prediction index, the data request, and the specific conditions of the computing power network module; the computing power network module performs computing power allocation and calculations based on the results of the computing power scheduling allocation calculations. According to the historical data request situation, a predictive analysis is performed on data requests for a preset time in the future, and the obtained trend prediction index is incorporated into the computing power allocation process, and the influence of the computing power network device itself and the specific category of the data request is considered during the computing power allocation, so as to fully improve the efficiency and ability of computing power resource mobilization. However, the above technical solutions cannot scientifically select the best computing power execution platform based on the data processing object. Summary of the invention

[0004] 1. Technical issues to be resolved

[0005] In order to solve the problem that the above-mentioned existing computing power scheduling platform cannot accurately evaluate the computing power capacity required for the tasks to be processed, nor can it screen the best cloud computing power platform for the tasks to be processed, the above purpose of accurately entering the characteristic information of the tasks to be processed, scientifically analyzing the computing power capacity parameters of the tasks to be processed, intelligently searching for the cloud computing power platform required for the tasks to be processed, accurately screening the best cloud computing power platform, efficiently and accurately executing computing power scheduling operations, and improving the efficiency and accuracy of computing power scheduling is achieved.

[0006] (II) Technical solution

[0007] The present invention is implemented by the following technical solution: a multi-cloud heterogeneous computing power scheduling method, the method comprising the following steps:

[0008] S1. Collect characteristic data of the task to be processed;

[0009] S2. Performing identification processing on the computing capacity required for the task to be processed based on the feature data of the task to be processed and the computing capacity data corresponding to the features of different tasks, and generating computing capacity data for the task to be processed;

[0010] S3, analyzing and processing the cloud computing platform object type required for the task to be processed according to the computing capacity data of the task to be processed and the cloud computing platform object data corresponding to different computing capacity, and constructing the cloud computing platform object data of the task to be processed;

[0011] S4, based on the pending task cloud computing platform object data and different cloud computing platform object service price data, perform service price search processing on the pending task cloud computing platform to generate pending task cloud computing platform object service price data;

[0012] S5. Analyze and process the best cloud computing platform object for the task to be processed based on the cloud computing platform object data for the task to be processed and the service price data of the cloud computing platform object for the task to be processed, and construct the best cloud computing platform object data for the task to be processed;

[0013] S6. Perform a search and processing of the operation instruction of the cloud computing platform for the task to be processed according to the object data of the best cloud computing platform for the task to be processed and the operation instruction data of the cloud computing platform object, and generate the operation instruction data of the best cloud computing platform object;

[0014] S7. Execute the computing power scheduling operation of the task to be processed according to the characteristic data of the task to be processed and the running instruction data of the optimal cloud computing power platform object.

[0015] Preferably, the steps of collecting the feature data of the task to be processed are as follows:

[0016] S11. Collect the characteristic information of the data task to be processed online through the data input dialog box and generate the characteristic data of the task to be processed The characteristic data of the task to be processed includes data type, data capacity, data task processing field and preset processing time length; the data type includes any one of text, video, image, and audio, and the data task processing field includes any one of banking, e-commerce, health, education, communication, and weather.

[0017] Preferably, the computing capacity required for the task to be processed is identified based on the feature data of the task to be processed and the computing capacity data corresponding to the features of different tasks, and the operation steps for generating the computing capacity data of the task to be processed are as follows:

[0018] S21. Establish a data set of computing capacity corresponding to the characteristics of different tasks , ;in Indicates The characteristics of different tasks correspond to the computing capacity data. Indicates the maximum value of the number of different task feature types. The computing capacity data corresponding to the features of different tasks indicates the minimum value of the computing capacity required for data processing of different task features. The unit of is operations per second;

[0019] S22: The task feature data to be processed Computing capacity data set corresponding to the characteristics of the different tasks The characteristics of different tasks described in the table correspond to the computing capacity data Perform task feature data matching to search for the task feature data to be processed The corresponding computing capacity data of the characteristics of the different tasks And generate the computing capacity data of the task to be processed through data identification , execute and generate the computing capacity data of the task to be processed The specific steps are as follows:

[0020] S221, initialization , that is, the characteristics of different tasks correspond to the computing capacity data set The characteristics of different tasks correspond to the computing capacity data Set as ; Set the number of nectar sources , the current number of iterations t and the maximum number of iterations T; Initial position , The formula for generating the initial position is as follows: Computing capacity data set corresponding to the characteristics of different tasks The search space is randomly generated, where express The random number inside and Represents the computing capacity data set corresponding to the characteristics of different tasks Upper and lower bounds of the search space;

[0021] S222. Computing capacity data set corresponding to the characteristics of different tasks The search space contains the computing capacity data corresponding to the features of different tasks. Honey source location Assign hired bees. At the beginning of the search, hired bees are at the location of the nectar source. Around, according to the formula Search for new nectar sources , that is, the computing capacity data set corresponding to the characteristics of different tasks Search the search space for the feature data of the task to be processed Computing capacity data corresponding to the characteristics of different tasks ;in Indicated in Randomly select a nectar source that is not equal to i from the nectar sources. Represents a randomly selected nectar source Computing capacity data set corresponding to the characteristics of different tasks The position in the search space, is a random number uniformly distributed between [-1,1];

[0022] S223, based on the location of the new nectar source The fitness of the task is determined by the greedy selection method to determine the retained honey source, that is, the corresponding computing capacity data set of the features of different tasks is retained. Search the search space for the feature data of the task to be processed Computing capacity data corresponding to the characteristics of different tasks with the largest fitness values ;

