A data processing method, device and readable storage medium based on edge computing

Abstract

The present application discloses a data processing method, device and readable storage medium based on edge computing. The method includes: obtaining the occupied computing power resource information of a target application for an edge computing node; predicting new service objects, offline service objects and scenario switching service objects for the target application within a target time period, and determining the expected required computing power resource information of the target application for the edge computing node within the target time period according to the new service objects, offline service objects and scenario switching service objects; obtaining the current operating frequency of the edge computing node, and determining the target operating frequency of the edge computing node within the target time period according to the maximum computing power resource information corresponding to the current operating frequency, the occupied computing power resource information and the expected required computing power resource information. By adopting the present application, the operating cost of the edge computing node can be reduced while meeting the computing power requirements.

Description

Technical Field

[0001] The present application relates to the field of computer technology, and in particular to a data processing method, device and readable storage medium based on edge computing. Background Art

[0002] Cloud gaming refers to the process of running a game on a remote server, compressing and encoding the rendered game screen, and then sending it to the terminal through the network in the form of audio and video streams. Cloud gaming does not need to consider terminal configuration, which completely solves the technical problem of insufficient terminal performance and inability to run heavy games. However, cloud gaming has very high requirements for network latency. In order to provide more stable network conditions for the object, cloud gaming servers are generally closer to the object through large-scale deployment of edge computing nodes.

[0003] However, the online number of game objects has a relatively obvious tidal phenomenon. In order to provide a better experience for the objects, the operating frequency of the edge computing nodes is generally set according to the maximum number of online game objects, and computing power is prepared for the game objects with the highest number of online games. During non-peak hours, the actual number of online game objects is far less than the maximum number of online games, so this part of the computing power resources will be idle during non-peak hours.

[0004] That is to say, if the operating frequency is set according to the maximum number of online users, the edge computing node does not need to run at the operating frequency corresponding to the maximum number of online users during non-peak hours. At this time, the operating frequency of the edge computing node will be too high, and the power consumption of the edge computing node will also be large, resulting in a waste of operating costs for the edge computing node. From the above, it can be seen that if the computing power is prepared according to the maximum number of online users of the game object, the operating cost will be greatly increased. Summary of the invention

[0005] The embodiments of the present application provide a data processing method, device, and readable storage medium based on edge computing, which can reduce the operating costs of edge computing nodes while meeting computing power requirements.

[0006] On the one hand, an embodiment of the present application provides a data processing method based on edge computing, including:

[0007] Obtain the computing power resource information occupied by the target application for the edge computing node; the occupied computing power resource information is the computing power resource information of the edge computing node occupied when running the target application;

[0008] Predict the new business objects, offline business objects, and scene switching business objects for the target application within the target time period, and determine the expected demand computing power resource information for the edge computing nodes of the target application within the target time period based on the new business objects, offline business objects, and scene switching business objects;

[0009] Obtain the current operating frequency of the edge computing node, and determine the target operating frequency of the edge computing node within the target time period according to the maximum computing power resource information, the occupied computing power resource information, and the expected required computing power resource information corresponding to the current operating frequency; the maximum computing power resource information corresponding to the target operating frequency meets the expected required computing power resource information.

[0010] On the one hand, an embodiment of the present application provides a data processing device based on edge computing, including:

[0011] An occupied computing power acquisition module, configured to obtain the occupied computing power resource information of the target application for the edge computing node; the occupied computing power resource information is the computing power resource information of the edge computing node occupied when running the target application;

[0012] An object prediction module, configured to predict new service objects, offline service objects, and scenario switching service objects for the target application within the target time period;

[0013] An expected computing power determination module, configured to determine the expected required computing power resource information of the target application for the edge computing node within the target time period according to the new service objects, offline service objects, and scenario switching service objects;

[0014] A frequency determination module, configured to obtain the current operating frequency of the edge computing node;

[0015] The frequency determination module is further configured to determine the target operating frequency of the edge computing node within the target time period according to the maximum computing power resource information, the occupied computing power resource information, and the expected required computing power resource information corresponding to the current operating frequency; the maximum computing power resource information corresponding to the target operating frequency meets the expected required computing power resource information.

[0016] In one embodiment, the frequency determination module includes:

[0017] An idle computing power determination unit, configured to determine the current idle computing power resource information of the edge computing node according to the maximum computing power resource information and the occupied computing power resource information corresponding to the current operating frequency;

[0018] An operating state determination unit, configured to determine the operating state of the edge computing node according to the current idle computing power resource information;

[0019] A target frequency determination unit, configured to, if the operating state of the edge computing node is a full-load operating state, determine the target operating frequency of the edge computing node within the target time period according to the occupied computing power resource information and the expected required computing power resource information;

[0020] The target frequency determination unit is further configured to, if the operating state of the edge computing node is an under-loaded operating state and the expected required computing power resource information is greater than the occupied computing power resource information, determine the target operating frequency of the edge computing node within the target time period according to the current idle computing power resource information, the occupied computing power resource information, and the expected required computing power resource information.

[0021] In one embodiment, the operating state determination unit includes:

[0022] A matching subunit, configured to match the current idle computing power resource information with the idle resource threshold;

[0023] A state determination subunit, configured to, if the current idle computing power resource information is greater than the idle resource threshold, determine the operating state of the edge computing node as an under-loaded operating state;

[0024] The state determination subunit is further configured to, if the current idle computing power resource information is less than the idle resource threshold, determine the operating state of the edge computing node as a fully-loaded operating state.

[0025] In one embodiment, the target frequency determination unit includes:

[0026] A computing power comparison subunit, configured to, if the operating state of the edge computing node is a fully-loaded operating state, compare the expected required computing power resource information with the occupied computing power resource information;

[0027] A frequency acquisition subunit, configured to, if the expected required computing power resource information is greater than the occupied computing power resource information, acquire a mapping table; the mapping table includes N mapping relationships; one mapping relationship includes a corresponding relationship between a configured operating frequency and a configured maximum computing power resource information;

[0028] The frequency acquisition subunit is further configured to, in the mapping table, determine the configured operating frequency corresponding to the configured maximum computing power resource information greater than the expected required computing power resource information as the target operating frequency;

[0029] A first frequency adjustment subunit, configured to adjust the operating frequency of the edge computing node from the current operating frequency to the target operating frequency.

[0030] In one embodiment, the target frequency determination unit further includes:

[0031] A first frequency determination subunit, configured to, if the expected required computing power resource information is less than the occupied computing power resource information, determine the current operating frequency as the target operating frequency.

[0032] In one embodiment, the target frequency determination unit includes:

[0033] A difference determination subunit, configured to determine the absolute value of the resource difference between the expected required computing power resource information and the occupied computing power resource information if the operating state of the edge computing node is an under-loaded operating state and the expected required computing power resource information is greater than the occupied computing power resource information;

[0034] A difference comparison subunit, configured to obtain a mapping table if the current idle computing power resource information is less than the absolute value of the resource difference; the mapping table includes N mapping relationships; one mapping relationship includes a corresponding relationship between a configured operating frequency and a configured maximum computing power resource information;

[0035] The difference comparison subunit is further configured to determine, in the mapping table, the configured operating frequency corresponding to the configured maximum computing power resource information greater than the expected required computing power resource information as the target operating frequency;

[0036] A second frequency adjustment subunit, configured to adjust the operating frequency of the edge computing node from the current operating frequency to the target operating frequency.

[0037] In one embodiment, the target frequency determination unit further includes:

[0038] A second frequency determination subunit, further configured to determine the current operating frequency as the target operating frequency if the current idle computing power resource information is greater than the absolute value of the resource difference.

[0039] In one embodiment, the frequency determination module further includes:

[0040] A table acquisition unit, configured to obtain a mapping table if the operating state of the edge computing node is an under-loaded operating state and the expected required computing power resource information is less than the occupied computing power resource information; the mapping table includes N mapping relationships; one mapping relationship includes a corresponding relationship between a configured operating frequency, a configured maximum computing power resource information, and a configured operating consumption;

[0041] A consumption acquisition unit, configured to determine the configured operating consumption corresponding to the current operating frequency in the mapping table as the to-be-compared configured operating consumption;

[0042] The consumption acquisition unit is further configured to determine the target operating frequency of the edge computing node within a target time period according to the to-be-compared configured operating consumption.

[0043] In one embodiment, the consumption acquisition unit is further specifically configured to determine, in the mapping table, the configured operating consumption corresponding to the configured maximum computing power resource information greater than the expected required computing power resource information as the target operating consumption;

[0044] The consumption acquisition unit is further specifically configured to compare the target operating consumption with the to-be-compared configured operating consumption;

[0045] The consumption acquisition unit is further specifically configured to, if the target running consumption is less than the to-be-compared configured running consumption, determine the configured running frequency corresponding to the target running consumption as the target running frequency, and adjust the running frequency of the edge computing node from the current running frequency to the target running frequency;

[0046] The consumption acquisition unit is further specifically configured to, if the target running consumption is greater than the configured running consumption, determine the current running frequency as the target running frequency.

[0047] In one embodiment, the used computing power acquisition module includes:

[0048] The running information acquisition unit is configured to acquire N current running information of the target application; one current running information includes an application activity scenario where an online service object is located, and the resource occupation information of the service object in the application activity scenario;

[0049] The used resource statistics unit is configured to acquire the resource occupation information of the service object corresponding to each current running information, and obtain N pieces of resource occupation information of the service object;

[0050] The used resource statistics unit is further configured to determine the total sum of the N pieces of resource occupation information of the service object as the occupied computing power resource information of the target application for the edge computing node.

[0051] In one embodiment, the number of application activity scenarios included in the N current running information is one or more; the one or more application activity scenarios include application activity scenario M i ; i is a positive integer;

[0052] The expected computing power determination module includes:

[0053] The object computing power statistics unit is configured to determine the online service object whose application activity scenario is application activity scenario M among the N online service objects as the to-be-statistic service object; i The object computing power statistics unit is further configured to statistically calculate the total resource occupation information of the service object in the application activity scenario M

[0054] The object computing power statistics unit is further configured to statistically calculate the total resource occupation information of the service object in the application activity scenario M i ;

[0055] The average computing power determination unit is configured to obtain the number of objects corresponding to the to-be-statistic service object, and determine the average computing power requirement information corresponding to the application activity scenario M according to the total resource occupation information of the service object and the number of objects; i The corresponding average computing power requirement information;

[0056] An expected computing power determination unit, configured to, when determining the average computing power demand information corresponding to one or more application activity scenarios respectively, determine the expected demand computing power resource information of the target application for the edge computing node within the target time period according to the average computing power demand information corresponding to one or more application activity scenarios respectively, the newly added service objects, the offline service objects, and the scenario switching service objects.

[0057] In one embodiment, the expected computing power determination unit is further specifically configured to obtain the newly added quantity corresponding to the newly added service objects, the offline quantity corresponding to the offline service objects, and the switching quantity corresponding to the scenario switching objects;

[0058] The expected computing power determination unit is further specifically configured to predict the initial login scenario corresponding to the newly added service objects; one or more application activity scenarios include the initial login scenario;

[0059] The expected computing power determination unit is further specifically configured to determine the offline application activity scenario as the application activity scenario in which the offline service objects are located when the offline behavior occurs;

[0060] The expected computing power determination unit is further specifically configured to determine the initial application activity scenario as the application activity scenario in which the scenario switching objects are located before the scenario switching, and determine the target application activity scenario as the application activity scenario in which the scenario switching objects are located after the scenario switching;

[0061] The expected computing power determination unit is further specifically configured to determine the expected demand computing power resource information of the target application for the edge computing node within the target time period according to the average computing power demand information corresponding to the initial login scenario, the offline application activity scenario, the initial application activity scenario, the target application activity scenario, as well as the switching quantity, the offline quantity, and the newly added quantity.

[0062] In one embodiment, the expected computing power determination unit is further specifically configured to determine the first changed computing power resource information corresponding to the newly added service objects according to the newly added quantity and the average computing power demand information corresponding to the initial login scenario;

[0063] The expected computing power determination unit is further specifically configured to determine the second changed computing power resource information corresponding to the offline service objects according to the offline quantity and the average computing power demand information corresponding to the online application activity scenario;

[0064] The expected computing power determination unit is further specifically configured to determine the third changed computing power resource information corresponding to the scenario switching objects according to the average computing power demand information corresponding to the initial application activity scenario, the average computing power demand information corresponding to the target application activity scenario, and the switching quantity;

[0065] The expected computing power determination unit is further specifically configured to determine the expected required computing power resource information of the target application for the edge computing node within the target time period according to the first variable computing power resource information, the second variable computing power resource information, the third variable computing power resource information, and the occupied computing power resource information of the edge computing node.

