Computing power network resource allocation method and device and nonvolatile storage medium
By mapping the computing, network and storage resources in the computing power network to the same dimension and combining the demand strategy, the problem of uneven resource allocation is solved, and the uniform allocation of computing power resources and the improvement of resource utilization is achieved.
Patent Information
- Application Number
- CN202411066469.5
- Authority / Receiving Office
- CN · China
- Patent Type
- Applications(China)
- Current Assignee / Owner
- Filing Date
- 2024-08-05
- Publication Date
- 2024-11-19
- Estimated Expiration
- 2044-08-05
AI Technical Summary
In computing power network, the problem of uneven resource allocation has not been effectively solved because computing resources, storage resources and network resources are allocated separately.
By obtaining the total amount of computing resources, network resources and storage resources in the computing power network, and mapping them to the same resource dimension, combining the resource demand strategies of multiple computing power demand sides, computing power resources are allocated to achieve resource Distribute evenly.
It realizes a more even distribution of computing power resources, solves the problem of uneven resource allocation, and improves resource utilization and maximizes the benefits of multiple computing power demand sides.
Smart Images

Figure CN118984271A_ABST
Abstract
Description
Technical Field
[0001] The present invention relates to the field of computing power resource allocation, and in particular, to a computing power network resource allocation method, device and non-volatile storage medium. Background Art
[0002] In the process of allocating computing power network resources, since relevant technologies allocate resources separately based on the computing resources, storage resources and network resources existing in the computing power network, without taking into account that the computing resources, storage resources and network resources in the computing power network are in different dimensions, there is an uneven distribution of computing power in the process of allocating computing power resources in the computing power network.
[0003] To address the above-mentioned problems, no effective solution has been proposed yet. Summary of the invention
[0004] Embodiments of the present invention provide a computing power network resource allocation method, device and non-volatile storage medium to at least solve the technical problem of uneven computing power resource allocation caused by the related technology of allocating resources separately according to computing resources, storage resources and network resources in the computing power network from the perspective of computing power demand.
[0005] According to one aspect of an embodiment of the present invention, a method for allocating computing power network resources is provided, comprising: obtaining a total amount of computing resources, a total amount of network resources, and a total amount of storage resources to be allocated in a computing power network, wherein the total amount of computing resources is used to indicate the total amount of resources used to perform computing tasks in the computing power network, the network resources are used to indicate the total amount of resources used for network transmission in the computing power network, and the storage resources are used to indicate the total amount of resources used to perform storage tasks in the computing power network; mapping the total amount of computing resources, the total amount of network resources, and the total amount of storage resources to the same resource dimension to obtain a resource quantification result, wherein the resource quantification result is used to indicate the total amount of resources after the total amount of computing resources, the total amount of network resources, and the total amount of storage resources are mapped to the same dimension; obtaining computing power demand strategies corresponding to multiple computing power demand ends, wherein the computing power demand strategies are used to indicate the computing power resource demands corresponding to the corresponding computing power demand ends in multiple different time periods; based on the resource quantification result and the computing power demand strategies corresponding to the multiple computing power demand ends, computing power resources are allocated to the computing power demand ends to obtain computing power resource allocation results corresponding to the multiple computing power demand ends.
[0006] According to another aspect of an embodiment of the present invention, a computing power network resource allocation device is also provided, including: a total resource acquisition module, used to acquire the total amount of computing resources, the total amount of network resources and the total amount of storage resources to be allocated in the computing power network, wherein the total amount of computing resources is used to indicate the total amount of resources used to perform computing tasks in the computing power network, the network resources are used to indicate the total amount of resources used for network transmission in the computing power network, and the storage resources are used to indicate the total amount of resources used to perform storage tasks in the computing power network; a dimension mapping module is used to map the total amount of computing resources, the total amount of network resources and the total amount of storage resources to the same resource dimension to obtain Resource quantification results, wherein the resource quantification results are used to indicate the total amount of resources after mapping the total amount of computing resources, the total amount of network resources and the total amount of storage resources to the same dimension; a demand strategy acquisition module, used to obtain the computing power demand strategies corresponding to multiple computing power demand ends, wherein the computing power demand strategies are used to indicate the computing power resource demands corresponding to the corresponding computing power demand ends in multiple different time periods; a resource allocation module, used to allocate computing power resources to the computing power demand ends based on the resource quantification results and the computing power demand strategies corresponding to the multiple computing power demand ends, and obtain the computing power resource allocation results corresponding to the multiple computing power demand ends.
[0007] According to another aspect of an embodiment of the present invention, a non-volatile storage medium is provided, wherein the non-volatile storage medium stores a plurality of instructions, wherein the instructions are suitable for being loaded by a processor and executing any one of the computing power network resource allocation methods.
[0008] According to another aspect of an embodiment of the present invention, a computer program product is also provided, including a computer program, which implements the steps of any one of the computing power network resource allocation methods when executed by a processor.
[0009] In an embodiment of the present invention, the total amount of computing resources, the total amount of network resources and the total amount of storage resources to be allocated in the computing power network are obtained, wherein the total amount of computing resources is used to indicate the total amount of resources used to perform computing tasks in the computing power network, the network resources are used to indicate the total amount of resources used for network transmission in the computing power network, and the storage resources are used to indicate the total amount of resources used to perform storage tasks in the computing power network; the total amount of computing resources, the total amount of network resources and the total amount of storage resources are mapped to the same resource dimension to obtain a resource quantification result, wherein the resource quantification result is used to indicate the total amount of resources after the total amount of computing resources, the total amount of network resources and the total amount of storage resources are mapped to the same dimension; computing power demand strategies corresponding to multiple computing power demand ends are obtained, wherein the computing power demand strategies are used to indicate the computing power resource demands corresponding to the corresponding computing power demand ends in multiple different time periods; based on the resource quantification result and the computing power demand strategies corresponding to the multiple computing power demand ends, computing power resources are allocated to the computing power demand ends to obtain computing power resource allocation results corresponding to the multiple computing power demand ends. The purpose of mapping the total amount of computing resources to be allocated, the total amount of network resources and the total amount of storage resources in the computing power network to the same dimension, comprehensively considering the total amount of resources in the computing power network, and allocating computing power network resources from the two aspects of resource demand and supply is achieved, thereby effectively achieving the technical effect of more uniform distribution of computing power resources in the network, and further solving the technical problem of uneven distribution of computing power resources caused by the fact that in related technologies, computing resources, storage resources and network resources in the computing power network are allocated separately from the perspective of computing power demand. BRIEF DESCRIPTION OF THE DRAWINGS
[0010] The drawings described herein are used to provide a further understanding of the present invention and constitute a part of this application. The exemplary embodiments of the present invention and their descriptions are used to explain the present invention and do not constitute an improper limitation of the present invention. In the drawings:
[0011] Figure 1 is a flow chart of a computing power network resource allocation method according to an embodiment of the present invention;
[0012] Figure 2 is a flow chart of an optional computing power network resource allocation method according to an embodiment of the present invention;
[0013] Figure 3 It is a schematic diagram of the deployment of a computing power network resource allocation method based on game theory according to an optional embodiment of the present invention;
[0014] Figure 4 It is a schematic diagram of a computing power network resource allocation device according to an embodiment of the present invention. DETAILED DESCRIPTION
[0015] In order to enable those skilled in the art to better understand the scheme of the present invention, the technical scheme in the embodiments of the present invention will be clearly and completely described below in conjunction with the drawings 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 should fall within the scope of protection of the present invention.