[0023] S224, employed bees share nectar information, and follower bees correspond to the computing capacity data set according to the probability values ​​of different tasks. Search the search space for the feature data of the task to be processed Computing capacity data corresponding to the characteristics of different tasks Honey source, follow the bee according to the honey source information shared by the employed bees, corresponding to the computing capacity data set of different tasks The search space searches for characteristic data related to the task to be processed. Computing capacity data corresponding to the characteristics of different tasks ; According to the probability formula , calculate the characteristic data of the task to be processed found by the employed bees Computing capacity data corresponding to the characteristics of different tasks The probability of a nectar source being followed, where Indicates the location of a new nectar source The probability of being found and followed by hired bees, Indicates the location of the new nectar source shared by the hired bees according to the follower bees Adaptability;

[0024] S225, follower bees search in the same way as employed bees, and determine the nectar source to be retained based on the greedy selection method. , that is, retain the corresponding computing capacity data set of features in different tasks Search the search space for the feature data of the task to be processed Computing capacity data corresponding to the characteristics of different tasks ;

[0025] S226. Data set of computing capacity corresponding to the characteristics of scout bees in different tasks Random search in the search space and the feature data of the task to be processed Computing capacity data corresponding to the characteristics of different tasks If the location of the nectar source After t iterations, the search reaches the iteration threshold No better honey source was found , which is related to the feature data of the task to be processed Computing capacity data corresponding to the characteristics of different tasks Honey source location The hired bees will be abandoned, and the corresponding hired bees will become scout bees; the characteristics of scout bees in different tasks correspond to the computing capacity data set The search space randomly generates the feature data of the task to be processed Computing capacity data corresponding to the characteristics of different tasks The new nectar source location, the new location calculation formula is as follows: ,in Indicated in After iterations, a new Computing capacity data set corresponding to the characteristics of different tasks The location of the search space, Indicated in After iterations, a new Computing capacity data set corresponding to the characteristics of different tasks The location of the search space;

[0026] S227, when the maximum number of iterations T is met, output the feature data of the task to be processed Computing capacity data corresponding to the characteristics of different tasks And generate the computing capacity data of the task to be processed through data identification , The unit is operations per second.

[0027] Preferably, the cloud computing platform object type required for the task to be processed is analyzed and processed according to the computing capacity data of the task to be processed and the cloud computing platform object data corresponding to different computing capacity, and the operation steps for constructing the cloud computing platform object data of the task to be processed are as follows:

[0028] S31. Establishing cloud computing platform object data sets corresponding to different computing capacity , ;in Indicates The different computing capacities corresponding to the computing capacity correspond to the cloud computing platform object data. Indicates the maximum value of the number of computing capacity types. The cloud computing platform object data corresponding to different computing capacity types indicates the cloud computing platform object data corresponding to different computing capacity required for data processing tasks. The cloud computing platform object types include Baidu Cloud, Alibaba Cloud, Huawei Cloud and Tencent Cloud.

[0029] S32, using a breadth-first algorithm to convert the computing capacity data of the task to be processed Cloud computing platform object data set corresponding to the different computing capacity The different computing capacity described in the above corresponds to the cloud computing platform object data According to the comparison of computing power capacity values, search for the computing power capacity data of the task to be processed The corresponding different computing capacity corresponds to the cloud computing platform object data And through data identification, a data set of cloud computing platform objects to be processed is constructed , ,in Indicates The cloud computing platform object data of the tasks to be processed. Indicates The cloud computing platform object data of the tasks to be processed.

[0030] Preferably, based on the pending task cloud computing platform object data and different cloud computing platform object service price data, the service price search process of the pending task cloud computing platform is performed to generate the pending task cloud computing platform object service price data as follows:

[0031] S41. Establish a data set of service prices for different cloud computing platforms ,in Indicates the cloud computing platform object data corresponding to different computing capacity The corresponding service price data of different cloud computing platform objects, wherein the service price data of different cloud computing platform objects represent the service price data of different cloud computing platforms for data processing tasks, The unit is yuan;

[0032] S42, using a bidirectional search algorithm to collect the object data of the task cloud computing platform to be processed The cloud computing platform object data of the pending tasks described in to A collection of service price data for different cloud computing platforms The price data of the service objects of different cloud computing platforms described in According to the cloud computing platform object character matching, search for the cloud computing platform object data of the task to be processed to The corresponding service price data of different cloud computing platforms And generate the service price data set of the cloud computing platform object to be processed through data identification ,in Represents the cloud computing platform object data of the task to be processed The corresponding service price data of the cloud computing platform object for the task to be processed, Represents the cloud computing platform object data of the task to be processed The corresponding service price data of the cloud computing platform object for the task to be processed, and The unit is Yuan.

[0033] Preferably, the optimal cloud computing platform object for the task to be processed is analyzed and processed based on the cloud computing platform object data for the task to be processed and the service price data of the cloud computing platform object for the task to be processed, and the operation steps for constructing the optimal cloud computing platform object data for the task to be processed are as follows:

[0034] S51: collect the service price data of the cloud computing platform object for the task to be processed The service price data of the cloud computing platform object to be processed are compared with the service price values, and the service price data of the cloud computing platform object to be processed with the smallest value is analyzed and marked as the minimum service price data of the cloud computing platform object to be processed. , The unit mean of ;

[0035] S52: The minimum service price data of the cloud computing platform object of the task to be processed According to the cloud computing platform object character matching and the cloud computing platform object data set of the task to be processed The cloud computing platform object data of the pending task is matched with the best cloud computing platform object of the pending task, and the minimum service price data of the cloud computing platform object of the pending task is searched out. The corresponding cloud computing platform object data of the task to be processed is constructed through data identification to obtain the optimal cloud computing platform object data of the task to be processed. .