[0066] On the one hand, an embodiment of the present application provides a computer device, including: a processor and a memory;

[0067] The memory stores a computer program, and when the computer program is executed by the processor, the processor is caused to execute the method in the embodiment of the present application.

[0068] On the one hand, an embodiment of the present application provides a computer-readable storage medium, which stores a computer program, and the computer program includes program instructions. When the program instructions are executed by the processor, the method in the embodiment of the present application is executed.

[0069] In one aspect of the present application, there is provided a computer program product or a computer program. The computer program product or the computer program includes computer instructions, and the computer instructions are stored in a computer-readable storage medium. The processor of the computer device reads the computer instructions from the computer-readable storage medium, and the processor executes the computer instructions, so that the computer device executes the method provided in one aspect of the embodiment of the present application.

[0070] In the embodiments of the present application, since the main factors affecting the change of computing power resources of edge computing nodes are factors such as the online, offline, and online of business objects, the present application can predict new business objects, offline business objects, and scenario switching objects for a target application within a target time period. Subsequently, the expected demand computing power resource information of the target application for the edge computing node within the target time period can be determined based on the new business objects, offline business objects, and scenario switching business objects. Thus, based on the expected demand computing power resource information of the target time period, the occupied computing power resource information of the edge computing node, and the maximum computing power resource information corresponding to the current operating frequency of the edge computing node, it can be determined whether to increase, decrease, or keep the current operating frequency unchanged within the target time period (i.e., determine the target operating frequency within the target time period). By predicting the target operating frequency within the target time period, the target operating frequency within the target time period can be determined in advance, so that the current operating frequency of the edge computing node can be adaptively adjusted, so that the operating frequency of the edge computing node within the target time period can meet the target operating frequency, that is, the computing power resources provided by the edge computing node within the target time period match the expected demand computing power resource information. By predicting the target operating frequency within the target time period, the operating frequency of the edge computing node can be adaptively and dynamically adjusted, thereby reducing the waste of power consumption, and thus reducing the waste of operating costs; at the same time, by dynamically adjusting the operating frequency, the expected demand computing power resource information can also be satisfied. That is, the present application can reduce the operating cost of the edge computing node while meeting the computing power demand by dynamically adjusting the operating frequency of the edge computing node. BRIEF DESCRIPTION OF THE DRAWINGS

[0071] In order to more clearly illustrate the technical solutions in the embodiments of the present invention or the prior art, the following will briefly introduce the drawings required for use in the description of the embodiments or the prior art. Obviously, the following drawings are only some embodiments of the present invention. For those of ordinary skill in the art, without creative efforts, other drawings can also be obtained based on these drawings.

[0072] Figure 1 is a schematic diagram of a network architecture provided by an embodiment of the present application;

[0073] Figures 2a - 2b is a scenario diagram for predicting the target operating frequency within a target time period provided by an embodiment of the present application;

[0074] Figure 3 is a schematic flowchart of a data processing method based on edge computing provided by an embodiment of the present application;

[0075] Figure 4It is a logic flowchart for adjusting the operating frequency of an edge computing node provided by an embodiment of the present application;

[0076] Figure 5 It is a schematic flowchart for determining the target operating frequency within a target time period provided by an embodiment of the present application;

[0077] Figure 6 It is a system architecture diagram provided by an embodiment of the present application;

[0078] Figure 7 It is a schematic structural diagram of a data processing device based on edge computing provided by an embodiment of the present application;

[0079] Figure 8 It is a schematic structural diagram of a computer device provided by an embodiment of the present application. Detailed implementation manners

[0080] Next, the technical solutions in the embodiments of the present application will be clearly and completely described in conjunction with the accompanying drawings in the embodiments of the present application. Obviously, the described embodiments are only a part of the embodiments of the present application, rather than all the embodiments. All other embodiments obtained by those of ordinary skill in the art based on the embodiments of the present application without creative efforts shall fall within the protection scope of the present application.

[0081] The present application relates to technologies such as cloud computing, cloud gaming, and edge computing. The following will first elaborate on related concepts such as cloud computing, cloud gaming, and edge computing.

[0082] Cloud computing refers to the delivery and usage model of IT infrastructure, which means obtaining the required resources in a on-demand and easily scalable manner through the network; in a broad sense, cloud computing refers to the delivery and usage model of services, which means obtaining the required services in a on-demand and easily scalable manner through the network. Such services can be related to IT and software, the Internet, or other services. Cloud computing is the product of the development and integration of traditional computer and network technologies such as grid computing, distributed computing, parallel computing, utility computing, network storage technologies, virtualization, and load balance.

[0083] With the development of the Internet, real-time data streams, and the diversification of connected devices, as well as the driving forces of demands such as search services, social networks, mobile commerce, and open collaboration, cloud computing has developed rapidly. Different from previous parallel distributed computing, the emergence of cloud computing will, in concept, promote revolutionary changes in the entire Internet model and enterprise management model.

[0084] Cloud gaming, also known as gaming on demand, is an online gaming technology based on cloud computing technology. Cloud gaming technology enables thin clients with relatively limited graphics processing and data computing capabilities to run high-quality games. In the cloud gaming scenario, the game does not run on the player's game terminal but on the cloud server, and the cloud server renders the game scene into a video and audio stream and transmits it to the player's game terminal through the network. The player's game terminal only needs to have basic streaming media playback capabilities and the ability to obtain the player's input instructions and send them to the cloud server.

[0085] Edge computing refers to an open platform that integrates network, computing, storage, and application core capabilities on the side close to the object or data source to provide the nearest-end services nearby. Its application programs are initiated on the edge side, generating faster network service responses and meeting the basic requirements in aspects such as real-time services, application intelligence, security, and privacy protection.

[0086] A cloud gaming edge computing node, a node used for edge computing, is generally composed of multiple servers with graphics processing unit (GPU) computing capabilities. A single server among them can be called a computing node.

[0087] Computing power, as the name implies, is the computing ability of a device. From mobile phones and computers to supercomputers, computing power exists in various hardware devices. Computing power resources are the hardware or network resources required when a device executes a computing task, and usually can include central processing unit (CPU) computing power resources, GPU computing power resources, memory resources, network bandwidth resources, and disk resources.

[0088] The solution provided in the embodiments of this application relates to cloud computing and cloud gaming technologies in the field of cloud technology, and the specific process is described through the following embodiments.

[0089] Please refer to Figure 1 , Figure 1 which is a schematic diagram of a network architecture provided by the embodiments of this application. As Figure 1As shown in the figure, the network architecture may include a management server 100 and edge nodes 11, 12, …, 1n. Among them, edge node 11 may include multiple computing servers such as computing servers 11a and 11b, and edge node 12 may include multiple computing servers such as computing servers 12a and 12b. As Figure 1 shown in the figure, the computing servers such as computing servers 11a and 11b in edge node 11 can communicate with each other, and the computing servers such as computing servers 12a and 12b in edge node 12 can communicate with each other. Any computing server in edge node 11, any computing server in edge node 12, …, any computing server in edge node 1n can be respectively network-connected to the above management server 100, so that each computing server can perform data interaction with the management server 100 through the network connection, and each computing server can receive management data from the above management server 100. It can be understood that the computing servers in the edge nodes are usually deployed in the same area, while different edge nodes are usually deployed in different areas.

[0090] As Figure 1 shown in the figure, the computing servers in the above edge nodes can all correspond to a terminal device cluster, and a target application can be integrally installed on each terminal device in the terminal device cluster. When the target application runs on each terminal device, it can perform data interaction with the computing server allocated by the management server 100. Among them, the target application may include one or more applications such as game applications, video editing applications, social applications, instant messaging applications, live broadcast applications, short video applications, video applications, music applications, shopping applications, novel applications, payment applications, browsers, etc., which have functions of displaying data information such as text, images, audio, and video. The computing server provides corresponding functional services for the target application running in the terminal device, but at the same time consumes corresponding computing power resources. The computing power resources of one computing server can correspond to different terminal devices at the same time. When the terminal device accessing the computing server runs a target application (for example, a cloud game application), the target application will occupy the computing power resources of the computing server. Each computing server can also be called an edge computing node.

[0091] In order for the objects of the target application to run the target application smoothly, usually the computing server will prepare computing power resources for the target application according to the maximum online number of the target application (the number of objects logged in to the target application) (set the operating frequency according to the maximum online number and prepare computing power resources according to the operating frequency); among them, the operating frequency can refer to the frequency of the computing module (such as CPU, GPU, etc.) in the computing server when it is running. For example, usually the maximum online number of the target application is 500, then the operating frequency set by the computing server for the target application is the frequency that can provide functional services for 500 objects, and the prepared computing power resources are the computing power resources corresponding to this operating frequency. If the number of online objects of the target application is much less than the maximum online number of 500 within a certain period of time, then the operating frequency of the computing server is too large (the operating frequency corresponding to the maximum online number of 500), and the actually required operating frequency can be smaller. Then, at this time, it will cause waste of the computing power resources of the computing server and also cause excessive power consumption of the computing server; similarly, if the number of online objects of the target application is much more than the maximum online number of 500 within a certain period of time, then the functional services that can be provided by the operating frequency of the computing server at this time may not be sufficient to support the operation of the target application, and then it will cause a situation where the requirements of the target application cannot be met. In order to enable the computing server to better meet the requirements of the target application for computing power resources while reducing the power consumption of the computing server, the management server 100 will perform data interaction with each computing server to obtain the current operating frequency of each computing server, so as to obtain the total computing power resource information (i.e., the maximum computing power resource information) corresponding to the current operating frequency; the management server 100 can also obtain the occupied computing power resource information of the computing server (i.e., the computing power resources of the computing server occupied by the target application when it is running, such as the CPU computing power resources of the computing server occupied by the target application when running the target application), and predict the expected demand computing power resource information of the target application for the computing server within the target time period (i.e., within the target time period, the computing power resources of the computing server that the target application is expected to occupy, such as the CPU computing power resources of the computing server that the target application may occupy when running within the target time period); subsequently, the management server 100 can, according to the expected demand computing power resource information, the occupied computing power resource information, and the current operating frequency (the current operating frequency can refer to the frequency of the computing module (such as CPU, GPU, etc.) in the computing server when it is currently running.When the occupied computing power resource information and the desired required computing power resource information refer to CPU computing power resources, the current operating frequency may refer to the maximum computing power resource information corresponding to the frequency of the CPU in the computing server during current operation. Determine the target operating frequency of the computing server within the target time period (the target operating frequency may refer to the expected operating frequency of the computing server within the target time period. The target operating frequency can be obtained by increasing or decreasing the current operating frequency, and the target operating frequency may also be equal to the current operating frequency). It should be noted that the maximum computing power resource information corresponding to the target operating frequency satisfies the desired required computing power resource information. That is to say, by dynamically adjusting the operating frequency of the computing server, the maximum computing power resource information corresponding to the target operating frequency of the computing server within the target time period can satisfy the desired required computing power resource information, thereby reducing the situation where the operating frequency of the computing server within the target time period is too high or too low, and then causing excessive power consumption or inability to meet the computing power resource requirements.

[0092] For ease of understanding, taking the management server 100 to determine the target operating frequency of the computing server 11a within the target time period as an example for illustration, as Figure 1 shown, the management server 100 will obtain the occupied computing power resource information of the target application for the computing server 11a (that is, the computing power resource information of the computing server 11a occupied when running the target application). Among them, the occupied computing power resource information may refer to the quantitative index information of the computing power resources occupied when the computing server 11a provides functional services for the target application. Among them, the quantitative index information may include one or more of the central processing unit computing power information (CPU), graphics processing unit computing power information (GPU), memory usage information, network bandwidth usage information, disk read and write ability information, and other index information. Taking the quantitative index information including CPU computing power and GPU computing power as an example, the occupied computing power resource information is the value of the CPU computing power and GPU computing power of the computing server 11a occupied when running the target application.

[0093] Subsequently, the management server 100 may predict the expected required computing power resource information of the target application for the computing server 11a within the target time period (i.e., the computing power resource information that is expected to be occupied when running the target application within the target time period). The management server 100 may obtain the current operating frequency of the computing server 11a. Based on the maximum computing power resource information corresponding to the current operating frequency, the occupied computing power resource information, and the expected computing power demand resource information, the management server 100 may determine the target operating frequency of the computing server 11a within the target time period. Among them, the maximum computing power resource information corresponding to the target operating frequency satisfies the expected required computing power resource information. Then, when in the target time period, the operating frequency of the computing server 11a may be adjusted from the current operating frequency to the target operating frequency, so that the computing power resources provided by the computing server 11a for the target application within the target time period will not be too large or too small. Among them, for the specific implementation method of predicting the expected required computing power resource information of the target application for the computing server 11a within the target time period and determining the target operating frequency within the target time period, reference can be made to the description in the subsequent Figure 3 corresponding embodiments.