[0016] It should be noted that the terms "first", "second", etc. in the specification and claims of the present invention and the above-mentioned drawings are used to distinguish similar objects, and are not necessarily used to describe a specific order or sequence. It should be understood that the data used in this way can be interchanged where appropriate, so that the embodiments of the present invention described herein can be implemented in an order other than those illustrated or described herein. In addition, the terms "including" and "having" and any variations thereof are intended to cover non-exclusive inclusions, for example, a process, method, system, product or device that includes a series of steps or units is not necessarily limited to those steps or units that are clearly listed, but may include other steps or units that are not clearly listed or inherent to these processes, methods, products or devices.
[0017] First, to facilitate understanding of the embodiments of the present invention, some terms or nouns involved in the present invention are explained below:
[0018] Heterogeneity refers to the fact that different nodes (different users or different times) have different computing capabilities, storage capabilities, and network bandwidths during the resource allocation process of the computing network. The existence of heterogeneity may cause nodes in the computing network to have different efficiencies and performance when processing tasks.
[0019] Game theory refers to the mathematical theory that studies decision-making and behavior patterns in conflict or cooperative environments. Game theory focuses on the problem of how to formulate optimal strategies for various parties under circumstances of information asymmetry, conflict of interest, cooperation and competition.
[0020] A computing network is a network that is composed of multiple computing networks connected together to complete computing tasks. A computing network is usually composed of a large number of computing resources and can be used to perform various complex computing tasks.
[0021] Resource dimensions refer to the types and quantities of computing resources available in the computing network. These computing network resources can include computer processing power, storage capacity, bandwidth, etc. In the computing network, different computer processing power, storage capacity, and bandwidth may have different resource dimensions. Demanders can choose the appropriate resource dimension to perform tasks according to their needs.
[0022] According to an embodiment of the present invention, a method embodiment of computing power network resource allocation is provided. It should be noted that the steps shown in the flowchart of the accompanying drawings can be executed in a computer system such as a set of computer executable instructions, and although a logical order is shown in the flowchart, in some cases, the steps shown or described can be executed in an order different from that shown here.
[0023] Figure 1 is a flow chart of a computing power network resource allocation method according to an embodiment of the present invention. Figure 1 As shown, the method comprises the following steps:
[0024] Step S102, obtaining the total amount of computing resources, network resources, and storage resources to be allocated in the computing network, wherein the total amount of computing resources is used to indicate the total amount of resources used to perform computing tasks in the computing network, the network resources are used to indicate the total amount of resources used for network transmission in the computing network, and the storage resources are used to indicate the total amount of resources used to perform storage tasks in the computing network;
[0025] Optionally, during the computing network resource allocation process, computing resources, storage resources, and network resources are in different dimensions, which can easily lead to uneven resource allocation in the computing network. By modeling based on the three dimensions corresponding to computing resources, network resources, and storage resources in the computing network, the three resources are mapped to the same resource dimension to achieve a more even distribution of computing network resources. Specifically, the computing resources, network resources, and storage resources in the computing network are obtained separately. First, all computing resources in the computing network are mapped to the same dimension to form the total amount of computing resources; all network resources are mapped to the same dimension to form the total amount of network resources; all storage resources are mapped to the same dimension to form the total amount of storage resources, thereby achieving a preliminary unification of resource dimensions.
[0026] In an optional embodiment, the total amount of computing resources, the total amount of network resources and the total amount of storage resources to be allocated in the computing power network are obtained, including: obtaining multiple initial computing resources, multiple initial network resources and multiple initial storage resources to be allocated in the computing power network, wherein the multiple initial computing resources include at least two of the following: logical computing resources, parallel computing resources, and neural network computing resources; the multiple initial network resources include at least two of the following: network bandwidth resources, latency resources, and packet loss rate resources; the multiple initial storage resources include at least two of the following: read-only storage resources and random storage resources; mapping the multiple initial computing resources to the same resource dimension to obtain the total amount of computing resources; mapping the multiple initial network resources to the same resource dimension to obtain the total amount of network resources; mapping the multiple initial storage resources to the same resource dimension to obtain the total amount of storage resources.
[0027] Optionally, computing resources can be mainly divided into three categories: general logic computing resources, parallel computing resources, and neural network computing resources. Different types of computing resources are preliminarily mapped to the same resource dimension based on the measurement function, and the computing power of the computing network is measured by calculating the weighted average value as the total amount of computing resources to eliminate interference in the allocation of computing resources due to different resource dimensions.
[0028] Specifically, assume that a computing resource includes n-dimensional logic computing resources, m-dimensional parallel computing resources, and p-dimensional neural network computing resources. Based on the computing resource modeling, a computing capacity quantification value is obtained as the total amount of computing resources. The computing capacity quantification value can be obtained in the following way:
[0029]
[0030] Where C represents the quantified value of the computing power of the computing network, and f1(x) represents the first mapping function, which is uniformly set by the computing network administrator. Specifically, f1(a i ) represents the first mapping function related to the logical computing resources, f1(b j ) represents the first mapping function related to parallel computing resources, f1(c k ) represents the first mapping function associated with the neural network computing resources.
[0031] Optionally, network resources can be mainly divided into four categories: network bandwidth, end-to-end delay, delay jitter, and packet loss rate. Among them, network bandwidth is used to indicate the amount of data that can be sent and received per unit time; end-to-end delay is used to indicate the total delay of user service response, delay jitter is used to indicate the robustness of user service network services, and delay jitter can be ignored in the process of network resource quantification. Transmission packet loss rate is used to indicate the reliability of network services. Similar to computing power resource modeling, the above three types of network resources (without considering delay jitter) are mapped to a unified dimension through a measurement function to obtain a network capacity quantification value as the total amount of network resources, where the network capacity quantification value can be obtained in the following way:
[0032] N=f2(Bandwidth)+f2(Delay)+f2(PLR)
[0033] Among them, N represents the quantified value of the network capacity of the computing power network, f2(x) represents the second mapping function, Bandwidth represents the bandwidth variable, Delay represents the end-to-end delay variable, and PLR represents the packet loss rate variable.