[0036] Preferably, the operation instruction search process of the cloud computing platform for the task to be processed is performed according to the optimal cloud computing platform object data of the task to be processed and the cloud computing platform object operation instruction data, and the operation steps of generating the optimal cloud computing platform object operation instruction data are as follows:

[0037] S61. Establishing a cloud computing platform object operation instruction data set ,in Indicates the cloud computing platform object data corresponding to different computing capacity The corresponding cloud computing platform object operation instruction data, wherein the cloud computing platform object operation instruction data represents the program instruction data used by the computing power scheduling center to control the specific cloud computing platform to perform data processing tasks;

[0038] S62, using a depth-restricted algorithm to convert the optimal cloud computing platform object data of the task to be processed A set of command data for running the cloud computing platform object The cloud computing platform object running instruction data described in According to the cloud computing platform object character matching, search for the best cloud computing platform object data for the task to be processed The corresponding cloud computing platform object running instruction data And generate the best cloud computing platform object running instruction data through data identification .

[0039] Preferably, the operation steps of executing the computing power scheduling operation of the task to be processed according to the feature data of the task to be processed and the operation instruction data of the optimal cloud computing power platform object are as follows:

[0040] S71, the computing power dispatching center runs the instruction data according to the optimal cloud computing power platform object Control the optimal cloud computing platform object data for the task to be processed The corresponding cloud computing platform runs and executes the characteristic data of the task to be processed The computing power scheduling job of the pending tasks.

[0041] A multi-cloud heterogeneous computing power scheduling system, used to implement the multi-cloud heterogeneous computing power scheduling method, the system includes a computing power capacity analysis module, a cloud computing power platform object evaluation module, and a computing power scheduling execution module;

[0042] The computing capacity analysis module includes a task feature information collection unit, a computing capacity storage unit corresponding to different task features, and a task computing capacity identification unit.

[0043] The task feature information collection unit to be processed collects the task feature data to be processed through a data input dialog box; the different task feature corresponding computing power capacity storage unit is used to store the feature corresponding computing power capacity data of different tasks; the task feature capacity identification unit to be processed performs the required computing power capacity identification processing for the task to be processed based on the task feature data to be processed and the feature corresponding computing power capacity data of different tasks, and generates the computing power capacity data for the task to be processed;

[0044] The cloud computing platform object evaluation module includes a cloud computing platform object storage unit corresponding to different computing power capacities, a cloud computing platform object analysis unit for pending tasks, a storage unit for different cloud computing platform object service prices, a cloud computing platform object service price search unit for pending tasks, and an optimal cloud computing platform object identification unit for pending tasks;

[0045] The cloud computing platform object storage unit corresponding to different computing power capacities stores cloud computing platform object data corresponding to different computing power capacities based on multi-cloud heterogeneity; the cloud computing platform object analysis unit for pending tasks performs analysis and processing on the cloud computing platform object type required for pending tasks according to the computing power capacity data of pending tasks and the cloud computing platform object data corresponding to different computing power capacities, and constructs the cloud computing platform object data for pending tasks; the different cloud computing platform object service price storage unit is used to store different cloud computing platform object service price data; the cloud computing platform object service price search unit for pending tasks performs service price search processing on the cloud computing platform for pending tasks based on the cloud computing platform object data for pending tasks and the service price data of different cloud computing platform objects, and generates service price data of the cloud computing platform object for pending tasks; the optimal cloud computing platform object identification unit for pending tasks performs analysis and processing on the optimal cloud computing platform object for pending tasks according to the cloud computing platform object data for pending tasks and the service price data of the cloud computing platform object for pending tasks, and constructs the optimal cloud computing platform object data for pending tasks;

[0046] The computing power scheduling execution module includes a cloud computing power platform object running instruction storage unit, a cloud computing power platform object running instruction search unit, and a pending task computing power scheduling job execution unit;

[0047] The cloud computing power platform object running instruction storage unit is used to store cloud computing power platform object running instruction data; the cloud computing power platform object running instruction search unit searches for the running instructions of the cloud computing power platform for the task to be processed based on the optimal cloud computing power platform object data for the task to be processed and the cloud computing power platform object running instruction data, and generates the optimal cloud computing power platform object running instruction data; the computing power scheduling job execution unit for the task to be processed executes the computing power scheduling job for the task to be processed based on the characteristic data of the task to be processed and the optimal cloud computing power platform object running instruction data.

[0048] (III) Beneficial effects

[0049] The present invention provides a multi-cloud heterogeneous computing power scheduling system and scheduling method. It has the following beneficial effects:

[0050] 1. Through the task feature information collection unit, the data input dialog box is used to conveniently obtain the task feature information online, so as to provide real data support for the subsequent accurate assessment of the computing capacity requirements of the tasks to be processed; the computing capacity storage unit corresponding to different task features and the computing capacity identification unit for the tasks to be processed cooperate with each other, and the computing capacity parameters corresponding to different task features are scientifically preset. The intelligent identification algorithm and the task feature information to be processed are combined to conduct a scientific analysis of the computing capacity required for the tasks to be processed, so as to realize the accurate assessment of the target computing capacity in the computing power scheduling process and realize the intelligent evaluation of the computing power of the computing power scheduling system scheduling operations.

[0051] 2. Through the cooperation of cloud computing platform object storage units corresponding to different computing power capacities and cloud computing platform object analysis units for pending tasks, standard storage of cloud platform objects corresponding to different computing power capacities is combined with intelligent search algorithms and computing power capacity parameters of pending tasks to efficiently analyze and process the target cloud computing platform type, so as to achieve accurate search of target cloud computing platform objects based on multi-cloud heterogeneity; different cloud computing platform object service price storage units, pending task cloud computing platform object service price search units and pending task best cloud computing platform object identification units cooperate with each other to perform digital analysis of multiple target cloud computing platform object service prices, scientifically analyze the cloud computing platform with the lowest service price, and accurately screen the best cloud computing platform object that meets the pending tasks, thereby improving the accuracy and reliability of the computing power scheduling system's scheduling operations and the scheduling operation quality of the computing power scheduling system.