[0094] It can be understood that the above processing process may be executed by the management server alone, or by the computing server alone, or jointly by the management server and the computing server. The specific implementation may be adjusted according to actual requirements and is not limited here.

[0095] It can be understood that the method provided in the embodiments of the present application may be executed by a computer device, and the computer device includes but is not limited to a terminal device, a computing server, or a management server. Among them, the management server may be an independent physical server, or a server cluster or distributed system composed of multiple physical servers, or a cloud server that provides basic cloud computing services such as cloud services, cloud databases, cloud computing, cloud functions, cloud storage, network services, cloud communications, middleware services, domain name services, security services, CDN, and big data and artificial intelligence platforms.

[0096] It can be understood that the above devices (such as the above management server 100, computing server 11a, computing server 11b, computing server 12a, …, computing server 12b, etc.) can be a node in a distributed system, where the distributed system can be a blockchain system, and the blockchain system can be a distributed system formed by connecting the multiple nodes in the form of network communication. Among them, the nodes can form a peer-to-peer (P2P) network, and the P2P protocol is an application layer protocol running on top of the Transmission Control Protocol (TCP). In the distributed system, any form of computer device, such as servers, terminal devices and other electronic devices, can become a node in the blockchain system by joining the peer-to-peer network.

[0097] Among them, the terminal devices in the above terminal device cluster can include mobile phones, tablet computers, laptop computers, palmtop computers, smart speakers, mobile internet devices (MIDs), POS (Point Of Sales) machines, wearable devices (such as smart watches, smart bracelets, etc.), in-vehicle devices, etc.

[0098] For easy understanding, please refer to Figures 2a - 2b , Figures 2a - 2b is a scenario diagram provided by an embodiment of the present application for predicting the target operating frequency within a target time period. As Figure 2a shown in the scenario, taking a multiplayer competitive cloud game application as an example of the target application, service object A can run the cloud game application through the corresponding terminal device 10a, service object B can run the cloud game application through the corresponding terminal device 10b, …, service object N can run the cloud game application through the corresponding terminal device 10n. And the computing server 11b provides service support for the cloud game applications in the terminal devices 10a, terminal devices 10b, …, terminal devices 10n. Among them, the computing server 11b can also be called a cloud game server or an edge computing node.

[0099] It can be understood that different application activity scenarios may be included in the cloud game application. The business objects logging in to the cloud game application (such as business object A, business object B, …, business object N; a business object may refer to the bound account of a business user who uses a terminal device to run the cloud game application in the cloud game application. The business user can use the bound account to log in to the cloud game application, and the cloud game application can also determine whether the business user has logged in through the bound account) can enter different application activity scenarios and play games in different application activity scenarios. The computing power resources required for different application activity scenarios are usually different. Among them, the application activity scenario may refer to the scenario type to which the picture displayed on the terminal device belongs when the cloud game application runs on the terminal device. An application activity scenario can provide corresponding application functions for the cloud game application. For example, the application activity scenario may include a game home page hall scenario (that is, when the business user logs in or opens the cloud game application, the cloud game application usually presents a default picture, which can be used for presenting game characters, changing character costumes, selecting competitive modes, etc. This default presented picture is usually called the game home page hall), a single-player competitive scenario (that is, a mode in the cloud game application where single-player combat can be carried out. This single-player competitive scenario is used for single-player parachuting, single-player level breaking, etc.), a multi-player competitive scenario (that is, a mode in the cloud game application where multi-player combat can be carried out. This multi-player competitive scenario can be used for multi-player teaming up to fight monsters to improve the game level, multi-player teaming up to break levels, etc.), and so on. Usually, in the game home page hall scenario, simple pictures and controls such as presenting the game background, introducing the game characters of the business object, and selecting game competitive modes need to be presented to the business object, so the computing power resources required for the game home page hall scenario are also relatively small (the computing server 11b can run at a relatively low frequency); while in the single-player competitive scenario, the business object can carry out single-player competitive games (for example, operating the game character alone to increase game experience or obtain game coins, etc.). Compared with the game home page hall scenario, there are more operation instructions from the business object in the single-player competitive scenario, so there will be more corresponding calculations, and thus the computing power resources required in the single-player competitive scenario are also more (the computing server 11b needs to run at a higher frequency); while in the multi-player competitive scenario, the business object can perform a large number of game operations (such as sliding the direction wheel, releasing the skills of the game character, clicking the retreat control). In addition, voice communication and text communication can be carried out between business objects. That is to say, in addition to meeting the game operation instructions of the business object, the multi-player competitive scenario also needs to provide computing services for communication between business objects. Therefore, compared with the single-player competitive scenario, the computing power resources required in the multi-player competitive scenario are more (the computing server 11b needs to run at a higher frequency).

[0100] such as Figure 2aAs shown in the figure, taking business objects A, B, …, N (including business objects 1, 2, 3, …, 10) as an example, business objects 1, 2, 3, …, 10 are all in an online state, and business objects 1, 2, 3, …, 10 can be referred to as current online business objects. The management server 100 can obtain the application activity scenarios in which each of business objects 1, 2, 3, …, 10 is located, and when each business object is in the corresponding application activity scenario, obtain the computing power resources occupied thereby (hereinafter referred to as business object occupied resource information); taking the application activity scenarios where business objects 1 - 5 are located as the game hall scenario, the application activity scenarios where business objects 6 - 8 are located as the single-player competition scenario, and the application activity scenarios where business objects 9 - 10 are located as the multi-player competition scenario as an example, among which, the business object occupied resource information corresponding to business object 1 is 70, the business object occupied resource information corresponding to business object 2 is 71, the business object occupied resource information corresponding to business object 3 is 72, the business object occupied resource information corresponding to business object 4 is 70, and the business object occupied resource information corresponding to business object 5 is 73. Then, the management server 100 can add up the business object occupied resource information corresponding to business objects 1 - 5, and thus the total occupied computing power resources of the game home page hall scenario can be obtained as 356 (i.e., 70 + 71 + 72 + 70 + 73).

[0101] Among them, the business object occupied resource information corresponding to business object 6 is 101, the business object occupied resource information corresponding to business object 7 is 110, and the business object occupied resource information corresponding to business object 8 is 108. Then, the management server 100 can add up the business object occupied resource information corresponding to business objects 6 - 8, and thus the total occupied computing power resources of the single-player competition scenario can be obtained as 319 (i.e., 101 + 110 + 108); the business object resource information corresponding to business object 9 is 200, and the business object occupied resource information corresponding to business object 10 is 208. Then, the management server 100 can add up the business object occupied resource information corresponding to business objects 9 - 10, and thus the total occupied computing power resources of the multi-player competition scenario can be obtained as 408 (i.e., 200 + 208). Further, the management server 100 can add up the total occupied computing power resources corresponding to the game home page hall scenario, the single-player competition scenario, and the multi-player competition scenario respectively, and thus the occupied computing power resource information of the computing server 11b can be obtained as 1083.

[0102] Further, as Figure 2bAs shown, the management server 100 can predict new business objects for the cloud game application within the target time period (i.e., the bound accounts corresponding to business users who log in or open the cloud game application for the first time within the target time period, and these new business objects do not belong to the above-mentioned currently online business objects), offline business objects (i.e., among the above-mentioned currently online business objects, the bound accounts corresponding to business users who exit the cloud game application within the target time period), and scene-switching business objects (i.e., among the above-mentioned currently online business objects, the bound accounts corresponding to business users who switch the application activity scene within the target time period. These scene-switching business objects are in different application activity scenes in the current and target time periods. For example, the current application activity scene is the game home lobby scene, but the application activity scene in the target time period is the single-player competitive scene). Taking the new business objects within the target time period including business object 11 and business object 12 (neither business object 11 nor business object 12 belongs to the currently online business objects 1 - 10), the offline business objects including business object 7 and business object 8 among the currently online business objects 1 - 10, and the scene-switching objects including business object 9 among the currently online business objects 1 - 10 as an example, usually, when a new business object starts running the cloud game application through a terminal device, the scene displayed on the terminal device is usually the game home lobby scene. That is to say, when business object 11 and business object 12 log in to the cloud game application, the initial login scene is the game home lobby scene; at this time, the total occupied computing power resources corresponding to this game home lobby scene (i.e., 356) and business object 11 and business object 12 can be used to determine the first changed computing power resource information corresponding to the new business objects, and this first changed computing power resource information is the newly demanded computing power resource information corresponding to business object 11 and business object 12.

[0103] The specific method for determining the first changed computing power resource information can be as follows: First, the total number of objects of business objects 1 - 5 (i.e., 5) can be obtained. Through the total occupied computing power resources corresponding to the game home lobby scene (i.e., 356) and this total number of objects 5, the average computing power demand information of this game home lobby scene (i.e., 356 / 5 = 71.2) can be determined. This average computing power demand information refers to the computing power resource information required by an average of each business object among business objects 1 - 5 in this game home lobby scene; subsequently, the new number of new business objects (including business object 11 and business object 12) (i.e., 2) can be obtained, and this new number can be multiplied by the average computing power demand information 71.2 of the game home lobby scene and the new number 2, thereby obtaining the first changed computing power resource information as 142.4.

[0104] It should be understood that the management server 100 can obtain the application activity scenarios in which the service objects 7 and 8 are located respectively when they go offline (as can be seen from the above, both are in the single-player competitive scenario). After the service objects 7 and 8 go offline, the computing server 11b no longer needs to provide computing power resources for them, so the computing server 11b can release the corresponding computing power resources. At this time, the management server 100 can determine the second changed computing power resource information corresponding to the offline service object, and this second changed computing power resource information is the computing power resource information that should be released corresponding to the service objects 7 and 8.

[0105] The specific method for determining the second changed computing power resource information can be as follows: First, the total number of objects of the service objects 6 - 8 (i.e., 3) can be obtained. Through the total occupied computing power resources corresponding to the single-player competitive scenario (i.e., 319) and this total number of objects 3, the average computing power demand information of this single-player competitive scenario (i.e., 319 / 3 = 106.3) can be determined. This average computing power demand information refers to the average computing power resource information required by each of the service objects 6 - 8 in this single-player competitive scenario; Subsequently, the offline quantity of the offline service objects (including the service objects 7 and 8) (i.e., 2) can be obtained, and this offline quantity can be multiplied by the average computing power demand information 106.3 of the single-player competitive scenario and the offline quantity 2. Thus, the second changed computing power resource information can be obtained as 212.6.

[0106] It should be understood that the management server 100 can obtain the application activity scenario in which the service object 9 is located when a scenario switching behavior occurs (as can be seen from the above, it is a multi-player competitive scenario), and the application activity scenario in which it is located after the scenario switching behavior occurs (for example, the scenario after switching is the game home page hall scenario). Because different application activity scenarios require different computing power resources, after the service object 9 performs scenario switching, the computing power resources required by the service object 9 will also change. At this time, the management server 100 can determine the third changed computing power resource information corresponding to the service object 9. This third changed computing power resource information is the changed computing power resource information corresponding after the scenario switching (this changed computing power resource information can be positive or negative. When this changed computing power resource information is positive, it can indicate that the computing server 11b should add computing power resources after the scenario switching; when this changed computing power resource information is negative, it can indicate that the computing server 11b should release computing power resources after the scenario switching).

[0107] The specific method for determining the third variable computing power resource information can be as follows: First, the total number of objects of service object 9 - service object 10 (i.e., 2) can be obtained. Through the total occupied computing power resources corresponding to the multi-player competitive scenario (i.e., 408), and the total number of objects 2, the average computing power demand information of the multi-player competitive scenario (i.e., 408 / 2 = 204) can be determined. This average computing power demand information refers to the average computing power resource information required by each service object in service object 9 - service object 10 in the multi-player competitive scenario. Subsequently, the switching quantity of the scenario switching service object (including service object 9) (i.e., 1) can be obtained, and this switching quantity can be multiplied by the average computing power demand information 204 of the multi-player competitive scenario and the switching quantity 1. Thus, the computing power resources that should be released in the multi-player competitive scenario can be obtained as 204. Subsequently, as known above, the average computing power demand information of the game home page hall scenario is 71.2. Then, the management server 100 can multiply the average computing power demand information 71.2 of the game home page hall scenario by the switching quantity 1. Thus, the computing power resources that should be increased in the game home page hall scenario can be obtained as 71.2. As known above, because the computing power resources that should be released after the service object 9 switches scenarios are greater than the computing power resources that should be increased, the total variable computing power resource information (i.e., the third variable computing power resource information) corresponding to the service object 9 should be -132.8 (i.e., 71.2 - 204).