[0034] Optionally, storage resources can be divided into two categories: Read-Only Memory (ROM) and Random Access Memory (RAM). ROM is used to store fixed system data and applications, and RAM is used as a data cache area currently used by the operating system and applications, with strong read and write capabilities. The above two types of storage resources are mapped to a unified dimension through a measurement function to obtain a storage capacity quantification value as the total amount of network resources. The storage capacity quantification value is calculated as follows:
[0035] S=f3(ROM)+f3(RAM)
[0036] Among them, S represents the quantitative value of the storage capacity of the computing power network, f3(x) represents the third mapping function, ROM is the read-only memory capacity variable, and RAM is the random access memory capacity variable.
[0037] Step S104, mapping the total amount of computing resources, the total amount of network resources, and the total amount of storage resources to the same resource dimension to obtain a resource quantification result, wherein the resource quantification result is used to indicate the total amount of resources after the total amount of computing resources, the total amount of network resources, and the total amount of storage resources are mapped to the same dimension;
[0038] Optionally, after completing the quantification of the above three resources (computing resources, network resources, and storage resources), a preliminary unification of resource dimensions is formed. Subsequently, weighted calculation can be performed on the quantified values of the above three resources to achieve unified quantification of all resources in the computing power network and obtain resource quantification results. The resource quantification results are calculated as follows:
[0039] R=αC+βN+γS
[0040] Among them, C is the quantified value of computing power, N is the quantified value of network capacity, S is the quantified value of storage capacity, R is the resource quantification result after unified quantification of all resources in the computing power network, α is the first proportional coefficient, β is the second proportional coefficient, and γ is the third proportional coefficient. The specific first proportional coefficient, second proportional coefficient and third proportional coefficient are all configured by the computing power network administrator.
[0041] Step S106, obtaining computing power demand strategies corresponding to the plurality of computing power demand terminals, wherein the computing power demand strategies are used to indicate computing power resource demands corresponding to the corresponding computing power demand terminals in a plurality of different time periods;
[0042] Optionally, the computing power demand side is the computing power task, and the computing power resource demand required by the computing power task corresponding to the computing power demand side is the computing power resource demand required by the computing power demand side. Among them, when each computing power task is executed in different time periods, different computing power resource demands may appear. Specifically, affected by the performance of the computing power network, the total amount of computing power resources provided by the computing power network in each time period is relatively fixed. In different time periods, the computing power tasks of multiple computing power demand sides share the resources in the computing power network. For example, in the first preset time period, there are three computing power demand sides whose computing power tasks share the resources of the above-mentioned computing power network. At this time, the computing power network resources allocated to each computing power task are relatively sufficient. In the second preset time period, there are twelve computing power demand sides whose computing power tasks share the above-mentioned computing power network resources. At this time, the computing power network resources are relatively tight, and the computing power resource demand may be adjusted so that the computing power network resources are reasonably allocated to multiple computing power demand sides. In summary, in different time periods, based on the different numbers of computing power demand sides, different computing power demand strategies can be obtained, so that limited computing power resources can be relatively evenly allocated to different computing power demand sides.
[0043] In an optional embodiment, computing power demand strategies corresponding to multiple computing power demand ends are obtained, including: obtaining the computing power resource demand of any computing power demand end in any time period among the multiple computing power demand ends in the following manner: determining a target task to be performed by any computing power demand end in any time period; dividing the target task into multiple subtasks, and obtaining resource demands corresponding to the multiple subtasks; determining resource mapping functions corresponding to the multiple subtasks, wherein the resource mapping function is used to indicate a mapping relationship between the resource demand of the corresponding subtask and a predetermined resource dimension, and the predetermined resource dimension is a resource dimension that is the same as a resource quantification result; based on the resource mapping functions corresponding to the multiple subtasks, Function, and the resource requirements corresponding to multiple subtasks respectively, to obtain the resource mapping results corresponding to the multiple subtasks respectively; based on the resource mapping results corresponding to the multiple subtasks respectively, to obtain the computing power resource demand of any computing power demand end in any time period; adopt the method of obtaining the computing power resource demand of any computing power demand end in any time period to obtain the computing power resource demand corresponding to any computing power demand end in multiple time periods; based on the computing power resource demand corresponding to any computing power demand end in multiple time periods, to obtain the computing power demand strategy of any computing power demand end; adopt the method of obtaining the computing power demand strategy of any computing power demand end to obtain the computing power demand strategies corresponding to multiple computing power demand ends respectively.
[0044] Optionally, the target task executed by the computing power demand side in any time period is divided into multiple subtasks, wherein the multiple subtasks have execution progress priorities. Specifically, the priorities can be determined according to the order in which the subtasks need to be executed. Each subtask requires computing resources, network resources, and storage resources. The computing resources, network resources, and storage resources occupied by the multiple subtasks are mapped and quantified to form resource mapping results corresponding to the subtasks. Each subtask is mapped and quantified to obtain the corresponding resource mapping results. The resource mapping results of the multiple subtasks correspond to the resource mapping results corresponding to the target task of the computing power demand side in the above selected time period.
[0045] Optionally, taking N computing power demand ends and T time periods as an example, the computing power demand strategies corresponding to multiple computing power demand ends are obtained as follows: wherein, the set of multiple computing power demand ends is represented as N = {1, 2, ..., N}, and at the same time, each computing power demand end has significant differences in the demand for computing power in different time periods of a day. Therefore, a day is divided into T time periods, where T = {1, 2, ..., t}, which accurately expresses the demand for computing power network resources by the computing power demand end in different time periods.
[0046] Specifically, the computing power resource demand of each computing power demand terminal k∈N in the tth period is as follows:
[0047]
[0048] in, Indicates the computing power demand of the kth computing power demander in the tth period;
[0049] Specifically, the sum of the computing power demand of the kth computing power demand end in all time periods is as follows:
[0050]
[0051] in, The maximum computing power requirement preset for the kth computing power demander.
[0052] Specifically, the computing resource demand vector of each computing demand end k∈N is expressed as As shown below:
[0053]
[0054] in, Indicates the computing power demand strategy corresponding to the k-th computing power demand end.