[0052] 3. Through the cooperation between the cloud computing platform object operation instruction storage unit and the cloud computing platform object operation instruction search unit, the cloud computing platform object operation instruction parameters are scientifically preset, combined with the intelligent search algorithm and the optimal cloud computing platform object parameters of the task to be processed to autonomously retrieve the operation instructions of the target optimal cloud computing platform, so as to achieve efficient matching of the target optimal cloud computing platform operation instructions; the task computing power scheduling job execution unit autonomously and dynamically executes the computing power scheduling job of the task to be processed according to the characteristic parameters of the task to be processed and the target optimal cloud computing power platform object operation instructions, thereby improving the execution accuracy and efficiency of the computing power scheduling job and improving the scientific nature of the computing power scheduling. BRIEF DESCRIPTION OF THE DRAWINGS

[0053] Figure 1 A module schematic diagram of a multi-cloud heterogeneous computing power scheduling system provided by the present invention;

[0054] Figure 2 A flowchart of a multi-cloud heterogeneous computing power scheduling method provided by the present invention. DETAILED DESCRIPTION

[0055] The following will be combined with the drawings in the embodiments of the present invention to clearly and completely describe the technical solutions in the embodiments of the present invention. Obviously, the described embodiments are only part of the embodiments of the present invention, not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without creative work are within the scope of protection of the present invention.

[0056] The embodiments of the multi-cloud heterogeneous computing power scheduling system and scheduling method are as follows:

[0057] Embodiment 1:

[0058] See also Figure 1 - Figure 2 , a multi-cloud heterogeneous computing power scheduling method, the method comprising the following steps:

[0059] S1. Collect characteristic data of the task to be processed;

[0060] S2. Identify the computing capacity required for the task to be processed based on the feature data of the task to be processed and the computing capacity data corresponding to the features of different tasks, and generate computing capacity data for the task to be processed;

[0061] S3. Analyze and process the cloud computing platform object type required for the task to be processed according to the computing capacity data of the task to be processed and the cloud computing platform object data corresponding to different computing capacity, and construct the cloud computing platform object data of the task to be processed;

[0062] S4, based on the object data of the cloud computing platform for the task to be processed and the service price data of the objects of different cloud computing platforms, a service price search process is performed on the cloud computing platform for the task to be processed, and the service price data of the object of the cloud computing platform for the task to be processed is generated;

[0063] S5. Analyze and process the best cloud computing platform object for the task to be processed based on the cloud computing platform object data for the task to be processed and the service price data of the cloud computing platform object for the task to be processed, and construct the best cloud computing platform object data for the task to be processed;

[0064] S6. Search and process the operation instruction of the cloud computing platform for the task to be processed according to the object data of the best cloud computing platform for the task to be processed and the operation instruction data of the cloud computing platform object, and generate the operation instruction data of the best cloud computing platform object;

[0065] S7. Execute the computing power scheduling operation of the task to be processed based on the feature data of the task to be processed and the object running instruction data of the optimal cloud computing power platform.

[0066] For further information, see Figure 1 - Figure 2 ,The steps for collecting the feature data of the task to be processed are as follows:

[0067] S11. Collect the characteristic information of the data task to be processed online through the data input dialog box and generate the characteristic data of the task to be processed The characteristic data of the task to be processed include data type, data capacity, data task processing field and preset processing time length; the data type includes any one of text, video, image, and audio, and the data task processing field includes any one of banking, e-commerce, health, education, communication, and weather.

[0068] Based on the feature data of the task to be processed and the feature corresponding computing capacity data of different tasks, the computing capacity required for the task to be processed is identified and processed. The operation steps for generating the computing capacity data of the task to be processed are as follows:

[0069] S21. Establish a data set of computing capacity corresponding to the characteristics of different tasks , ;in Indicates The characteristics of different tasks correspond to the computing capacity data. Indicates the maximum number of different task feature types. The computing capacity data corresponding to the features of different tasks indicates the minimum computing capacity required for data processing of different task features. The unit of is operations per second;

[0070] S22: The feature data of the task to be processed Computing capacity data set corresponding to the characteristics of different tasks The characteristics of different tasks correspond to the computing capacity data Match task feature data and search for task feature data to be processed The corresponding computing capacity data of the characteristics of different tasks And generate the computing capacity data of the task to be processed through data identification , execute and generate computing capacity data of pending tasks The specific steps are as follows:

[0071] S221, initialization , that is, the characteristics of different tasks correspond to the computing capacity data set The characteristics of different tasks correspond to the computing capacity data Set as ; Set the number of nectar sources , the current number of iterations t and the maximum number of iterations T; Initial position , The formula for generating the initial position is as follows: Computing capacity data set corresponding to the characteristics of different tasks The search space is randomly generated, where express The random number inside, and Represents the computing capacity data set corresponding to the characteristics of different tasks Upper and lower bounds of the search space;

[0072] S222. Computing capacity data set corresponding to the characteristics of different tasks The search space contains the computing capacity data corresponding to the features of different tasks. Honey source location Assign hired bees. At the beginning of the search, hired bees are at the location of the nectar source. Around, according to the formula Search for new nectar sources , that is, the computing capacity data set corresponding to the characteristics of different tasks Search the search space to find the feature data of the task to be processed Computing capacity data corresponding to the characteristics of different tasks ;in Indicated in Randomly select a nectar source that is not equal to i from the nectar sources. Represents a randomly selected nectar source Computing capacity data set corresponding to the characteristics of different tasks The position in the search space, is a random number uniformly distributed between [-1,1];

[0073] S223, based on the location of the new nectar source The fitness of , that is, retain the corresponding computing capacity data set of features in different tasks Search the search space to find the feature data of the task to be processed Computing capacity data corresponding to the characteristics of different tasks with the largest fitness values ;