[0108] Further, the management server 100 can add the occupied computing power resource information (1083) determined above to the first variable computing power resource information (i.e., 142.4) and the third variable computing power resource information (i.e., -132.8), and subtract the second variable computing power resource information (i.e., 212.6) from the result obtained by the addition (1092.6). The result obtained (i.e., 880) can be determined as the expected demand computing power resource information required by the cloud game application for the computing server 11b during the target time period.

[0109] Further, the management server 100 can obtain the current operating frequency of the computing server 11b. According to the current operating frequency, the current maximum computing power resource information of the computing server 11b can be determined. The management server 100 can determine the target operating frequency of the computing server 11b during the target time period based on the maximum computing power resource information corresponding to the current operating frequency, the above-mentioned expected demand computing power resource information, and the occupied computing power resource information. Among them, the maximum computing power resource information corresponding to the target operating frequency satisfies the expected demand computing power resource information. For the specific implementation method of determining the target operating frequency, reference can be made to the description in the subsequent Figure 4 corresponding embodiments.

[0110] After determining the target operating frequency, the operating frequency of the computing server 11b within the target time period can be adjusted from the current operating frequency to the target operating frequency. By dynamically adjusting the operating frequency of the computing server, the maximum computing power resource information corresponding to the target operating frequency of the computing server 11b within the target time period can meet the expected required computing power resource information, thereby reducing the situation where the operating frequency of the computing server is too high or too low within the target time period, resulting in excessive power consumption of the computing server 11b or the inability to meet the computing power resource requirements of cloud game applications.

[0111] It should be noted that the numerical values of the above-mentioned various computing power resources (such as 71, 70, 110, 108, etc.) are all illustrative examples for easy understanding and do not have practical reference significance.

[0112] Furthermore, please refer to Figure 3 , Figure 3 which is a schematic flowchart of a data processing method based on edge computing provided by an embodiment of the present application. Among them, this method can be executed by the computer device in the corresponding embodiment of the above Figure 1 , that is, it can be executed by the management server 100 in Figure 1 , or it can be executed by the computing servers in the edge node cluster in Figure 1 (including the computing server 11a, the computing server 11b, the computing server 12a, and the computing server 12b). As shown in Figure 3 , this data processing method based on edge computing may include the following steps S101 - step S103:

[0113] Step S101, obtain the occupied computing power resource information of the target application for the edge computing node; the occupied computing power resource information is the computing power resource information of the edge computing node occupied when running the target application.

[0114] In the present application, the edge computing node may refer to a computer device that can provide computing or application services, such as a server (for example, the computing server 11a, the computing server 11b, the computing server 12a, or the computing server 12b shown in the above Figure 1 ). The target application may refer to an application for which the edge computing node needs to complete relevant computing tasks. For example, the target application may be a cloud game application. Based on cloud computing technology, cloud games usually run on a remote server, and the terminal device only needs to receive the audio and video stream sent by the remote server and then decode and play it. At this time, the remote server can be the edge computing node.

[0115] It should be understood that when the target application runs on the terminal device, the edge computing node can provide corresponding computing services for it. When the edge computing node provides corresponding computing services for it, the target application will occupy the computing power resources of the edge computing node. Among them, the computing power can refer to the computing ability of the edge computing node. In this application, the computing power of the edge computing node is usually measured by CPU computing power and GPU computing power. Among them, the CPU computing power is generally measured by the number of operations per second (OPS); while the GPU computing power can have multiple measurement indicators according to the type of calculation. Generally, it is measured from two indicators: computing ability (according to the type of operation, measured by the number of floating-point operations per second (FLOPS), OPS, half-precision peak computing ability, and double-precision peak computing ability) and data read throughput. For computing power resources, in addition to CPU computing power resources and CPU computing power resources, of course, other computing power resources can also be included, such as memory resources, network bandwidth resources, disk resources, and so on. This application does not limit the content included in the computing power resources. The following will take the computing power resources including CPU computing power resources and GPU computing power resources as an example for illustration.

[0116] Specifically, the target application can include one or more application activity scenarios. The application activity scenario can refer to the scenario type to which the screen displayed on the terminal device belongs when the target application runs on the terminal device. An application activity scenario can provide corresponding application functions for the target application. For example, when the target application is a social application, the application activity scenarios can include voice call scenarios, video call scenarios, and text chat scenarios. When the business user logs in to the social application through the bound account of the social application (which can be called the business object), in the voice call scenario, the social application provides the voice communication function between different business objects (that is, business users can conduct voice communication); in the video call scenario, the social application provides the video communication function between different business objects (that is, business users can conduct video communication); in the text chat scenario, the social application provides the text communication function between different business objects (that is, business users can conduct text communication). For example, when the target application is a cloud game application, the application activity scenarios can include the game home page lobby scenario, single-player competition scenario, and multi-player competition scenario. In the game home page lobby scenario, the cloud game application provides functions such as game background introduction, game character introduction, data display, and competition mode selection for the business object; in the single-player competition scenario, the cloud game application provides the single-player combat competition function for the business object; in the multi-player competition scenario, the cloud game application provides the multi-player combat competition function, multi-player communication function, and so on for the business object.

[0117] It can be understood that when the application activity scenarios of business objects in the target application are different, the required computing power resources are usually different. Therefore, this application can obtain the application activity scenarios of online business objects in the target application, and obtain the computing power resources occupied by each online business object in the application activity scenarios where they are located. Thus, the computing power resources of the edge computing nodes occupied by all online business objects can be statistically obtained. The computing power resources of the edge computing nodes occupied by all online business objects in total can be called the occupied computing power resource information. The specific method for determining the occupied computing power resource information of the edge computing nodes can be as follows: N current running information of the target application can be obtained; among them, one current running information can include the application activity scenario where an online business object is located, and the business object occupancy resource information in the application activity scenario; obtain the business object occupancy resource information corresponding to each current running information to obtain N business object occupancy resource information; determine the sum of the N business object occupancy resource information as the occupied computing power resource information of the target application for the edge computing node.

[0118] It should be understood that the above-mentioned business object occupancy resource information can be understood as the computing power resources of the edge computing nodes occupied when an online business object is in a certain application activity scenario; after each business user logs in to the target application through a bound account (hereinafter referred to as a business object), the login status of its business object in the target application is the online state. At this time, the business object can also be called an online business object. The computer device will obtain the application activity scenarios of each online business object in the target application in real time, and statistically calculate the computing power resources occupied by the online business object when it is in a certain application activity scenario. The application activity scenario where an online business object is currently located and the computing power resources occupied when it is in the application activity scenario can form a current running information. Because one current running information includes the application activity scenario where an online business object is located and the computing power resources occupied when it is in this application activity scenario; then N current running information can include the application activity scenarios where N online business objects are respectively located and the computing power resources respectively occupied by the N online business objects when they are in the corresponding application activity scenarios. Adding up and summing these N business object occupancy resource information (that is, the computing power resources respectively occupied by N online business objects when they are in the corresponding application activity scenarios), the total computing power resources of the edge computing nodes occupied by N online business objects (that is, the occupied computing power resource information) can be obtained.

[0119] It should be understood that the computing power resources include CPU computing power resources and GPU computing power resources. Then, when statistically calculating the occupied computing power resource information, the CPU computing power resources and GPU computing power resources of the edge computing nodes occupied by N online business objects can be statistically calculated. The occupied CPU computing power resources and GPU computing power resources can both be called the occupied computing power resource information.

[0120] Step S102, predict the newly added service objects, offline service objects, and scenario switching service objects for the target application within the target time period, and determine the expected required computing power resource information for the target application with respect to the edge computing node within the target time period according to the newly added service objects, offline service objects, and scenario switching service objects.

[0121] In this application, the newly added service object may refer to the bound account (service object) of the service user who starts the target application within the target time period, and the edge computing node can provide functional services for it; the offline service object may refer to the service object corresponding to the service user who will close the target application among the service users corresponding to the above N online service objects; the scenario switching service object may refer to the service object among the above N online service objects that will switch the application activity scenario within the target time period (for example, for a cloud game application, an online service object is in a single-player competitive scenario at the current moment and will exit the single-player competitive scenario and enter the game home page lobby scenario within the target time period, and this online service object can be called a scenario switching object). This application can predict the newly added service objects, offline service objects, and scenario switching objects within the target time period, and determine the expected required computing power resource information within the target time period according to the newly added service objects, offline service objects, and scenario switching objects.

[0122] Among them, the specific method for predicting the newly added service objects, offline service objects, and scenario switching objects within the target time period may be: the computer device will obtain the historical service behavior data of the target application, and then will obtain the operation activity information of the target application within the target time period; according to the operation activity information and historical service behavior data within the target time period, the newly added service objects for the target application by the edge computing node within the target time period can be determined; according to the application activity scenario where the service object is located, the operation activity information within the target time period, and the historical service behavior data, the offline service objects and scenario switching service objects for the target application by the edge computing node within the target time period can be determined. Among them, the historical service behavior data may include the relevant behavior data of the historical online service objects, historical online service objects, historical offline service objects, and historical scenario switching service objects at each time node within the historical time period of the target application, etc. Among them, the relevant behavior data may include the application activity scenario where it is located, operation behavior, application running duration, etc.

[0123] It should be understood that the operation activity information within the target time period may refer to special activities launched in the target application during holidays (such as May Day, Qixi Festival, Dragon Boat Festival), specific festivals, version release days, etc. For example, taking the target application as a cloud game application, when it is the Dragon Boat Festival, new Dragon Boat Festival limited-time activities (such as a game character race activity) and a limited-time purchase activity for character costumes will be launched in the cloud game application. There is a high probability that business objects will choose to log in to the cloud game application during the target time period to participate in this Dragon Boat Festival limited-time activity, purchase character costumes, etc. Those business objects that did not log in to the cloud game application (i.e., not among the N online business objects) but are predicted to log in to the cloud game application during the target time period can be referred to as new business objects. At the same time, the computer device can, based on the operation activity information and historical business behavior data, predict the business objects among the N online business objects that will close and run the target application during the target time period. These business objects can be referred to as offline business objects; the computer device can also, based on the operation activity information and historical business behavior data, predict the business objects among the N online business objects that will switch scenarios. These business objects can be referred to as scenario-switching business objects.

[0124] Optionally, the solution provided in the embodiments of the present application may involve the machine learning technology of artificial intelligence. Machine learning (ML) is a multi-disciplinary subject that involves multiple disciplines such as probability theory, statistics, approximation theory, convex analysis, and algorithm complexity theory. It specifically studies how computers simulate or implement human learning behaviors to acquire new knowledge or skills and reorganize the existing knowledge structure to continuously improve their own performance. Machine learning is the core of artificial intelligence and the fundamental way to make computers intelligent, and its applications cover all fields of artificial intelligence. Machine learning and deep learning usually include technologies such as artificial neural networks, belief networks, reinforcement learning, transfer learning, inductive learning, and rote learning. It is specifically illustrated through the following embodiments: When predicting the offline business users of the edge computing node for the target application during the target time period, the computer device can generate the application activity scenario where the business user is located, the operation activity information corresponding to the target application, and the offline behavior characteristics corresponding to the offline behavior of the business user according to the offline prediction model, and then output the predicted offline label corresponding to the offline behavior characteristics in the offline prediction model, and then determine the offline business users according to the predicted offline label. Among them, the offline prediction model is a machine learning model trained based on historical business behavior data and is used to simulate the offline behaviors of business users at different time periods and infer the behavioral states such as what application activity scenarios and time nodes the users will be in when they go offline. Correspondingly, the computer device can also predict the scenario-switching business users through a corresponding machine learning model.

[0125] Further, based on the newly added service objects, the offline service objects, and the scenario-switching service objects, the expected required computing power resource information of the target application for the edge computing nodes within the target time period can be determined. The following will take the number of application activity scenarios included in the N current running information being one or more, and one or more application activity scenarios including application activity scenario M i (i being a positive integer) as an example to illustrate the specific method for determining the expected required computing power resource information. The specific method can be: Among the N online service objects, the online service objects in the application activity scenario being application activity scenario M i can be determined as the service objects to be counted; subsequently, the total occupied resource information of the service objects to be counted in application activity scenario M i can be counted; subsequently, the corresponding object quantity of the service objects to be counted can be obtained. Based on the total occupied resource information of the service objects and the object quantity, the average computing power demand information corresponding to application activity scenario M i can be determined; when the average computing power demand information corresponding to one or more application activity scenarios is determined respectively, the expected required computing power resource information of the target application for the edge computing nodes within the target time period can be determined based on the average computing power demand information corresponding to one or more application activity scenarios respectively, the newly added service objects, the offline service objects, and the scenario-switching service objects.