[0055] In an optional embodiment, when resource requirements include computing resource requirements, network resource requirements and storage resource requirements, and resource mapping functions include computing resource mapping functions, network resource mapping functions and storage resource mapping functions, resource mapping results corresponding to multiple subtasks are obtained based on the resource mapping functions corresponding to multiple subtasks and the resource requirements corresponding to multiple subtasks, including: obtaining the resource mapping result of any subtask among the multiple subtasks in the following manner: obtaining the computing resource mapping result of any subtask based on the computing resource requirement of any subtask and the corresponding computing resource mapping function; obtaining the network resource mapping result of any subtask based on the network resource requirement of any subtask and the corresponding network resource function; obtaining the storage resource mapping result of any subtask based on the storage resource requirement of any subtask and the corresponding storage resource mapping function; performing weighted calculation on the computing resource mapping result, the network resource mapping result and the storage resource mapping result to obtain the resource mapping result of any subtask; and obtaining the resource mapping results corresponding to multiple subtasks by adopting the method of obtaining the resource mapping result of any subtask.
[0056] Optionally, taking the i-th subtask of the k-th computing power demand side as an example, the demand of the i-th subtask for computing power network resources is measured from the three dimensions of computing resources, network resources, and storage resources, and the three types of resources corresponding to the i-th subtask are normalized, so that not only the dimensions of the computing power network are unified, but also the dimensions of the computing power network required by the computing power demand side are unified, which is convenient for subsequent computing power network resource allocation. Specifically, the i-th subtask is normalized based on the three types of resources, and the calculation method of obtaining the resource mapping results corresponding to the i-th subtask is as follows:
[0057]
[0058] in, represents the resource mapping result corresponding to the i-th subtask of the k-th computing power demand side, f 4i (C) is the fourth mapping function of the computing resource requirements of the i-th subtask, f 4i (N) is the fourth mapping function of the network resource requirements of the i-th subtask, f 4i (S) is the fourth mapping function of the storage resource requirement of the i-th subtask.
[0059] Based on the resource mapping results corresponding to the above-mentioned i-th subtask, the resource mapping results required by the target task of the k-th computing power demand side in any period are calculated. The specific calculation is as follows:
[0060]
[0061] Among them, D represents the total resource mapping result of n subtasks on the kth computing power demand side in any period of time, and n is the number of subtasks.
[0062] Step S108, based on the resource quantification results and the computing power demand strategies corresponding to the multiple computing power demand ends, allocate computing power resources to the computing power demand ends, and obtain computing power resource allocation results corresponding to the multiple computing power demand ends.
[0063] Optionally, the resource quantification results in the computing power network can be allocated based on the computing power demand strategies corresponding to multiple computing power demand ends, and the computing power resource allocation results corresponding to the multiple computing power demand ends can be obtained. The corresponding tasks can be executed according to the computing power resource allocation results obtained by the multiple computing power demand ends at the current moment, so as to effectively improve the resource utilization rate in the computing power network, maximize the benefit values of multiple computing power demand ends in the computing power network, and ensure the stability of services of multiple computing power demand ends in the computing power network.
[0064] In an optional embodiment, based on the resource quantification results and the computing power demand strategies corresponding to the multiple computing power demand ends, computing power resources are allocated to the computing power demand ends to obtain the computing power resource allocation results corresponding to the multiple computing power demand ends, including: determining the initial resource allocation results corresponding to the multiple computing power demand ends; constructing utility functions corresponding to the multiple computing power demand ends, wherein the utility functions are used to indicate the relationship between the computing power resource demand and the computing power resource income of the corresponding computing power demand ends; based on the computing power demand strategies corresponding to the multiple computing power demand ends, and the utility functions corresponding to the multiple computing power demand ends, determining the utility values corresponding to the multiple computing power demand ends; based on the resource quantification results, taking the maximum sum of the utility values corresponding to the multiple computing power demand ends as the optimization goal, using a game theory algorithm to optimize the initial resource allocation results corresponding to the multiple computing power demand ends to obtain the computing power resource allocation results corresponding to the multiple computing power demand ends.
[0065] Optionally, the utility function is determined as The utility function reaches its maximum value, which is: When the goal is to maximize the computing power utility value, it is also necessary to comprehensively consider the probability change rate of the computing power demand side selecting strategy c, that is, to minimize the change in the computing power demand strategy, and the computing power demand side selecting strategy The specific calculation is as follows:
[0066]
[0067] Among them, x c It is represented by the preset variable value, which is set to 10 divided by the number of optional strategies. The excess return brought by strategy c is The calculation is as follows:
[0068]
[0069] in, is the average utility value corresponding to all strategies, The calculation method is as follows:
[0070]
[0071] Among them, m is the total number of all optional strategies in all computing power demand ends, p is the set of optional strategies of any computing power demand end, and N is the set of computing power demand ends.
[0072] Optionally, the goal of computing network resource allocation is to match appropriate resources to each computing demand side to maximize the benefits of all computing demand sides. The computing network resource allocation problem can be a process of multiple computing demand sides competing for resources. Considering that computing demand sides cannot obtain global information, evolutionary game theory is used to complete resource competition and achieve the optimal match between computing network resources and computing demand. Specifically, the game algorithm for computing network resource allocation can be expressed as G= <w,(p i ) i∈w ,(F i ) i∈w >, where w is the set of game participants, p i represents the set of optional strategies for the i-th computing power demand side, F i Represents the utility function of the i-th computing power demand side. The computing power demand side can dynamically adjust the expected computing power demand value of each time slot through game theory. And use the dynamic replication equation to represent the game process, by making the change rate The two goals of minimizing and maximizing the utility value are set, and the Nash equilibrium of the game is found based on the evolutionary game method, so as to output the computing power resource allocation results corresponding to multiple computing power demand ends corresponding to the Nash equilibrium.
[0073] In an optional embodiment, when each of the multiple computing power demand ends corresponds to multiple computing power demand strategies, constructing utility functions corresponding to the multiple computing power demand ends respectively includes: constructing a utility function corresponding to any one of the multiple computing power demand ends in the following manner:
[0074]
[0075] in, Represents the utility function corresponding to any computing power demand end when adopting any computing power demand strategy among multiple computing power demand strategies; Represents any computing power demand strategy; Indicates the computing power resource demand in any period of time in any computing power demand strategy; i indicates the identifier corresponding to any computing power demand end; k indicates the identifier corresponding to any computing power demand strategy; P t Indicates the price of computing power resources in any period of time; It represents the computing power resource consumption valuation function corresponding to any computing power demand end. The computing power resource consumption valuation function is used to indicate the relationship between the computing power resource consumption of the corresponding computing power demand end and the corresponding computing power demand strategy and demand priority. T represents the total number of multiple time periods included in any computing power demand strategy. The utility function corresponding to any computing power demand end is obtained to obtain the utility functions corresponding to multiple computing power demand ends respectively.