[0074] S224, employed bees share nectar information, and follower bees correspond to the computing capacity data set according to the probability values ​​of different tasks. Search the search space to find the feature data of the task to be processed Computing capacity data corresponding to the characteristics of different tasks Honey source, follow the bee according to the honey source information shared by the employed bees, corresponding to the computing capacity data set of different tasks The search space searches for characteristic data related to the task to be processed. Computing capacity data corresponding to the characteristics of different tasks Honey source information; according to probability formula , calculate the characteristic data of the tasks to be processed found by the employed bees Computing capacity data corresponding to the characteristics of different tasks The probability of a nectar source being followed, where Indicates the location of a new nectar source The probability of being found and followed by hired bees, Indicates the location of the new nectar source shared by the hired bees according to the follower bees Adaptability;

[0075] S225, follower bees search in the same way as employed bees, and determine the retained ones according to the greedy selection method. , that is, retain the corresponding computing capacity data set of features in different tasks Search the search space to find the feature data of the task to be processed Computing capacity data corresponding to the characteristics of different tasks ;

[0076] S226. Data set of computing capacity corresponding to the characteristics of scout bees in different tasks Random search in the search space and feature data of the task to be processed Computing capacity data corresponding to the characteristics of different tasks If the location of the nectar source After t iterations, the search reaches the iteration threshold No better one found , which is related to the feature data of the task to be processed Computing capacity data corresponding to the characteristics of different tasks Honey source location The hired bees will be abandoned, and the corresponding hired bees will become scout bees; the characteristics of scout bees in different tasks correspond to the computing capacity data set Random generation of search space and feature data of tasks to be processed Computing capacity data corresponding to the characteristics of different tasks The new nectar source location, the new location calculation formula is as follows: ,in Indicated in After iterations, a new Computing capacity data set corresponding to the characteristics of different tasks The location of the search space, Indicated in After iterations, a new Computing capacity data set corresponding to the characteristics of different tasks The location of the search space;

[0077] S227, when the maximum number of iterations T is met, the feature data of the task to be processed is output Computing capacity data corresponding to the characteristics of different tasks And generate the computing capacity data of the task to be processed through data identification , The unit is operations per second.

[0078] Through the pending task feature information collection unit, the data input dialog box is used to conveniently obtain the pending task feature information online, providing real data support for the subsequent accurate assessment of the computing capacity requirements of the pending tasks; the computing capacity storage unit corresponding to different task features and the pending task computing capacity identification unit cooperate with each other, and the computing capacity parameters corresponding to different task features are scientifically preset. Combined with the intelligent identification algorithm and the pending task feature information, a scientific analysis of the computing capacity required for the pending tasks is performed, so as to achieve accurate assessment of the target computing capacity in the computing power scheduling process and realize intelligent evaluation of the computing power of the computing power scheduling system scheduling operations.

[0079] For further information, see Figure 1 - Figure 2 , according to the computing capacity data of the tasks to be processed and the cloud computing platform object data corresponding to different computing capacity, the cloud computing platform object type required for the tasks to be processed is analyzed and processed, and the operation steps for constructing the cloud computing platform object data of the tasks to be processed are as follows:

[0080] S31. Establishing cloud computing platform object data sets corresponding to different computing capacity , ;in Indicates The different computing capacities corresponding to the computing capacity correspond to the cloud computing platform object data. Indicates the maximum number of computing capacity types. The cloud computing platform object data corresponding to different computing capacity indicates the cloud computing platform object data corresponding to different computing capacity required for data processing tasks. The cloud computing platform object types include Baidu Cloud, Alibaba Cloud, Huawei Cloud, and Tencent Cloud.

[0081] S32, use the breadth-first algorithm to convert the computing capacity data of the task to be processed Cloud computing platform object data set corresponding to different computing capacity The different computing capacity corresponds to the cloud computing platform object data According to the comparison of computing power capacity values, search for the computing power capacity data of the tasks to be processed The corresponding different computing capacity corresponds to the cloud computing platform object data And through data identification, a data set of cloud computing platform objects to be processed is constructed , ,in Indicates The cloud computing platform object data of the tasks to be processed. Indicates The cloud computing platform object data of the tasks to be processed.

[0082] Based on the object data of the cloud computing platform for the task to be processed and the service price data of different cloud computing platform objects, the service price search and processing of the cloud computing platform for the task to be processed is performed. The operation steps for generating the service price data of the cloud computing platform for the task to be processed are as follows:

[0083] S41. Establish a data set of service prices for different cloud computing platforms ,in Indicates the cloud computing platform object data corresponding to different computing capacity The corresponding service price data of different cloud computing platforms represent the service price data of different cloud computing platforms for data processing tasks. The unit is yuan;

[0084] S42, using a bidirectional search algorithm to collect the object data of the cloud computing platform to be processed Object data of the cloud computing platform for pending tasks to A collection of service price data from different cloud computing platforms Price data of object services of different cloud computing platforms According to the cloud computing platform object character matching, search for the cloud computing platform object data of the task to be processed to Corresponding service price data for different cloud computing platforms And generate the service price data set of the cloud computing platform object to be processed through data identification ,in Represents the cloud computing platform object data of the task to be processed The corresponding service price data of the cloud computing platform object for the task to be processed, Represents the cloud computing platform object data of the task to be processed The corresponding service price data of the cloud computing platform object for the task to be processed, and The unit is Yuan.

[0085] According to the object data of the cloud computing platform for pending tasks and the service price data of the cloud computing platform for pending tasks, the best cloud computing platform for pending tasks is analyzed and processed, and the operation steps for constructing the object data of the best cloud computing platform for pending tasks are as follows:

[0086] S51. Price data set of cloud computing platform objects for processing tasks The service price value of the pending task cloud computing platform object service price data is compared, and the service price data of the pending task cloud computing platform object with the smallest value is analyzed and marked as the minimum service price data of the pending task cloud computing platform object through data. , The unit mean of ;

[0087] S52: The minimum service price data of the cloud computing platform object of the task to be processed According to the cloud computing platform object character matching and the pending task cloud computing platform object data set The cloud computing platform object data of pending tasks is used to match the best cloud computing platform object for pending tasks, and the minimum service price data of the cloud computing platform object for pending tasks is searched out. The corresponding cloud computing platform object data of the pending task is constructed through data identification to obtain the optimal cloud computing platform object data of the pending task. .