[0126] Among them, the specific method for determining the expected required computing power resource information of the target application for the edge computing nodes within the target time period based on the average computing power demand information corresponding to one or more application activity scenarios respectively, the newly added service objects, the offline service objects, and the scenario-switching service objects can be: The newly added quantity corresponding to the newly added service objects, the offline quantity corresponding to the offline service objects, and the switching quantity corresponding to the scenario-switching objects can be obtained; the initial login scenario corresponding to the newly added service objects can be predicted; among them, one or more application activity scenarios include the initial login scenario; subsequently, the application activity scenario where the offline service objects are located when the offline behavior occurs can be determined as the offline application activity scenario; subsequently, the application activity scenario where the scenario-switching objects are located before the scenario switching can be determined as the initial application activity scenario, and the application activity scenario where the scenario-switching objects are located after the scenario switching can be determined as the target application activity scenario; based on the average computing power demand information corresponding to the initial login scenario, the offline application activity scenario, the initial application activity scenario, and the target application activity scenario respectively, as well as the switching quantity, the offline quantity, and the newly added quantity, the expected required computing power resource information of the target application for the edge computing nodes within the target time period can be determined.

[0127] Among them, according to the average computing power demand information corresponding to the initial login scenario, the offline application activity scenario, the initial application activity scenario, and the target application activity scenario, as well as the switching quantity, the offline quantity, and the new quantity, the specific method for determining the expected required computing power resource information of the target application for the edge computing node within the target time period can be as follows: The first changed computing power resource information corresponding to the new business object can be determined according to the new quantity and the average computing power demand information corresponding to the initial login scenario; the second changed computing power resource information corresponding to the offline business object can be determined according to the offline quantity and the average computing power demand information corresponding to the online application activity scenario; the third changed computing power resource information corresponding to the scenario switching object can be determined according to the average computing power demand information corresponding to the initial application activity scenario, the average computing power demand information corresponding to the target application activity scenario, and the switching quantity; according to the first changed computing power resource information, the second changed computing power resource information, the third changed computing power resource information, and the occupied computing power resource information of the edge computing node, the expected required computing power resource information of the target application for the edge computing node within the target time period can be determined.

[0128] It should be understood that when a business object logs in to or opens the target application, it usually enters a default interface, and the scenario corresponding to this default interface can be called the initial login scenario; among them, one or more application activity scenarios included in the target application include this initial login scenario. For example, when a business object logs in to a cloud game application, it usually enters the game home page hall scenario, and this game home page hall scenario can be called the initial login scenario. The computer device can obtain the average computing power demand information corresponding to the initial login scenario (that is, in the initial login scenario, the computing power resources required by each online business object on average), and according to the new quantity of the new business object and the average computing power demand information corresponding to this initial login scenario (for example, multiplying the new quantity by the average computing power demand information corresponding to the initial login scenario), the new computing power resource information corresponding to the new business object (such as the product result of the new quantity and the average computing power demand information corresponding to the initial login scenario) can be determined. This new computing power resource information is the computing power resource that needs to be increased when the edge computing node needs to provide functional services for the new business object; this new computing power resource information can be called the first changed computing power resource information.

[0129] It should be understood that the computer device can also obtain the application activity scenario (hereinafter referred to as the offline application activity scenario) in which the offline service object is located when the offline behavior occurs (such as the behavior of exiting or closing the target application). According to the average computing power demand information corresponding to the offline application activity scenario (that is, in the offline application activity scenario, the computing power resources required by each online service object on average) and the offline quantity of the offline service object, the released computing power resource information corresponding to the offline service object can be determined. That is to say, when these offline service users close the target application, the edge computing node no longer needs to provide functional services for them and can release the corresponding computing power resources. These released computing power resources can be called the second variable computing power resource information. The second variable computing power resource information can be the product of the offline quantity and the average computing power demand information corresponding to the offline application activity scenario.

[0130] It should be understood that since the computing power resources required by different application activity scenarios are usually different, when the service object switches scenarios, the required computing power resources will also change (maybe increase or maybe decrease). Then, the computer device can predict the service objects among the N online service objects that will switch scenarios during the target time period, and predict the application activity scenario (which can be called the initial application activity scenario) in which these scenario-switching objects are located before the scenario switch, and the application activity scenario (which can be called the target application activity scenario) in which they are located after the scenario switch; for the initial application activity scenario, the edge computing node no longer needs to provide the corresponding computing power resources for these scenario-switching users and should release the corresponding computing power resources; for the target application activity scenario, the edge computing node needs to provide the corresponding computing power resources for it and should add new computing power resources. Then, the average computing power demand information corresponding to the initial application activity scenario (in the initial application activity scenario, the computing power resources required by each online service object on average) and the average computing power demand information corresponding to the target application activity scenario can be obtained. The product of the average computing power demand information corresponding to the initial application activity scenario and the switching quantity of the scenario-switching user can be determined as the computing power resources that should be released corresponding to the initial application activity scenario; the product of the average computing power demand information corresponding to the target application activity scenario (in the target application activity scenario, the computing power resources required by each online service object on average) and the switching quantity of the scenario-switching user can be determined as the computing power resources that should be newly added corresponding to the target application activity scenario. The computing power resources that should be released corresponding to the initial application activity scenario and the computing power resources that should be newly added corresponding to the target application activity scenario can be added together, and the result can be used as the third variable computing power resource information corresponding to the scenario-switching user (which may be a positive value or a negative value).

[0131] Further, the above-mentioned occupied computing power resource information can be added to the newly added computing power resources, and the result obtained by the addition is then subtracted by the computing power resources that should be released, so as to obtain the final desired required computing power resource information. That is, the occupied computing power resource information, the first variable computing power resource information, and the third variable computing power resource information can be added, and then the second variable computing power resource information is subtracted from the result of the addition, thereby obtaining the desired required computing power resource information of the target application for the edge computing node during the target time period.

[0132] Step S103: Obtain the current operating frequency of the edge computing node, and determine the target operating frequency of the edge computing node during the target time period according to the maximum computing power resource information, the occupied computing power resource information, and the desired required computing power resource information corresponding to the current operating frequency; the maximum computing power resource information corresponding to the target operating frequency meets the desired required computing power resource information.

[0133] In this application, the computer device can obtain the current operating frequency of the edge computing node. Subsequently, the maximum computing power resource information corresponding to the current operating frequency can be obtained according to the current operating frequency. According to the maximum computing power resource information, the occupied computing power resource information, and the desired required computing power resource information corresponding to the current operating frequency, the target operating frequency of the edge computing node during the target time period can be determined. Among them, determining the target operating frequency of the edge computing node is to judge whether to increase, decrease, or keep the current operating frequency unchanged during the target time period. For ease of understanding, please also refer to Figure 4 , Figure 4 which is a logic flowchart for adjusting the operating frequency of an edge computing node provided by an embodiment of this application. As Figure 4 shown, this logic process may include at least the following steps S41 - step S49:

[0134] Step S41: Judge whether the computing power of the edge computing node is fully loaded.

[0135] Specifically, the occupied computing power resource information of the edge computing node can be compared with the maximum computing power resource information corresponding to the current operating frequency. If the occupied computing power resource information is less than the maximum computing power resource information corresponding to the current operating frequency, and the absolute value of the difference between the occupied computing power resource information and the maximum computing power resource information corresponding to the current operating frequency is greater than a threshold value (which can be set manually), it can be determined that the computing power of the edge computing node is not fully loaded and there are redundant idle computing power resources. If the occupied computing power is less than the maximum computing power resource information corresponding to the current operating frequency, but the absolute value of the difference between the occupied computing power resource information and the maximum computing power resource information corresponding to the current operating frequency is less than the threshold value, it can be determined that the computing power of the edge computing node is in a fully loaded state and there are no redundant idle computing power resources. Optionally, if the occupied computing power resource is equal to the maximum computing power resource information corresponding to the current operating frequency, it can also be determined that the computing power of the edge computing node is in a fully loaded state and there are no redundant idle computing power resources.

[0136] If it is determined that the computing power of the edge computing node is in a fully loaded state, the subsequent step S42 can be executed. If it is determined that the computing power of the edge computing node is not in a fully loaded state (i.e., in an unloaded state), the subsequent step S45 can be executed.

[0137] Step S42: Determine whether the expected computing power increases.

[0138] Specifically, the occupied computing power resource information can be compared with the expected required computing power resource information, so as to determine whether the expected required computing power resource information has increased, decreased (or remained unchanged) compared with the occupied computing power resource information. If the expected required computing power resource information shows an increasing situation, the subsequent step S44 can be executed. If the expected required computing power resource information shows a non-increasing situation, the subsequent step S43 can be executed.

[0139] Step S43: Keep the current operating frequency unchanged.

[0140] Specifically, when the computing power of the edge computing node is in a fully loaded state and the expected required computing power resource information shows a non-increasing situation, it indicates that the computing power resources provided by the edge computing node at the current operating frequency are sufficient to meet the expected required computing power resources. To reduce the number of adjustments to the operating frequency and thus reduce the losses caused by the adjustment frequency, the operating frequency can be kept unchanged at this time, and this unchanged current operating frequency is the target operating frequency within the target time period.

[0141] Optionally, the current operating frequency can also be appropriately reduced (for example, reduced to the operating frequency corresponding to the expected required computing power resources).

[0142] Step S44: Increase the operating frequency.

[0143] Specifically, when the computing power of the edge computing node is in a fully loaded state and the expected demand computing power resource information shows a growth trend, it indicates that the computing power resources provided by the edge computing node at the current operating frequency are insufficient to meet the expected demand computing power resources. In this case, the operating frequency can be appropriately increased (for example, increased to the operating frequency corresponding to the expected demand computing power resources). The increased operating frequency is the target operating frequency within the target time period.

[0144] Step S45: Determine whether the expected computing power is growing.

[0145] Specifically, when it is determined that the computing power of the edge computing node is not in a fully loaded state, the occupied computing power resource information can also be compared with the expected demand computing power resource information, so as to determine whether the expected demand computing power resource information has increased, decreased (or remained unchanged) compared with the occupied computing power resource information. If the expected demand computing power resource information shows a growth trend, the subsequent step S46 can be executed; if the expected demand computing power resource information does not show a growth trend, the subsequent step S48 can be executed.

[0146] Step S46: Determine whether the idle computing power meets the expected computing power.

[0147] Specifically, when it is determined that the computing power of the edge computing node is not in a fully loaded state and the expected demand computing power resource information is in a growth trend, the maximum computing power resource information corresponding to the current operating frequency can be subtracted from the occupied computing power resource information, so as to obtain the idle computing power resource information of the edge computing node. Subsequently, the resource sum of the above first variable computing power resource information, second variable computing power resource information, and third variable computing power resource information can be calculated. Here, the sum is the additional demand computing power resources required by the edge computing node. The idle computing power resource information can be compared with this resource sum. If the idle computing power resource information is greater than this resource sum, it can be determined that the idle computing power resource information can meet the additional demand computing power resources, and step S43 (i.e., keeping the current operating frequency unchanged) can be executed; if the idle computing power resource information is less than this resource sum, it can be determined that the idle computing power resource information is insufficient to meet the additional demand computing power resources, and the subsequent step S47 can be executed. Optionally, if the idle computing power resource information is equal to this resource sum, it can also be determined that the idle computing power resource information can meet the additional demand computing power resources, and step S43 can be executed.

[0148] Step S47: Increase the operating frequency.

[0149] Specifically, when the idle computing power resource information is insufficient to meet the additional demand computing power resources, the operating frequency can be appropriately increased (for example, increased to the operating frequency corresponding to the expected demand computing power resources within the target time period). The increased operating frequency is the target operating frequency within the target time period.

[0150] Step S48: Determine whether the power consumption changes after the frequency is reduced.

[0151] Specifically, when it is determined that the computing power of the edge computing node is not fully loaded and the expected demand computing power resource information is not increasing, the maximum computing power resource corresponding to the current operating frequency is sufficient to meet the expected demand computing power resource information at this time, and the operating frequency can be appropriately reduced. However, optionally, in order to reduce the number of adjustments to the operating frequency (to avoid frequent adjustment of the operating frequency) and thus reduce the problem of system operation instability caused by frequent frequency adjustments, before reducing the frequency in this application, it can first be determined whether the power consumption of the edge computing node changes (i.e., whether it will decrease) after the frequency is reduced. When the power consumption changes, the operating frequency is then reduced. When the power consumption does not change, the frequency can be kept unchanged. Although the operating frequency is not reduced at this time, since the situation of frequent adjustment of the operating frequency is avoided during the entire process of adjusting the frequency (it is not that the operating frequency must be reduced whenever the frequency needs to be reduced), the stability of the system operation can be improved in this way, which is also a way to reduce the operating cost.