[0076] Optionally, first determine the utility functions corresponding to multiple computing power demand ends In the game scenario, each computing power demand side in the computing power network hopes to obtain the maximum benefit. By defining the utility function of the computing power demand side Specifically, the utility obtained by considering the relationship between subtask priorities and the utility obtained without considering the relationship between subtask priorities are calculated. The utility obtained by considering the subtask priorities is less than the utility obtained without considering the subtask priorities, that is, the utility value obtained by the utility function is a negative value, so that the maximum utility value is the optimization target, that is:
[0077]
[0078] The corresponding constraints are as follows:
[0079]
[0080] In an optional embodiment, a computing resource consumption estimation function corresponding to any computing demand end is constructed in the following manner:
[0081]
[0082] in, Represents the computing resource consumption estimation function corresponding to any computing demand side, Indicates the computing resource consumption value of any computing demand end in any period of time; Indicates the demand priority of any computing power demand side in any period of time; Indicates the demand for computing power resources in any period of time in any computing power demand strategy; the computing power resource price in any period of time is obtained by the following method:
[0083]
[0084] Among them, P t Indicates the price of computing power resources in any period of time; It represents the computing power demand of any computing power demand end in any time period, represents any computing power demand end, the sum of N computing power demand ends represents the total number of the multiple computing power demand ends, and N is an integer greater than or equal to 1; k1 represents the first proportional coefficient; k2 represents the second proportional coefficient.
[0085] Optionally, in order to express the different computing resource requirements of the computing demand side, the computing resource consumption estimation function corresponding to multiple computing demand sides is determined as Among them, the priority The specific computing resource consumption valuation function should satisfy the following properties:
[0086] 1. The computing power resource consumption valuation function is non-decreasing.
[0087] 2. In order to prevent the waste of computing power resources on the computing power demand side, and It should satisfy the law of diminishing marginal utility, that is, and is a nonlinear positive correlation, and as The bigger, The smaller the growth rate.
[0088] 3. In order to reduce the complexity of computing the utility of the computing power demand side during the game, and There is a linear positive correlation.
[0089] Based on the above properties, the total estimation function of computing resource consumption of subtask priority is defined as:
[0090]
[0091] Optionally, based on the actual situation of computing network resources, the computing resource price in the computing resource circulation platform will change over time, using the computing resource price function P t (DC t ) represents the price of a unit of computing power in time slot t, where the computing power resource price function is specified to satisfy the following properties: the computing power resource price function changes with the total computing power demand DC in time slot t t Increase and increase, among which At the same time, the marginal cost of the computing power resource price function will surge, which means that the greater the demand for computing power resources, the higher the computing power resource price. Based on this property, the computing power resource price function is as follows:
[0092] P t (DC t )=k1(DC t ) 2 +k2DC t
[0093] Among them, k1 represents the first proportional coefficient; k2 represents the second proportional coefficient, both of which are set by the computing power network administrator.
[0094] Through the above steps S102 to S108, the total amount of computing resources to be allocated in the computing power network, the total amount of network resources and the total amount of storage resources can be mapped to the same dimension, and the total amount of resources in the computing power network can be comprehensively considered to allocate computing power network resources from the two aspects of resource demand and supply, thereby effectively achieving the technical effect of more uniform network distribution of computing power resources, and further solving the technical problem of uneven distribution of computing power resources caused by the allocation of resources according to computing resources, storage resources and network resources in the computing power network from the perspective of computing power demand in related technologies.
[0095] Based on the above embodiments and optional embodiments, the present invention proposes an optional implementation mode: Figure 2 is a flowchart of an optional computing power network resource allocation method according to an embodiment of the present invention, such as Figure 2 As shown, computing power measurement and computing power modeling are the first step in the unified management and control of computing power network resources, which can provide guarantee for the flexible matching and allocation of computing network resources. This embodiment models the computing power provider resources and computing power consumer needs from the perspective of computing power providers and computing power consumers, respectively. The method includes:
[0096] S1, based on computing power network resources and demand modeling solutions, maps heterogeneous resources and resource requirements of different computing power demand ends to the same dimension.
[0097] S11, computing power provider resource modeling solution (computing power network resource modeling solution): starting from the three dimensions of computing, network, and storage, the computing power network resources are modeled separately.
[0098] S12, demand modeling solution for computing power consumers (computing power demand-side modeling solution):
[0099] With the development of artificial intelligence and mobile Internet, a large number of new applications such as autonomous driving and large models have emerged. These applications usually consume a large amount of multi-dimensional resources and are usually deployed in the form of microservices. At this time, a service can be split into multiple sub-task sets. In the computing network scenario, the services are divided according to resource requirements. After the service division is completed, the resource requirements of the sub-tasks are measured from the three dimensions of computing, network, and storage, and finally the resource requirements are normalized.
[0100] S2, on the basis of completing the same modeling measurement of resources and demands, converts the resource allocation problem into a process in which multiple users compete for computing resources.
[0101] S3, based on the idea of demand response, completes the construction of the computing power resource price function, and uses the fireworks game theory to describe the process of allocating computing power resources on the multi-computing power demand side.
[0102] A mathematical model of the computing power demand side response in the computing power network is constructed. The model includes heterogeneous computing power resource providers (such as cloud service providers, communication operators, big data centers, etc.), computing power resource circulation platforms, and multiple widely distributed computing power demand sides. Based on the above computing power network resource modeling scheme, computing power resources can be regarded as infinitely divisible commodities. The computing power demand side uses the computing power resource circulation platform to reasonably compete for computing power network resources that meet service needs based on game theory.
[0103] S4, construct utility functions corresponding to multiple computing power demand ends, and use dynamic replication equations to iteratively seek the optimal resource allocation strategy until the Nash equilibrium is reached.
[0104] S41, each computing power demand side calculates the benefits of the corresponding selected strategy based on the utility function and sends it to the computing power network resource provider platform;
[0105] S42, the computing power network resource circulation platform calculates the average income and replicates the dynamic equation Update the computing power network resource allocation strategy and send it to the computing power demand side;
[0106] S43, until the updated computing power network resource allocation strategy reaches Nash equilibrium.
[0107] Figure 3 is a schematic diagram of the deployment of a computing power network resource allocation method based on game theory according to an optional embodiment of the present invention. This embodiment can be used in Figure 3 It is used in the computing power network system shown in the figure. The computing power resource provider registers the computing power resource information in the computing power network platform and completes the authentication. The computing power network platform connects the computing power resource provider and the computing power resource consumer (computing power demand side) in series. On the basis of completing the quantification of computing power resources and demand, the computing power resource consumer sends a resource allocation request to the computing power network platform. At the same time, computing power resource consumers use evolutionary game to allocate computing power resources, realize the optimal allocation of computing power network resources, improve the utilization rate of computing power network resources, and improve the overall interests of consumers on the basis of ensuring the stable operation of consumers' businesses.