[0088] Through the cooperation between the cloud computing platform object storage units corresponding to different computing power capacities and the cloud computing platform object analysis units for pending tasks, the standard storage of cloud platform objects corresponding to different computing power capacities is combined with the intelligent search algorithm and the computing power capacity parameters of the pending tasks to efficiently analyze and process the target cloud computing platform type, so as to realize accurate search of target cloud computing platform objects based on multi-cloud heterogeneity; the service price storage units of different cloud computing platform objects and the service price search units of cloud computing platform objects for pending tasks cooperate with each other, and the best cloud computing platform object identification units for pending tasks cooperate with each other to perform digital analysis on the service prices of multiple target cloud computing platform objects, and scientifically analyze the cloud computing platform with the lowest service price, so as to realize accurate screening of the best cloud computing platform objects that meet the pending tasks, improve the accuracy and reliability of the scheduling operations of the computing power scheduling system, and improve the scheduling operation quality of the computing power scheduling system.

[0089] For further information, see Figure 1 - Figure 2 , according to the object data of the best cloud computing platform for the task to be processed and the running instruction data of the cloud computing platform object, the running instruction search and processing of the cloud computing platform for the task to be processed is performed, and the operation steps for generating the running instruction data of the best cloud computing platform object are as follows:

[0090] S61. Establishing a cloud computing platform object operation instruction data set ,in Indicates the cloud computing platform object data corresponding to different computing capacity The corresponding cloud computing platform object operation instruction data, which represents the program instruction data used by the computing power dispatching center to control the specific cloud computing platform to perform data processing tasks;

[0091] S62, using a depth-constrained algorithm to optimize the cloud computing platform object data for the task to be processed A collection of command data that runs with cloud computing platform objects Zhongyun computing platform object running instruction data According to the cloud computing platform object character matching, search for the best cloud computing platform object data for the task to be processed The corresponding cloud computing platform object running instruction data And generate the best cloud computing platform object running instruction data through data identification .

[0092] The steps for executing the computing power scheduling job for the pending tasks based on the pending task feature data and the optimal cloud computing power platform object operation instruction data are as follows:

[0093] S71, the computing power dispatch center runs the instruction data according to the best cloud computing power platform object Control the best cloud computing platform object data for pending tasks The corresponding cloud computing platform runs and executes the feature data of the task to be processed The computing power scheduling job of the pending tasks.

[0094] Through the cooperation between the cloud computing platform object running instruction storage unit and the cloud computing platform object running instruction search unit, the cloud computing platform object running instruction parameters are scientifically preset and combined with the intelligent search algorithm and the optimal cloud computing platform object parameters of the task to be processed to autonomously retrieve the running instructions of the target optimal cloud computing platform, so as to achieve efficient matching of the target optimal cloud computing platform running instructions; the task computing power scheduling job execution unit autonomously and dynamically executes the computing power scheduling job of the task to be processed according to the characteristic parameters of the task to be processed and the target optimal cloud computing power platform object running instructions, thereby improving the execution accuracy and efficiency of the computing power scheduling job and improving the scientific nature of the computing power scheduling.

[0095] Embodiment 2:

[0096] See also Figure 1 - Figure 2 , a multi-cloud heterogeneous computing power scheduling system, a multi-cloud heterogeneous computing power scheduling method for implementing, the system includes a computing power capacity analysis module, a cloud computing power platform object evaluation module, and a computing power scheduling execution module;

[0097] The computing capacity analysis module includes a task feature information collection unit, a computing capacity storage unit corresponding to different task features, and a task computing capacity identification unit.

[0098] The feature information collection unit of the task to be processed collects the feature data of the task to be processed through the data input dialog box; the computing capacity storage unit corresponding to the features of different tasks is used to store the computing capacity data corresponding to the features of different tasks; the computing capacity identification unit of the task to be processed performs the computing capacity identification processing required by the task to be processed based on the feature data of the task to be processed and the computing capacity data corresponding to the features of different tasks, and generates the computing capacity data of the task to be processed;

[0099] The cloud computing platform object evaluation module includes a cloud computing platform object storage unit corresponding to different computing power capacities, a cloud computing platform object analysis unit for pending tasks, a storage unit for different cloud computing platform object service prices, a cloud computing platform object service price search unit for pending tasks, and an optimal cloud computing platform object identification unit for pending tasks;

[0100] A cloud computing platform object storage unit corresponding to different computing power capacities stores cloud computing platform object data corresponding to different computing power capacities based on multi-cloud heterogeneity; a cloud computing platform object analysis unit for pending tasks analyzes and processes the cloud computing platform object types required for pending tasks based on the computing power capacity data of pending tasks and the cloud computing platform object data corresponding to different computing power capacities, and constructs the cloud computing platform object data for pending tasks; a different cloud computing platform object service price storage unit is used to store different cloud computing platform object service price data; a cloud computing platform object service price search unit for pending tasks searches and processes the service price of the cloud computing platform for pending tasks based on the cloud computing platform object data for pending tasks and the service price data of different cloud computing platform objects, and generates the service price data of the cloud computing platform object for pending tasks; an optimal cloud computing platform object identification unit for pending tasks analyzes and processes the optimal cloud computing platform object for pending tasks based on the cloud computing platform object data for pending tasks and the service price data of the cloud computing platform object for pending tasks, and constructs the optimal cloud computing platform object data for pending tasks;

[0101] The computing power scheduling execution module includes a cloud computing power platform object running instruction storage unit, a cloud computing power platform object running instruction search unit, and a pending task computing power scheduling job execution unit;

[0102] The cloud computing power platform object operation instruction storage unit is used to store the cloud computing power platform object operation instruction data; the cloud computing power platform object operation instruction search unit searches for the operation instructions of the cloud computing power platform for the pending task according to the optimal cloud computing power platform object data of the pending task and the cloud computing power platform object operation instruction data, and generates the optimal cloud computing power platform object operation instruction data; the pending task computing power scheduling job execution unit executes the pending task computing power scheduling job according to the pending task feature data and the optimal cloud computing power platform object operation instruction data.