[0152] However, since the reduction of the operating frequency and the reduction of the power consumption are not linearly related (i.e., it is not that the power consumption will definitely decrease when the operating frequency is reduced), due to different hardware implementations, when the operating frequency is slightly reduced, the power consumption may not decrease accordingly. Therefore, in order to accurately count the relationship between the frequencies and power consumptions of different models of hardware, the corresponding relationship between the operating frequency and the power consumption can be continuously corrected according to historical data and real-time operating conditions, so as to obtain a relationship mapping table of the computing power, frequency, and power consumption of the hardware (including CPU and GPU). The relationship mapping table can be as shown in Table 1 below.

[0153] Table 1

[0154]

[0155] Among them, the relationship mapping table shown in Table 1 may include the mapping relationship between the operating frequency of the CPU, the maximum computing power corresponding to the operating frequency, and the power consumption corresponding to the operating frequency; it may also include the mapping relationship between the operating frequency of the GPU, the maximum computing power corresponding to the operating frequency, and the power consumption corresponding to the operating frequency. When determining the expected demand computing power resource information within the target time period, the operating frequency corresponding to the expected demand computing power resource information can be obtained according to the relationship mapping table. When the computing power of the edge computing node is not fully loaded and the expected demand computing power resource does not increase, the operating frequency can be appropriately reduced at this time. However, before that, through the relationship mapping table, it can be determined whether the power consumption will also decrease when the frequency is reduced. If the power consumption corresponding to the expected demand computing power resource information does not decrease after the frequency is reduced compared with the power consumption corresponding to the current operating frequency, then the operating frequency can not be reduced at this time and can remain unchanged; continue to use the current operating frequency as the target operating frequency within the target time period.

[0156] Among them, if the power consumption corresponding to the expected demand computing power resource information changes (such as decreases) after the frequency is reduced compared with the power consumption corresponding to the current operating frequency, then the operating frequency can be reduced at this time (such as reducing the operating frequency to the operating frequency corresponding to the expected demand computing power resource through the relationship mapping table); the reduced operating frequency can be used as the target operating frequency within the target time period.

[0157] It should be understood that by judging whether there is a change in power consumption, the present application dynamically adjusts the operating frequency of the edge computing node, which can reduce the situation of frequently changing the operating frequency but without obvious benefits, and at the same time can improve the stability of system operation.

[0158] In the embodiments of the present application, since the main factors affecting the change of computing power resources of edge computing nodes are factors such as the online, offline, and online of objects, the present application can predict the newly added service objects, offline service objects, and scenario switching objects for the target application within the target time period. Subsequently, the expected required computing power resource information of the target application for the edge computing node within the target time period can be determined based on the newly added service objects, offline service objects, and scenario switching service objects. Thus, based on the expected required computing power resource information of the target time period, the occupied computing power resource information of the edge computing node, and the maximum computing power resource information corresponding to the current operating frequency of the edge computing node, it can be determined whether to increase, decrease, or keep the current operating frequency unchanged within the target time period (i.e., determine the target operating frequency within the target time period). Because the maximum computing power resource information of the edge computing node corresponds to the operating frequency of the edge computing node (for example, the greater the operating frequency, the greater the maximum computing power resource information), at the same time, there is also a corresponding relationship between the operating frequency and the operating consumption (i.e., power consumption) of the edge computing node. Then, by timely adjusting the operating frequency of the edge computing node (increasing, decreasing, or keeping it unchanged), the maximum computing power resource information corresponding to the target operating frequency within the target time period can meet the expected required computing power resource information. Thus, the problem of mismatch between the computing power resources provided by the edge computing node and the actual required computing power resources can be reduced (for example, the problem that the computing power resources provided by the edge computing node are too large, but the actual required computing power resources are very small). It can make the computing power resources provided by the edge computing node match the required computing power resources within the target time period. Thus, the operating consumption of the edge computing node will not be too large, and the operating cost can be reduced. That is, the present application can make the computing power resources provided by the edge computing node not be much greater than the expected required computing power resource information (i.e., reduce the operating cost) or much less than the expected required computing power resource information, resulting in unmet requirements, by adjusting the operating frequency of the edge node. Thus, the computing power demand and the operating cost can be dynamically balanced. That is, the present application can meet the computing power demand while reducing the operating cost of the edge computing node.

[0159] Further, please refer to Figure 5 , Figure 5 which is a schematic flowchart of a process for determining the target operating frequency within the target time period provided by the embodiments of the present application. This process can correspond to the process of determining the target operating frequency in step S103 in the corresponding embodiment of the above Figure 3 . As Figure 5 shown, this process can include the following steps S201 - step S203:

[0160] Step S201: Determine the current idle computing power resource information of the edge computing node based on the maximum computing power resource information corresponding to the current operating frequency and the occupied computing power resource information (i.e., the computing power resources of the edge computing node that are not occupied by the target application when the target application is running currently, that is, the unutilized computing power resource information in the edge computing node), and determine the operating state of the edge computing node according to the current idle computing power resource information.

[0161] Specifically, the maximum computing power resource information corresponding to the current operating frequency can be subtracted from the occupied computing power resource information (such as using the maximum computing power resource information corresponding to the current operating frequency to subtract the occupied computing power resource information), thereby obtaining the current idle computing power resource information of the edge computing node. According to the current idle computing power resource information, the operating state of the edge computing node can be determined. The specific method can be: match the current idle computing power resource information with the idle resource threshold; if the current idle computing power resource information is greater than the idle resource threshold, determine the operating state of the edge computing node as the non-full-load operating state; if the current idle computing power resource information is less than the idle resource threshold, determine the operating state of the edge computing node as the full-load operating state. Optionally, if the current idle computing power resource information is equal to the idle resource threshold, the operating state of the edge computing node can also be determined as the non-full-load operating state. Among them, the idle resource threshold here can correspond to the threshold in the above Figure 4 corresponding embodiment, and the non-full-load operating state can be understood as the above-mentioned non-full-load state of computing power; the full-load operating state can be understood as the above-mentioned full-load state of computing power. For its specific determination method, reference can be made to the description of step S41 in the above Figure 4 corresponding embodiment, and details will not be elaborated here.

[0162] Step S202: If the operating state of the edge computing node is the full-load operating state, determine the target operating frequency of the edge computing node within the target time period according to the occupied computing power resource information and the expected required computing power resource information.

[0163] Specifically, if the operating state of the edge computing node is a full-load operating state, the expected demand computing power resource information can be compared with the occupied computing power resource information; if the expected demand computing power resource information is greater than the occupied computing power resource information, the mapping table can be obtained; where the mapping table includes N mapping relationships; a mapping relationship includes a corresponding relationship between a configured operating frequency (the configured operating frequency can refer to the operating frequency in the mapping table) and a configured maximum computing power resource information (in the mapping table, an operating frequency corresponds to a maximum computing power resource information, the operating frequency in the mapping table can be called the configured operating frequency, and the maximum computing power resource information corresponding to each configured operating frequency can be called the configured maximum computing power resource information); in the mapping table, the configured operating frequency corresponding to the configured maximum computing power resource information greater than the expected demand computing power resource information can be determined as the target operating frequency (optionally, the configured operating frequency corresponding to the configured maximum computing power resource information equal to the expected demand computing power resource information can also be determined as the target operating frequency); subsequently, the operating frequency of the edge computing node can be adjusted from the current operating frequency to the target operating frequency. It should be noted that when the configured operating frequency corresponding to the configured maximum computing power resource information greater than the expected demand computing power resource information in the mapping table is determined as the target operating frequency, the configured maximum computing power resource information greater than the expected demand computing power resource information should be slightly greater than the expected demand computing power resource information (for example, the smallest one among the configured maximum computing power resource information greater than the expected demand computing power resource information in the mapping table), so as to avoid the computing power resources provided during the target time period being too large (that is, to avoid the edge computing node operating at too high an operating frequency during the target time period).

[0164] Optionally, it can be understood that after comparing the expected demand computing power resource information with the occupied computing power resource information, if the expected demand computing power resource information is less than the occupied computing power resource information, the current operating frequency can be determined as the target operating frequency. Optionally, after comparing the expected demand computing power resource information with the occupied computing power resource information, if the expected demand computing power resource information is equal to the occupied computing power resource information, the current operating frequency can also be determined as the target operating frequency at this time.

[0165] Step S203, if the operating state of the edge computing node is an under-full-load operating state and the expected demand computing power resource information is greater than the occupied computing power resource information, determine the target operating frequency of the edge computing node during the target time period according to the current idle computing power resource information, the occupied computing power resource information, and the expected demand computing power resource information.

[0166] Specifically, if the operating state of the edge computing node is an under-loaded operating state and the expected required computing power resource information is greater than the occupied computing power resource information, the absolute value of the resource difference between the expected required computing power resource information and the occupied computing power resource information can be determined; if the current idle computing power resource information is less than the absolute value of the resource difference, the mapping table can be obtained; the mapping table includes N mapping relationships; one mapping relationship includes the correspondence between a configured operating frequency and a configured maximum computing power resource information; in the mapping table, the configured operating frequency corresponding to the configured maximum computing power resource information greater than (or equal to) the expected required computing power resource information can be determined as the target operating frequency; the operating frequency of the edge computing node is adjusted from the current operating frequency to the target operating frequency. It should be noted that when the configured operating frequency corresponding to the configured maximum computing power resource information greater than the expected required computing power resource information in the mapping table is determined as the target operating frequency, the configured maximum computing power resource information greater than the expected required computing power resource information should be slightly greater than the expected required computing power resource information (for example, the smallest one among the configured maximum computing power resource information greater than the expected required computing power resource information in the mapping table), so as to avoid the computing power resources provided during the target time period being too large.

[0167] Optionally, it can be understood that after determining the absolute value of the resource difference between the expected required computing power resource information and the occupied computing power resource information, if the current idle computing power resource information is greater than or equal to the absolute value of the resource difference, the current operating frequency is determined as the target operating frequency.

[0168] Optionally, it can be understood that if the operating state of the edge computing node is an under-loaded operating state and the expected required computing power resource information is less than the occupied computing power resource information, the mapping table can be obtained; the mapping table includes N mapping relationships; one mapping relationship includes the correspondence between a configured operating frequency, a configured maximum computing power resource information, and a configured operating consumption (usually, when the edge computing node operates at a certain frequency, power consumption (power loss, usually in watts) will be generated, and different operating frequencies will generate different power consumptions. The operating consumption here can refer to power consumption, and the operating consumption corresponding to each configured operating frequency in the mapping table can be called the configured operating consumption); the configured operating consumption corresponding to the current operating frequency in the mapping table is determined as the configured operating consumption to be compared, and the target operating frequency of the edge computing node in the target time period is determined according to the configured operating consumption to be compared.

[0169] Among them, the specific method for determining the target operating frequency of the edge computing node within the target time period according to the operating consumption of the configuration to be compared may be as follows: In the mapping table, the configuration operating consumption corresponding to the maximum computing power resource information greater than or equal to the expected demand computing power resource information can be determined as the target operating consumption; the target operating consumption can be compared with the operating consumption of the configuration to be compared; if the target operating consumption is less than the operating consumption of the configuration to be compared, the configuration operating frequency corresponding to the target operating consumption is determined as the target operating frequency, and the operating frequency of the edge computing node is adjusted from the current operating frequency to the target operating frequency; if the target operating consumption is greater than or equal to the operating consumption of the configuration, the current operating frequency is determined as the target operating frequency. Among them, it should be noted that when the configuration operating frequency corresponding to the maximum computing power resource information greater than the expected demand computing power resource information in the mapping table is determined as the target operating frequency, the maximum computing power resource information greater than the expected demand computing power resource information here should be slightly greater than the expected demand computing power resource information (for example, the smallest one among the maximum computing power resource information greater than the expected demand computing power resource information in the mapping table), so as to avoid the computing power resources provided within the target time period being too large (that is, the operating frequency of the edge computing node within the target time period is too large).

[0170] Optionally, if the operating state of the edge computing node is not fully loaded and the expected demand computing power resource information is equal to the occupied computing power resource information, the current operating frequency can be kept unchanged (that is, the current operating frequency is used as the target operating frequency within the target time period).

[0171] Among them, the operating consumption can be understood as the power consumption of the edge computing node, and the mapping table can be understood as a relationship mapping table. For the specific implementation manners of steps S201 - S203, reference can be made to the descriptions of steps S41 - S49 in the corresponding embodiments above Figure 4 and will not be elaborated here.

[0172] Furthermore, please refer to Figure 6 , Figure 6 which is a system architecture diagram provided by an embodiment of the present application. As shown in Figure 6 , the system as a whole may include a management server and an edge computing node. Among them, the edge computing node may include a computing power information collection module, a frequency control module, a computing power guarantee module, and a power consumption information collection module. For ease of understanding, the functions corresponding to each module will be described below.