[0108] In this embodiment, a computing power network resource allocation device is also provided, which is used to implement the above embodiments and preferred implementation modes, and will not be repeated here. As used below, the terms "module" and "device" can implement a combination of software and / or hardware for a predetermined function. Although the devices described in the following embodiments are preferably implemented in software, the implementation of hardware, or a combination of software and hardware, is also possible and conceivable.
[0109] According to an embodiment of the present invention, there is also provided an apparatus embodiment for implementing the above-mentioned computing power network resource allocation method. Figure 4 is a schematic diagram of the structure of a computing power network resource allocation device according to an embodiment of the present invention. Figure 4 As shown, the computing power network resource allocation device includes: a total resource acquisition module 300, a dimension mapping module 301, a demand strategy acquisition module 302 and a resource allocation module 303, wherein:
[0110] The total resource acquisition module 300 is used to acquire the total amount of computing resources, network resources and storage resources to be allocated in the computing network, wherein the total amount of computing resources is used to indicate the total amount of resources used to perform computing tasks in the computing network, the network resources are used to indicate the total amount of resources used for network transmission in the computing network, and the storage resources are used to indicate the total amount of resources used to perform storage tasks in the computing network;
[0111] The dimension mapping module 301 is connected to the total resource acquisition module 300 and is used to map the total computing resources, the total network resources and the total storage resources to the same resource dimension to obtain a resource quantification result, wherein the resource quantification result is used to indicate the total resource amount after the total computing resources, the total network resources and the total storage resources are mapped to the same dimension;
[0112] The demand strategy acquisition module 302 is connected to the dimension mapping module 301 and is used to obtain the computing power demand strategies corresponding to the multiple computing power demand ends, wherein the computing power demand strategies are used to indicate the computing power resource demands corresponding to the corresponding computing power demand ends in multiple different time periods;
[0113] The resource allocation module 303 is connected to the demand strategy acquisition module 302, and is used to allocate computing power resources to the computing power demand end based on the resource quantification results and the computing power demand strategies corresponding to the multiple computing power demand ends, and obtain the computing power resource allocation results corresponding to the multiple computing power demand ends.
[0114] It should be noted that the above modules can be implemented by software or hardware. For example, for the latter, it can be implemented in the following ways: the above modules can be located in the same processor; or the above modules can be located in different processors in any combination.
[0115] It should be noted that the resource total amount acquisition module 300, the dimension mapping module 301, the demand strategy acquisition module 302 and the resource allocation module 303 correspond to steps S102 to S108 in the embodiment, and the examples and application scenarios implemented by the modules and the corresponding steps are the same, but are not limited to the contents disclosed in the above embodiments. It should be noted that the modules as part of the device can be run in a computer terminal.
[0116] It should be noted that the optional or preferred implementation of this embodiment can refer to the relevant description in the embodiment, which will not be repeated here.
[0117] The above-mentioned computing power network resource allocation device may also include a processor and a memory. The above-mentioned total resource acquisition module 300, dimension mapping module 301, demand strategy acquisition module 302 and resource allocation module 303 are all stored in the memory as program modules, and the processor executes the above-mentioned program modules stored in the memory to realize the corresponding functions.
[0118] The processor includes a kernel, which retrieves the corresponding program module from the memory. The kernel may be one or more. The memory may include non-permanent memory in a computer-readable medium, random access memory (RAM) and / or non-volatile memory, such as read-only memory (ROM) or flash RAM. The memory includes at least one memory chip.
[0119] According to an embodiment of the present application, an embodiment of a non-volatile storage medium is also provided. Optionally, in this embodiment, the non-volatile storage medium includes a stored program, wherein when the program is running, the device where the non-volatile storage medium is located is controlled to execute any of the computing power network resource allocation methods.
[0120] Optionally, in this embodiment, the non-volatile storage medium may be located in any computer terminal in a computer terminal group in a computer network, or in any mobile terminal in a mobile terminal group, and the non-volatile storage medium includes a stored program.
[0121] Optionally, when the program is running, the device where the non-volatile storage medium is located is controlled to perform the following functions: obtain the total amount of computing resources, the total amount of network resources, and the total amount of storage resources to be allocated in the computing power network, wherein the total amount of computing resources is used to indicate the total amount of resources used to perform computing tasks in the computing power network, the network resources are used to indicate the total amount of resources used for network transmission in the computing power network, and the storage resources are used to indicate the total amount of resources used to perform storage tasks in the computing power network; map the total amount of computing resources, the total amount of network resources, and the total amount of storage resources to the same resource dimension to obtain resource quantification results, wherein the resource quantification results are used to indicate the total amount of resources after the total amount of computing resources, the total amount of network resources, and the total amount of storage resources are mapped to the same dimension; obtain computing power demand strategies corresponding to multiple computing power demand ends, wherein the computing power demand strategies are used to indicate the computing power resource demands corresponding to the corresponding computing power demand ends in multiple different time periods; based on the resource quantification results and the computing power demand strategies corresponding to the multiple computing power demand ends, allocate computing power resources to the computing power demand ends to obtain computing power resource allocation results corresponding to the multiple computing power demand ends.
[0122] According to an embodiment of the present application, an embodiment of a processor is also provided. Optionally, in this embodiment, the processor is used to run a program, wherein when the program is run, any one of the computing power network resource allocation methods is executed.
[0123] According to an embodiment of the present application, an embodiment of a computer program product is also provided. Optionally, in this embodiment, the computer program product includes a computer program, and when the computer program is executed by a processor, the computer program implements any of the steps of the computing power network resource allocation method.
[0124] Optionally, the above-mentioned computer program product, when executed on a data processing device, is suitable for executing a program that initializes the following method steps: obtaining the total amount of computing resources, the total amount of network resources and the total amount of storage resources to be allocated in the computing power network, wherein the total amount of computing resources is used to indicate the total amount of resources used to perform computing tasks in the computing power network, the network resources are used to indicate the total amount of resources used for network transmission in the computing power network, and the storage resources are used to indicate the total amount of resources used to perform storage tasks in the computing power network; mapping the total amount of computing resources, the total amount of network resources and the total amount of storage resources to the same resource dimension to obtain a resource quantification result, wherein the resource quantification result is used to indicate the total amount of resources after the total amount of computing resources, the total amount of network resources and the total amount of storage resources are mapped to the same dimension; obtaining computing power demand strategies corresponding to multiple computing power demand ends respectively, wherein the computing power demand strategies are used to indicate the computing power resource demands corresponding to the corresponding computing power demand ends in multiple different time periods; based on the resource quantification results and the computing power demand strategies corresponding to the multiple computing power demand ends respectively, allocating computing power resources to the computing power demand ends to obtain computing power resource allocation results corresponding to the multiple computing power demand ends respectively.