[0103] Although embodiments of the present invention have been shown and described, it will be appreciated by those skilled in the art that various changes, modifications, substitutions and variations may be made to the embodiments without departing from the principles and spirit of the present invention, and that the scope of the present invention is defined by the appended claims and their equivalents.

Claims

1. A multi-cloud heterogeneous computing power scheduling method, characterized in that: The method comprises the following steps: S1, collecting feature data of the task to be processed, said S1 comprises the following steps: S11. Collect the characteristic information of the data task to be processed online through the data input dialog box and generate the characteristic data of the task to be processed , the characteristic data of the task to be processed includes data type, data capacity, data task processing field and preset processing time length; S2, performing identification processing on the computing capacity required for the task to be processed based on the feature data of the task to be processed and the computing capacity data corresponding to the features of different tasks, and generating computing capacity data for the task to be processed, wherein S2 comprises the following steps: S21. Establish a data set of computing capacity corresponding to the characteristics of different tasks , ;in Indicates The characteristics of different tasks correspond to the computing capacity data. Indicates the maximum number of different task feature types, The unit of is operations per second; S22: The task feature data to be processed Computing capacity data set corresponding to the characteristics of the different tasks The characteristics of different tasks described in the table correspond to the computing capacity data Perform task feature data matching to search for the task feature data to be processed The corresponding computing capacity data of the characteristics of the different tasks And generate the computing capacity data of the task to be processed through data identification , execute and generate the computing capacity data of the task to be processed The specific steps are as follows: S221, Initialize each honey source , that is, the characteristics of different tasks correspond to the computing capacity data set The characteristics of different tasks correspond to the computing capacity data Set as honey source ; Set the number of nectar sources , current number of iterations t and maximum number of iterations T; honey source Initial position ; S222. Computing capacity data set corresponding to the characteristics of different tasks The search space contains the computing capacity data corresponding to the features of different tasks. Honey source location Assign hired bees. At the beginning of the search, hired bees are at the location of the nectar source. Search around to generate new honey source locations , that is, the computing capacity data set corresponding to the characteristics of different tasks Search the search space for the feature data of the task to be processed Computing capacity data corresponding to the characteristics of different tasks ; S223, based on the location of the new nectar source The fitness of the nectar source to be retained is determined by the greedy selection method. , that is, retain the corresponding computing capacity data set of features in different tasks Search the search space for the feature data of the task to be processed Computing capacity data corresponding to the characteristics of different tasks with the largest fitness values ; S224, employed bees share nectar information, and follower bees correspond to the computing capacity data set according to the probability values ​​of different tasks. Search the search space for the feature data of the task to be processed Computing capacity data corresponding to the characteristics of different tasks Honey Source , the follower bee corresponds to the computing capacity data set of the characteristics of different tasks according to the nectar source information shared by the employed bees The search space searches for characteristic data related to the task to be processed. Computing capacity data corresponding to the characteristics of different tasks ; S225, follower bees search in the same way as employed bees, and determine the nectar source to be retained based on the greedy selection method. , that is, retain the corresponding computing capacity data set of features in different tasks Search the search space for the feature data of the task to be processed Computing capacity data corresponding to the characteristics of different tasks ; S226. Data set of computing capacity corresponding to the characteristics of scout bees in different tasks Random search in the search space and the feature data of the task to be processed Computing capacity data corresponding to the characteristics of different tasks If the location of the nectar source After t iterations, the search reaches the iteration threshold No better honey source was found , which is related to the feature data of the task to be processed Computing capacity data corresponding to the characteristics of different tasks Honey source location The bees will be abandoned, and the corresponding hired bees will become scout bees; the characteristics of scout bees in different tasks correspond to the computing capacity data set The search space randomly generates the feature data of the task to be processed Computing capacity data corresponding to the characteristics of different tasks New nectar source location; S227, when the maximum number of iterations T is met, output the feature data of the task to be processed Computing capacity data corresponding to the characteristics of different tasks And generate the computing capacity data of the task to be processed through data identification , The unit of is operations per second; S3: Analyze and process the cloud computing platform object type required for the task to be processed according to the computing capacity data of the task to be processed and the cloud computing platform object data corresponding to different computing capacity, and construct the cloud computing platform object data of the task to be processed. S3 includes the following steps: S31. Establishing cloud computing platform object data sets corresponding to different computing capacity , ;in Indicates The different computing capacities corresponding to the computing capacity correspond to the cloud computing platform object data. Indicates the maximum number of computing capacity types. Cloud computing platform object types include Baidu Cloud, Alibaba Cloud, Huawei Cloud, and Tencent Cloud. S32, using a breadth-first algorithm to convert the computing capacity data of the task to be processed Cloud computing platform object data set corresponding to the different computing capacity The different computing capacity described in the above corresponds to the cloud computing platform object data According to the comparison of computing power capacity values, search for the computing power capacity data of the task to be processed The corresponding different computing capacity corresponds to the cloud computing platform object data And through data identification, a data set of cloud computing platform objects to be processed is constructed , ,in Indicates The cloud computing platform object data of the tasks to be processed. Indicates The cloud computing platform object data of the tasks to be processed; S4: Based on the object data of the cloud computing platform for the task to be processed and the service price data of different cloud computing platform objects, a service price search process of the cloud computing platform for the task to be processed is performed to generate the service price data of the cloud computing platform for the task to be processed, wherein S4 comprises the following steps: S41. Establish a data set of service prices for different cloud computing platforms ,in Indicates the cloud computing platform object data corresponding to different computing capacity Corresponding service price data for different cloud computing platforms. The unit is yuan; S42, using a bidirectional search algorithm to collect the object data of the task cloud computing platform to be processed The cloud computing platform object data of the pending tasks described in to A collection of service price data for different cloud computing platforms The price data of the service objects of different cloud computing platforms described in According to the cloud computing platform object character matching, search for the cloud computing platform object data of the task to be processed to The corresponding service price data of different cloud computing platforms And generate the service price data set of the cloud computing platform object to be processed through data identification ,in Represents the cloud computing platform object data of the task to be processed The corresponding service price data of the cloud computing platform object for the task to be processed, Represents the cloud computing platform object data of the task to be processed The corresponding service price data of the cloud computing platform object for the task to be processed, and The unit of all is Yuan; S5: Analyze and process the best cloud computing platform object for the task to be processed based on the cloud computing platform object data for the task to be processed and the service price data of the cloud computing platform object for the task to be processed, and construct the best cloud computing platform object data for the task to be processed. S5 includes the following steps: S51: collect the service price data of the cloud computing platform object for the task to be processed The service price data of the cloud computing platform object to be processed are compared with the service price values, and the service price data of the cloud computing platform object to be processed with the smallest value is analyzed and marked as the minimum service price data of the cloud computing platform object to be processed. , The unit mean of ; S52: The minimum service price data of the cloud computing platform object of the task to be processed According to the cloud computing platform object character matching and the cloud computing platform object data set of the task to be processed The cloud computing platform object data of the pending task is matched with the best cloud computing platform object of the pending task, and the minimum service price data of the cloud computing platform object of the pending task is searched out. The corresponding cloud computing platform object data of the task to be processed is constructed through data identification to obtain the optimal cloud computing platform object data of the task to be processed. ; S6: Performing a search and processing of the operation instruction of the cloud computing platform for the task to be processed according to the object data of the optimal cloud computing platform for the task to be processed and the operation instruction data of the cloud computing platform object, and generating the operation instruction data of the optimal cloud computing platform object, wherein S6 comprises the following steps: S61. Establishing a cloud computing platform object operation instruction data set ,in Indicates the cloud computing platform object data corresponding to different computing capacity The corresponding cloud computing platform object running instruction data; S62, using a depth-restricted algorithm to convert the optimal cloud computing platform object data of the task to be processed A set of command data for running the cloud computing platform object The cloud computing platform object running instruction data described in According to the cloud computing platform object character matching, search for the best cloud computing platform object data for the task to be processed The corresponding cloud computing platform object running instruction data And generate the best cloud computing platform object running instruction data through data identification ; S7: Execute the computing power scheduling operation of the task to be processed according to the feature data of the task to be processed and the object operation instruction data of the optimal cloud computing power platform, wherein S7 includes the following steps: S71, the computing power dispatching center runs the instruction data according to the optimal cloud computing power platform object Control the optimal cloud computing platform object data for the task to be processed The corresponding cloud computing platform runs and executes the characteristic data of the task to be processed The computing power scheduling job of the pending tasks.