[0173] Computing power information collection module: It mainly includes CPU computing power information collection, GPU computing power information collection, and local information collection. Among them, CPU computing power information collection mainly collects the CPU model, current operating frequency, maximum computing power that can be achieved currently, current actual used computing power (i.e., occupied computing power), power consumption of the CPU itself, etc.; while GPU computing power information collection mainly collects the GPU model, current operating frequency, maximum computing power that can be achieved currently, current actual used computing power, and power consumption of the GPU itself; local information collection mainly collects the local model of the edge computing node and the current overall power consumption of the local machine.

[0174] Frequency control module: It is mainly used for CPU frequency control and GPU frequency control; among them, CPU frequency control mainly is: to increase or decrease the operating frequency of the CPU according to the requirements of the computing power guarantee module; GPU frequency control mainly is: to increase or decrease the operating frequency of the CPU according to the requirements of the computing power guarantee module.

[0175] Computing power guarantee module: According to the instructions of the management server and comprehensively considering the actual computing power situation of the local machine, it issues CPU and GPU frequency adjustment instructions to the frequency control module. The computing power guarantee module is mainly used for frequency adjustment and basic computing power guarantee. Among them, frequency adjustment mainly is: to increase or decrease the operating frequency of the CPU and GPU according to the instructions of the management server: when it is necessary to increase, it can be directly executed; when it is necessary to decrease, it can judge the change situation of the current CPU and GPU power consumption. If the power consumption does not decrease after reducing the frequency, the frequency reduction operation will not be performed, and the corresponding information will be notified to the management server. Among them, basic computing power guarantee mainly is: regularly check the load situation of the current CPU and GPU. If it has reached full load, the operating frequency will be increased in real time.

[0176] Power consumption information collection module: It is mainly used for CPU power consumption collection (collecting the real-time power consumption of the CPU), GPU power consumption collection (collecting the real-time power consumption of the GPU), overall machine power consumption collection (collecting the real-time power consumption of the overall machine), and computing power data. And after collection, it summarizes and reports to the management server.

[0177] The management server can include a data analysis module, a computing power prediction module, and a frequency adjustment strategy module. For the convenience of understanding, the functions corresponding to each module will be elaborated below.

[0178] Data analysis module: It is used to receive the data reported by the edge computing node and perform preliminary analysis for other modules to use.

[0179] Computing power prediction module: It is used to predict the change situation of computing power in the next stage according to new users and historical trends, that is, to predict the expected demand for computing power resources information in the next stage.

[0180] Frequency adjustment strategy module: Determine whether to increase or decrease the operating frequencies of the CPU and GPU according to the current computing power usage and the operating frequencies of the CPU and GPU.

[0181] Further, please refer to Figure 7 , Figure 7 which is a schematic structural diagram of a data processing device based on edge computing provided by an embodiment of the present application. The data processing device based on edge computing can be a computer program (including program code) running on a computer device. For example, the data processing device based on edge computing is an application software. The data processing device based on edge computing can be used to execute Figure 3 the method shown. As Figure 7 shown, the data processing device 1 based on edge computing may include: an occupied computing power acquisition module 11, an object prediction module 12, an expected computing power determination module 13, and a frequency determination module 14.

[0182] The occupied computing power acquisition module 11 is configured to acquire the occupied computing power resource information of the target application for the edge computing node. The occupied computing power resource information is the computing power resource information of the edge computing node occupied when the target application is running.

[0183] The object prediction module 12 is configured to predict new service objects, offline service objects, and scenario switching service objects for the target application within the target time period.

[0184] The expected computing power determination module 13 is configured to determine the expected required computing power resource information of the target application for the edge computing node within the target time period according to the new service objects, offline service objects, and scenario switching service objects.

[0185] The frequency determination module 14 is configured to acquire the current operating frequency of the edge computing node.

[0186] The frequency determination module 14 is further configured to determine the target operating frequency of the edge computing node within the target time period according to the maximum computing power resource information corresponding to the current operating frequency, the occupied computing power resource information, and the expected required computing power resource information. The maximum computing power resource information corresponding to the target operating frequency satisfies the expected required computing power resource information.

[0187] Among them, for the specific implementation manners of the occupied computing power acquisition module 11, the object prediction module 12, the expected computing power determination module 13, and the frequency determination module 14, reference can be made to the descriptions of steps S101 - S103 in the corresponding embodiments above, which will not be elaborated here. Figure 3 The description will not be repeated here.

[0188] In one embodiment, the frequency determination module 14 may include: an idle computing power determination unit 141, an operating state determination unit 142, and a target frequency determination unit 143.

[0189] The idle computing power determination unit 141 is configured to determine the current idle computing power resource information of the edge computing node according to the maximum computing power resource information corresponding to the current operating frequency and the occupied computing power resource information.

[0190] The operating state determination unit 142 is configured to determine the operating state of the edge computing node according to the current idle computing power resource information.

[0191] The target frequency determination unit 143 is configured to, if the operating state of the edge computing node is a full-load operating state, determine the target operating frequency of the edge computing node within the target time period according to the occupied computing power resource information and the desired demand computing power resource information.

[0192] The target frequency determination unit 143 is further configured to, if the operating state of the edge computing node is a non-full-load operating state and the desired demand computing power resource information is greater than the occupied computing power resource information, determine the target operating frequency of the edge computing node within the target time period according to the current idle computing power resource information, the occupied computing power resource information, and the desired demand computing power resource information.

[0193] Wherein, for the specific implementation manners of the idle computing power determination unit 141, the operating state determination unit 142, and the target frequency determination unit 143, reference may be made to the description of step S103 in the corresponding embodiment above, which will not be elaborated herein. Figure 3

[0194] In one embodiment, the operating state determination unit 142 may include: a matching subunit 1421 and a state determination subunit 1422.

[0195] The matching subunit 1421 is configured to match the current idle computing power resource information with the idle resource threshold.

[0196] The state determination subunit 1422 is configured to, if the current idle computing power resource information is greater than the idle resource threshold, determine the operating state of the edge computing node as a non-full-load operating state.

[0197] The state determination subunit 1422 is further configured to, if the current idle computing power resource information is less than the idle resource threshold, determine the operating state of the edge computing node as a full-load operating state.

[0198] Wherein, for the specific implementation manners of the matching subunit 1421 and the state determination subunit 1422, reference may be made to the description of step S103 in the corresponding embodiment above, which will not be elaborated herein. Figure 3 ​​

[0199] In one embodiment, the target frequency determination unit 143 may include: a computing power comparison subunit 1431, a frequency acquisition subunit 1432, and a first frequency adjustment subunit 1433.

[0200] The computing power comparison subunit 1431 is configured to compare the expected demand computing power resource information with the occupied computing power resource information if the operating state of the edge computing node is a full-load operating state;

[0201] The frequency acquisition subunit 1432 is configured to obtain a mapping table if the expected demand computing power resource information is greater than the occupied computing power resource information; the mapping table includes N mapping relationships; one mapping relationship includes a corresponding relationship between a configured operating frequency and a configured maximum computing power resource information;

[0202] The frequency acquisition subunit 1432 is further configured to determine, in the mapping table, the configured operating frequency corresponding to the configured maximum computing power resource information greater than the expected demand computing power resource information as the target operating frequency;

[0203] The first frequency adjustment subunit 1433 is configured to adjust the operating frequency of the edge computing node from the current operating frequency to the target operating frequency.

[0204] Among them, for the specific implementation manners of the computing power comparison subunit 1431, the frequency acquisition subunit 1432, and the first frequency adjustment subunit 1433, reference may be made to the description of step S103 in the corresponding embodiment above, and details will not be elaborated here. Figure 3 The description corresponding to the embodiment will not be repeated here.

[0205] In one embodiment, the target frequency determination unit 143 may further include: a first frequency determination subunit 1434.

[0206] The first frequency determination subunit 1434 is configured to determine the current operating frequency as the target operating frequency if the expected demand computing power resource information is less than the occupied computing power resource information.

[0207] In one embodiment, the target frequency determination unit 143 may include: a difference determination subunit 1435, a difference comparison subunit 1436, and a second frequency adjustment subunit 1437.

[0208] The difference determination subunit 1435 is configured to determine the absolute value of the resource difference between the expected demand computing power resource information and the occupied computing power resource information if the operating state of the edge computing node is not a full-load operating state and the expected demand computing power resource information is greater than the occupied computing power resource information;

[0209] The difference comparison subunit 1436 is configured to obtain a mapping table if the current idle computing power resource information is less than the absolute value of the resource difference; the mapping table includes N mapping relationships; one mapping relationship includes a corresponding relationship between a configured operating frequency and a configured maximum computing power resource information;

[0210] The difference comparison subunit 1436 is further configured to, in the mapping table, determine the configured operating frequency corresponding to the configured maximum computing power resource information greater than the expected required computing power resource information as the target operating frequency;

[0211] The second frequency adjustment subunit 1437 is configured to adjust the operating frequency of the edge computing node from the current operating frequency to the target operating frequency.

[0212] Among them, for the specific implementation manners of the difference determination subunit 1435, the difference comparison subunit 1436, and the second frequency adjustment subunit 1437, reference may be made to the description of step S103 in the corresponding embodiment above, which will not be elaborated here. Figure 3 The description of step S103 in the corresponding embodiment above will not be elaborated here.

[0213] In one embodiment, the target frequency determination unit 143 may further include: a second frequency determination subunit 1438.

[0214] The second frequency determination subunit 1438 is further configured to, if the current idle computing power resource information is greater than the absolute value of the resource difference, determine the current operating frequency as the target operating frequency.

[0215] In one embodiment, the frequency determination module 14 may further include: a table acquisition unit 144 and a consumption acquisition unit 145.

[0216] The table acquisition unit 144 is configured to obtain a mapping table if the operating state of the edge computing node is not fully loaded and the expected required computing power resource information is less than the occupied computing power resource information; the mapping table includes N mapping relationships; one mapping relationship includes a corresponding relationship between a configured operating frequency, a configured maximum computing power resource information, and a configured operating consumption;

[0217] The consumption acquisition unit 145 is configured to determine the configured operating consumption corresponding to the current operating frequency in the mapping table as the to-be-compared configured operating consumption;

[0218] The consumption acquisition unit 145 is further configured to determine the target operating frequency of the edge computing node within the target time period according to the to-be-compared configured operating consumption.

[0219] Among them, for the specific implementation manners of the table acquisition unit 144 and the consumption acquisition unit 145, reference may be made to the description of step S103 in the corresponding embodiment above, which will not be elaborated here. Figure 3 The description of step S103 in the corresponding embodiment above will not be elaborated here.

[0220] In one embodiment, the consumption acquisition unit 145 is further specifically configured to, in the mapping table, determine the configured running consumption corresponding to the maximum configured computing power resource information greater than the desired demand computing power resource information as the target running consumption;

[0221] The consumption acquisition unit 145 is further specifically configured to compare the target running consumption with the configured running consumption to be compared;

[0222] The consumption acquisition unit 145 is further specifically configured to, if the target running consumption is less than the configured running consumption to be compared, determine the configured running frequency corresponding to the target running consumption as the target running frequency, and adjust the running frequency of the edge computing node from the current running frequency to the target running frequency;

[0223] The consumption acquisition unit 145 is further specifically configured to, if the target running consumption is greater than the configured running consumption, determine the current running frequency as the target running frequency.

[0224] In one embodiment, the used computing power acquisition module 11 may include: an operation information acquisition unit 111 and a used resource statistics unit 112.

[0225] The operation information acquisition unit 111 is configured to acquire N current operation information of the target application; one current operation information includes an application activity scenario where an online business object is located, and the resource occupancy information of the business object in the application activity scenario;

[0226] The used resource statistics unit 112 is configured to acquire the resource occupancy information of the business object corresponding to each current operation information, and obtain N pieces of resource occupancy information of the business object;

[0227] The used resource statistics unit 112 is further configured to determine the sum of the N pieces of resource occupancy information of the business object as the occupied computing power resource information of the target application for the edge computing node.

[0228] Among them, for the specific implementation manners of the operation information acquisition unit 111 and the used resource statistics unit 112, reference may be made to the description of step S101 in the corresponding embodiment above, which will not be elaborated here. Figure 3 The description of step S101 in the corresponding embodiment above will not be elaborated here.

[0229] In one embodiment, the number of application activity scenarios included in the N current operation information is one or more; the one or more application activity scenarios include application activity scenario M i ; i is a positive integer;

[0230] The desired computing power determination module 13 may include: an object computing power statistics unit 131, an average computing power determination unit 132, and a desired computing power determination unit 133.