[0125] An embodiment of the present invention provides an electronic device, which includes a processor, a memory, and a program stored in the memory and executable on the processor. When the processor executes the program, the following steps are implemented: obtaining the total amount of computing resources, the total amount of network resources, and the total amount of storage resources to be allocated in the computing power network, wherein the total amount of computing resources is used to indicate the total amount of resources used to perform computing tasks in the computing power network, the network resources are used to indicate the total amount of resources used for network transmission in the computing power network, and the storage resources are used to indicate the total amount of resources used to perform storage tasks in the computing power network; mapping the total amount of computing resources, the total amount of network resources, and the total amount of storage resources to the same resource dimension to obtain a resource quantification result, wherein the resource quantification result is used to indicate the total amount of resources after the total amount of computing resources, the total amount of network resources, and the total amount of storage resources are mapped to the same dimension; obtaining computing power demand strategies corresponding to multiple computing power demand ends, wherein the computing power demand strategies are used to indicate the computing power resource demands corresponding to the corresponding computing power demand ends in multiple different time periods; based on the resource quantification result and the computing power demand strategies corresponding to the multiple computing power demand ends, computing power resources are allocated to the computing power demand ends to obtain computing power resource allocation results corresponding to the multiple computing power demand ends.
[0126] The above sequence of the embodiments of the present invention is for description only and does not represent the superiority or inferiority of the embodiments.
[0127] In the above embodiments of the present invention, the description of each embodiment has its own emphasis. For parts that are not described in detail in a certain embodiment, reference can be made to the relevant descriptions of other embodiments.
[0128] In the several embodiments provided in this application, it should be understood that the disclosed technical content can be implemented in other ways. Among them, the device embodiments described above are only schematic. For example, the division of the above modules can be a logical function division. There may be other division methods in actual implementation, such as multiple modules or components can be combined or integrated into another system, or some features can be ignored or not executed. Another point is that the mutual coupling or direct coupling or communication connection shown or discussed can be through some interfaces, modules or indirect coupling or communication connection of modules, which can be electrical or other forms.
[0129] The modules described above as separate components may or may not be physically separated, and the components shown as modules may or may not be physical modules, that is, they may be located in one place or distributed on multiple modules. Some or all of the modules may be selected according to actual needs to achieve the purpose of the present embodiment.
[0130] In addition, each functional module in each embodiment of the present invention may be integrated into one processing module, or each module may exist physically separately, or two or more modules may be integrated into one module. The above-mentioned integrated module may be implemented in the form of hardware or in the form of software functional modules.
[0131] If the above-mentioned integrated module is implemented in the form of a software function module and sold or used as an independent product, it can be stored in a computer-readable non-volatile storage medium. Based on this understanding, the technical solution of the present invention is essentially or the part that contributes to the prior art or all or part of the technical solution can be embodied in the form of a software product. The computer software product is stored in a non-volatile storage medium, including a number of instructions to enable a computer device (which can be a personal computer, server or network device, etc.) to perform all or part of the steps of the various embodiments of the present invention. The aforementioned non-volatile storage medium includes: U disk, read-only memory (ROM, Read-Only Memory), random access memory (RAM, Random Access Memory), mobile hard disk, magnetic disk or optical disk and other media that can store program codes.
[0132] The above are only preferred embodiments of the present invention. It should be pointed out that, for ordinary technicians in this technical field, several improvements and modifications can be made without departing from the principles of the present invention. These improvements and modifications should also be regarded as the scope of protection of the present invention.
Claims
1. A computing power network resource allocation method, characterized in that: include: Obtain the total amount of computing resources, network resources, and storage resources to be allocated in the computing network, wherein the total amount of computing resources is used to indicate the total amount of resources used to perform computing tasks in the computing network, the network resources are used to indicate the total amount of resources used for network transmission in the computing network, and the storage resources are used to indicate the total amount of resources used to perform storage tasks in the computing network; Mapping the total amount of computing resources, the total amount of network resources, and the total amount of storage resources to the same resource dimension to obtain a resource quantification result, wherein the resource quantification result is used to indicate the total amount of resources after the total amount of computing resources, the total amount of network resources, and the total amount of storage resources are mapped to the same dimension; Obtain computing power demand strategies corresponding to multiple computing power demand ends respectively, wherein the computing power demand strategies are used to indicate computing power resource demands corresponding to the corresponding computing power demand ends in multiple different time periods respectively; Based on the resource quantification result and the computing power demand strategies respectively corresponding to the multiple computing power demand ends, computing power resources are allocated to the computing power demand ends to obtain computing power resource allocation results respectively corresponding to the multiple computing power demand ends.
2. The method according to claim 1, characterized in that The obtaining of the total amount of computing resources, network resources and storage resources to be allocated in the computing power network includes: Acquire multiple initial computing resources, multiple initial network resources, and multiple initial storage resources to be allocated in the computing network, wherein the multiple initial computing resources include at least two of the following: logical computing resources, parallel computing resources, and neural network computing resources; the multiple initial network resources include at least two of the following: network bandwidth resources, latency resources, and packet loss rate resources; the multiple initial storage resources include at least two of the following: read-only storage resources and random storage resources; Mapping the multiple initial computing resources to the same resource dimension to obtain the total amount of computing resources; Mapping the multiple initial network resources to the same resource dimension to obtain the total amount of the network resources; The multiple initial storage resources are mapped to the same resource dimension to obtain the total amount of storage resources.
3. The method according to claim 1, characterized in that The obtaining of computing power demand strategies corresponding to the plurality of computing power demand terminals respectively includes: The computing power resource demand of any computing power demand end in any period of time among the multiple computing power demand ends is obtained by: Determine the target task to be performed by any computing power demand end in any time period; Divide the target task into multiple subtasks, and obtain resource requirements corresponding to the multiple subtasks respectively; Determine resource mapping functions corresponding to the multiple subtasks respectively, wherein the resource mapping functions are used to indicate a mapping relationship between resource requirements of the corresponding subtasks and predetermined resource dimensions, and the predetermined resource dimensions are the same resource dimensions as the resource quantization results; Based on the resource mapping functions respectively corresponding to the multiple subtasks and the resource requirements respectively corresponding to the multiple subtasks, obtaining the resource mapping results respectively corresponding to the multiple subtasks; Based on the resource mapping results respectively corresponding to the multiple subtasks, the computing power resource demand of any computing power demand end in any time period is obtained; The computing power resource requirements corresponding to each of the computing power demand end in multiple time periods are obtained by obtaining the computing power resource requirements of each of the computing power demand end in the respective time periods; Based on the computing resource demands of any computing power demand end corresponding to the multiple time periods, obtaining a computing power demand strategy of any computing power demand end; The computing power demand strategies corresponding to the multiple computing power demand ends are obtained by obtaining the computing power demand strategy of any computing power demand end.