2. A multi-cloud heterogeneous computing power scheduling system, used to implement a multi-cloud heterogeneous computing power scheduling method as described in claim 1, characterized in that: The system includes a computing power capacity analysis module, a cloud computing power platform object evaluation module, and a computing power scheduling execution module; The computing capacity analysis module includes a task feature information collection unit, a computing capacity storage unit corresponding to different task features, and a task computing capacity identification unit. The task feature information collection unit to be processed collects the task feature data to be processed through a data input dialog box; the different task feature corresponding computing power capacity storage unit is used to store the feature corresponding computing power capacity data of different tasks; the task feature capacity identification unit to be processed performs the required computing power capacity identification processing for the task to be processed based on the task feature data to be processed and the feature corresponding computing power capacity data of different tasks, and generates the computing power capacity data for the task to be processed; The cloud computing platform object evaluation module includes a cloud computing platform object storage unit corresponding to different computing power capacities, a cloud computing platform object analysis unit for pending tasks, a storage unit for different cloud computing platform object service prices, a cloud computing platform object service price search unit for pending tasks, and an optimal cloud computing platform object identification unit for pending tasks; The cloud computing platform object storage unit corresponding to different computing power capacities stores cloud computing platform object data corresponding to different computing power capacities based on multi-cloud heterogeneity; the cloud computing platform object analysis unit for pending tasks performs analysis and processing on the cloud computing platform object type required for pending tasks according to the computing power capacity data of pending tasks and the cloud computing platform object data corresponding to different computing power capacities, and constructs the cloud computing platform object data for pending tasks; the different cloud computing platform object service price storage unit is used to store different cloud computing platform object service price data; the cloud computing platform object service price search unit for pending tasks performs service price search processing on the cloud computing platform for pending tasks based on the cloud computing platform object data for pending tasks and the service price data of different cloud computing platform objects, and generates service price data of the cloud computing platform object for pending tasks; the optimal cloud computing platform object identification unit for pending tasks performs analysis and processing on the optimal cloud computing platform object for pending tasks according to the cloud computing platform object data for pending tasks and the service price data of the cloud computing platform object for pending tasks, and constructs the optimal cloud computing platform object data for pending tasks; The computing power scheduling execution module includes a cloud computing power platform object running instruction storage unit, a cloud computing power platform object running instruction search unit, and a pending task computing power scheduling job execution unit; The cloud computing power platform object running instruction storage unit is used to store cloud computing power platform object running instruction data; the cloud computing power platform object running instruction search unit searches for the running instructions of the cloud computing power platform for the task to be processed based on the optimal cloud computing power platform object data for the task to be processed and the cloud computing power platform object running instruction data, and generates the optimal cloud computing power platform object running instruction data; the computing power scheduling job execution unit for the task to be processed executes the computing power scheduling job for the task to be processed based on the characteristic data of the task to be processed and the optimal cloud computing power platform object running instruction data.

Images