4. The method according to claim 3, characterized in that When the resource requirements include computing resource requirements, network resource requirements, and storage resource requirements, and the resource mapping function includes a computing resource mapping function, a network resource mapping function, and a storage resource mapping function, obtaining resource mapping results corresponding to the multiple subtasks based on the resource mapping functions corresponding to the multiple subtasks and the resource requirements corresponding to the multiple subtasks includes: The resource mapping result of any subtask among the multiple subtasks is obtained in the following manner: Based on the computing resource requirement of any subtask and the corresponding computing resource mapping function, obtain the computing resource mapping result of any subtask; Based on the network resource requirement of any subtask and the corresponding network resource function, obtaining a network resource mapping result of any subtask; Based on the storage resource requirement of any subtask and the corresponding storage resource mapping function, obtaining a storage resource mapping result of any subtask; Performing weighted calculation on the computing resource mapping result, the network resource mapping result, and the storage resource mapping result to obtain a resource mapping result of any subtask; The resource mapping results respectively corresponding to the plurality of subtasks are obtained by adopting a method of obtaining the resource mapping result of any one of the subtasks.
5. The method according to claim 1, characterized in that The computing power resource allocation to the computing power demand end based on the resource quantification result and the computing power demand strategies respectively corresponding to the multiple computing power demand ends is performed to obtain the computing power resource allocation results respectively corresponding to the multiple computing power demand ends, including: Determine the initial resource allocation results corresponding to the multiple computing power demand ends respectively; Constructing utility functions corresponding to the plurality of computing power demand ends respectively, wherein the utility functions are used to indicate the relationship between the computing power resource demand amount and the computing power resource benefit of the corresponding computing power demand end; Determine the utility values corresponding to the multiple computing power demand ends respectively based on the computing power demand strategies corresponding to the multiple computing power demand ends respectively and the utility functions corresponding to the multiple computing power demand ends respectively; Based on the resource quantification results, taking the maximum sum of the utility values corresponding to the multiple computing power demand ends as the optimization goal, a game theory algorithm is used to optimize the initial resource allocation results corresponding to the multiple computing power demand ends, so as to obtain the computing power resource allocation results corresponding to the multiple computing power demand ends.
6. The method according to claim 5, characterized in that In the case where each of the multiple computing power demand ends corresponds to multiple computing power demand strategies, constructing utility functions corresponding to the multiple computing power demand ends respectively includes: The utility function corresponding to any computing power demand end among the multiple computing power demand ends is constructed in the following manner: in, represents a utility function corresponding to any computing power demand end when any computing power demand strategy among multiple computing power demand strategies is adopted; Represents any of the computing power demand strategies; represents the computing power resource demand in any period of time in any computing power demand strategy; i represents the identifier corresponding to any computing power demand end; k represents the identifier corresponding to any computing power demand strategy; P t Indicates the price of computing power resources in any period of time; represents a computing power resource consumption valuation function corresponding to any computing power demand end, and the computing power resource consumption valuation function is used to indicate the relationship between the computing power resource consumption of the corresponding computing power demand end and the corresponding computing power demand strategy and demand priority; T represents the total number of multiple time periods included in any computing power demand strategy; The utility functions corresponding to the multiple computing power demand ends are obtained by obtaining the utility function corresponding to any one of the computing power demand ends.
7. The method according to claim 6, characterized in that The computing power resource consumption estimation function corresponding to any computing power demand end is constructed in the following way: in, represents the computing resource consumption estimation function corresponding to any computing demand end, Indicates the computing power resource consumption value of any computing power demand end in any time period; Indicates the demand priority of any computing power demand end in any time period; Indicates the computing power resource demand in any period of time in any computing power demand strategy; The computing power resource price for any period of time is obtained as follows: Among them, P t Indicates the price of computing power resources in any period of time; It represents the computing power demand of any computing power demand end in any time period, i=1,2,…,N represents any computing power demand end, N represents the total number of the multiple computing power demand ends, and N is an integer greater than or equal to 1; k1 represents the first proportional coefficient; k2 represents the second proportional coefficient.
8. A computing power network resource allocation device, characterized in that: include: A resource total amount acquisition module, used to acquire the total amount of computing resources, network resources and storage resources to be allocated in the computing network, wherein the computing resources are used to indicate the total amount of resources used to perform computing tasks in the computing network, the network resources are used to indicate the total amount of resources used for network transmission in the computing network, and the storage resources are used to indicate the total amount of resources used to perform storage tasks in the computing network; A dimension mapping module, used to map the total amount of computing resources, the total amount of network resources, and the total amount of storage resources to the same resource dimension to obtain a resource quantification result, wherein the resource quantification result is used to indicate the total amount of resources after the total amount of computing resources, the total amount of network resources, and the total amount of storage resources are mapped to the same dimension; A demand strategy acquisition module is used to obtain computing power demand strategies corresponding to multiple computing power demand ends, wherein the computing power demand strategies are used to indicate the computing power resource demands corresponding to the corresponding computing power demand ends in multiple different time periods; The resource allocation module is used to allocate computing power resources to the computing power demand end based on the resource quantification results and the computing power demand strategies corresponding to the multiple computing power demand ends, and obtain the computing power resource allocation results corresponding to the multiple computing power demand ends.
9. A non-volatile storage medium, characterized in that: The non-volatile storage medium stores a plurality of instructions, and the instructions are suitable for being loaded by a processor and executing the computing power network resource allocation method described in any one of claims 1 to 7.
10. A computer program product, comprising a computer program, characterized in that When the computer program is executed by a processor, the steps of the computing power network resource allocation method described in any one of claims 1 to 7 are implemented.
Citation Information
Patent Citations
Block chain computing power optimization scheduling method oriented to intelligent data processing
CN112769641A
Method, device and equipment for scheduling computing power network resources and medium
CN116366576A
New-generation network service distribution method and device, electronic equipment and storage medium
CN116390169A
Optimization method and optimization device for computing power resource allocation, electronic equipment and medium
CN116541176A
Allocation of computational resources with policy selection
US20130246208A1