[0231] The object computing power statistics unit 131 is used to determine the online service objects to be statistically analyzed among N online service objects where the application activity scenario is application activity scenario M i as the service objects to be statistically analyzed;

[0232] The object computing power statistics unit 131 is also used to statistically analyze the total resource occupancy information of the service objects to be statistically analyzed in application activity scenario M i ;

[0233] The average computing power determination unit 132 is used to obtain the number of objects corresponding to the service objects to be statistically analyzed, and determine the average computing power demand information corresponding to application activity scenario M i according to the total resource occupancy information of the service objects and the number of objects;

[0234] The expected computing power determination unit 133 is used to, when determining the average computing power demand information corresponding to one or more application activity scenarios respectively, determine the expected required computing power resource information for the target application for the edge computing node within the target time period according to the average computing power demand information corresponding to one or more application activity scenarios respectively, newly added service objects, offline service objects, and scenario switching service objects.

[0235] Among them, for the specific implementation manners of the object computing power statistics unit 131, the average computing power determination unit 132, and the expected computing power determination unit 133, reference can be made to the description of step S102 in the corresponding embodiment above, which will not be elaborated here. Figure 3

[0236] In one embodiment, the expected computing power determination unit 133 is also specifically used to obtain the newly added quantity corresponding to the newly added service objects, the offline quantity corresponding to the offline service objects, and the switching quantity corresponding to the scenario switching objects;

[0237] The expected computing power determination unit 133 is also specifically used to predict the initial login scenario corresponding to the newly added service objects; one or more application activity scenarios include the initial login scenario;

[0238] The expected computing power determination unit 133 is also specifically used to determine the application activity scenario when the offline service objects perform the offline behavior as the offline application activity scenario;

[0239] The expected computing power determination unit 133 is also specifically used to determine the application activity scenario before the scenario switching objects perform scenario switching as the initial application activity scenario, and determine the application activity scenario after the scenario switching objects perform scenario switching as the target application activity scenario;

[0240] ​The expected computing power determination unit 133 is further specifically configured to determine the expected required computing power resource information of the target application for the edge computing node within the target time period according to the average computing power demand information corresponding to the initial login scenario, the offline application activity scenario, the initial application activity scenario, and the target application activity scenario, as well as the switching quantity, the offline quantity, and the new quantity.

[0241] In one embodiment, the expected computing power determination unit 133 is further specifically configured to determine the first changed computing power resource information corresponding to the new service object according to the new quantity and the average computing power demand information corresponding to the initial login scenario;

[0242] The expected computing power determination unit 133 is further specifically configured to determine the second changed computing power resource information corresponding to the offline service object according to the offline quantity and the average computing power demand information corresponding to the online application activity scenario;

[0243] The expected computing power determination unit 133 is further specifically configured to determine the third changed computing power resource information corresponding to the scenario switching object according to the average computing power demand information corresponding to the initial application activity scenario, the average computing power demand information corresponding to the target application activity scenario, and the switching quantity;

[0244] The expected computing power determination unit 133 is further specifically configured to determine the expected required computing power resource information of the target application for the edge computing node within the target time period according to the first changed computing power resource information, the second changed computing power resource information, the third changed computing power resource information, and the occupied computing power resource information of the edge computing node.

[0245] In the embodiments of the present application, since the main factors affecting the change of computing power resources of edge computing nodes are factors such as the online, offline, and online of objects, the present application can predict the newly added service objects, offline service objects, and scenario switching objects for the target application within the target time period. Subsequently, the expected required computing power resource information of the target application for the edge computing node within the target time period can be determined based on the newly added service objects, offline service objects, and scenario switching service objects. Thus, based on the expected required computing power resource information of the target time period, the occupied computing power resource information of the edge computing node, and the maximum computing power resource information corresponding to the current operating frequency of the edge computing node, it can be determined whether to increase, decrease, or keep the current operating frequency unchanged within the target time period (i.e., determine the target operating frequency within the target time period). Because the maximum computing power resource information of the edge computing node corresponds to the operating frequency of the edge computing node (for example, the greater the operating frequency, the greater the maximum computing power resource information), at the same time, there is also a corresponding relationship between the operating frequency and the operating consumption (i.e., power consumption) of the edge computing node. Then, by timely adjusting the operating frequency of the edge computing node (increasing, decreasing, or keeping it unchanged), the maximum computing power resource information corresponding to the target operating frequency within the target time period can meet the expected required computing power resource information. Thus, the problem of mismatch between the computing power resources provided by the edge computing node and the actual required computing power resources can be reduced (for example, the problem that the computing power resources provided by the edge computing node are too large, but the actual required computing power resources are very small). It can make the computing power resources provided by the edge computing node match the required computing power resources within the target time period. Thus, the operating consumption of the edge computing node will not be too large, and the operating cost can be reduced. That is, the present application can make the computing power resources provided by the edge computing node not be much greater than the expected required computing power resource information (i.e., reduce the operating cost) by adjusting the operating frequency of the edge node, nor be much less than the expected required computing power resource information resulting in unmet demands. Thus, the computing power demand and the operating cost can be dynamically balanced. That is, the present application can meet the computing power demand while reducing the operating cost of the edge computing node.

[0246] Further, please refer to Figure 8 , Figure 8 which is a schematic structural diagram of a computer device provided by an embodiment of the present application. As Figure 8 shown, the above Figure 7The device 1 in the corresponding embodiment can be applied to the aforementioned computer device 1000. The computer device 1000 may include: a processor 1001, a network interface 1004, and a memory 1005. In addition, the computer device 1000 further includes: an object interface 1003 and at least one communication bus 1002. Among them, the communication bus 1002 is used to realize the connection and communication between these components. Among them, the object interface 1003 may include a display screen (Display) and a keyboard (Keyboard). Optionally, the object interface 1003 may further include a standard wired interface and a wireless interface. The network interface 1004 may optionally include a standard wired interface and a wireless interface (such as a WI-FI interface). The memory 1005 may be a high-speed RAM memory or a non-volatile memory, such as at least one disk memory. Optionally, the memory 1005 may also be at least one storage device located far from the aforementioned processor 1001. As Figure 8 shown, the memory 1005, as a computer-readable storage medium, may include an operating system, a network communication module, an object interface module, and a device control application program.

[0247] In Figure 8 the computer device 1000 shown, the network interface 1004 can provide network communication functions; while the object interface 1003 is mainly used to provide an input interface for objects; and the processor 1001 can be used to call the device control application program stored in the memory 1005 to achieve:

[0248] Obtain the occupied computing power resource information of the target application for the edge computing node; the occupied computing power resource information is the computing power resource information of the edge computing node occupied when the target application is running;

[0249] Predict the new business objects, offline business objects, and scenario switching business objects for the target application within the target time period, and determine the expected demand computing power resource information of the target application for the edge computing node within the target time period according to the new business objects, offline business objects, and scenario switching business objects;

[0250] Obtain the current operating frequency of the edge computing node, and determine the target operating frequency of the edge computing node within the target time period according to the maximum computing power resource information corresponding to the current operating frequency, the occupied computing power resource information, and the expected demand computing power resource information; the maximum computing power resource information corresponding to the target operating frequency meets the expected demand computing power resource information.

[0251] It should be understood that the computer device 1000 described in the embodiments of the present application can execute the foregoing Figures 3 to 5The description of the edge-computing-based data processing method in the corresponding embodiment can also be executed as described above Figure 7 The description of the edge-computing-based data processing apparatus 1 in the corresponding embodiment will not be repeated here. In addition, the beneficial effects of using the same method will not be described again.

[0252] In addition, it should be noted here that: The embodiment of the present application also provides a computer-readable storage medium, and the computer-readable storage medium stores a computer program executed by the aforementioned computer device 1000 for data processing. The computer program includes program instructions. When the aforementioned processor executes the program instructions, it can execute the description of the aforementioned data processing method in the corresponding embodiment. Therefore, it will not be repeated here. In addition, the beneficial effects of using the same method will not be described again. For the technical details not disclosed in the embodiment of the computer-readable storage medium involved in the present application, please refer to the description of the method embodiment of the present application. Figures 3 to 5 The description of the aforementioned data processing method in the corresponding embodiment will not be repeated here. In addition, the beneficial effects of using the same method will not be described again. For the technical details not disclosed in the embodiment of the computer-readable storage medium involved in the present application, please refer to the description of the method embodiment of the present application.

[0253] The aforementioned computer-readable storage medium may be the internal storage unit of the edge-computing-based data processing apparatus or the aforementioned computer device provided in any of the foregoing embodiments, such as the hard disk or memory of the computer device. The computer-readable storage medium may also be an external storage device of the computer device, such as a plug-in hard disk, a smart media card (SMC), a secure digital (SD) card, a flash card, etc. equipped on the computer device. Further, the computer-readable storage medium may also include both the internal storage unit and the external storage device of the computer device. The computer-readable storage medium is used to store the computer program and other programs and data required by the computer device. The computer-readable storage medium may also be used to temporarily store data that has been output or is to be output.

[0254] In one aspect of the present application, there is provided a computer program product or a computer program. The computer program product or the computer program includes computer instructions, and the computer instructions are stored in a computer-readable storage medium. The processor of the computer device reads the computer instructions from the computer-readable storage medium, and the processor executes the computer instructions, so that the computer device executes the method provided in one aspect of the embodiment of the present application.

[0255] In the description, claims, and drawings of the embodiments of this application, the terms "first", "second", etc. are used to distinguish different objects, rather than to describe a specific order. In addition, the term "comprising" and any variations thereof are intended to cover non-exclusive inclusion. For example, a process, method, device, product, or equipment that includes a series of steps or units is not limited to the listed steps or modules, but may optionally further include steps or modules not listed, or may optionally further include other step units inherent to these processes, methods, devices, products, or equipment.

[0256] Those of ordinary skill in the art can realize that the units and algorithm steps of the examples described in combination with the embodiments disclosed herein can be implemented by electronic hardware, computer software, or a combination of both. To clearly illustrate the interchangeability of hardware and software, the components and steps of the examples have been generally described according to their functions in the above description. Whether these functions are executed in a hardware or software manner depends on the specific application and design constraints of the technical solution. A professional technician can use different methods to implement the described functions for each specific application, but such implementation should not be considered to exceed the scope of this application.

[0257] The methods and related devices provided by the embodiments of this application are described with reference to the method flowcharts and / or structural schematic diagrams provided by the embodiments of this application. Specifically, each process and / or block of the method flowchart and / or structural schematic diagram, and the combination of the processes and / or blocks in the flowchart and / or block diagram can be implemented by computer program instructions. These computer program instructions can be provided to the processor of a general-purpose computer, a special-purpose computer, an embedded processor, or other programmable data processing devices to generate a machine, so that the instructions executed by the processor of the computer or other programmable data processing devices generate a device for implementing the specified functions in Figure 1 one process or multiple processes and / or structural schematic Figure 1 one block or multiple blocks. These computer program instructions can also be stored in a computer-readable memory that can direct a computer or other programmable data processing device to work in a specific manner, so that the instructions stored in the computer-readable memory generate a manufactured article including an instruction device, and the instruction device implements the specified functions in Figure 1 one process or multiple processes and / or structural schematic Figure 1 one block or multiple blocks. These computer program instructions can also be loaded onto a computer or other programmable data processing device, so that a series of operation steps are executed on the computer or other programmable device to generate a computer-implemented process, and thus the instructions executed on the computer or other programmable device provide steps for implementing the specified functions in Figure 1 one process or multiple processes and / or structural schematic one block or multiple blocks.

[0258] The above disclosure is only for the preferred embodiments of the present application. Of course, the scope of rights of the present application cannot be limited thereby. Therefore, equivalent changes made according to the claims of the present application still fall within the scope covered by the present application.

Claims

1. A data processing method based on edge computing, characterized in that, it includes: Obtain the occupied computing power resource information of the target application for the edge computing node; The occupied computing power resource information is the computing power resource information of the edge computing node occupied when running the target application; Predict the new business objects, offline business objects, and scenario switching business objects for the target application within the target time period, and determine the expected required computing power resource information of the target application for the edge computing node within the target time period according to the new business objects, offline business objects, and scenario switching business objects; Determine the current idle computing power resource information of the edge computing node according to the maximum computing power resource information corresponding to the current operating frequency and the occupied computing power resource information, and determine the operating state of the edge computing node according to the current idle computing power resource information; If the operating state of the edge computing node is a full-load operating state, then determine the target operating frequency of the edge computing node within the target time period according to the occupied computing power resource information and the expected required computing power resource information; If the operating state of the edge computing node is a non-full-load operating state, and the expected required computing power resource information is greater than the occupied computing power resource information, then determine the target operating frequency of the edge computing node within the target time period according to the current idle computing power resource information, the occupied computing power resource information, and the expected required computing power resource information; the maximum computing power resource information corresponding to the target operating frequency meets the expected required computing power resource information.

2. The method according to claim 1, characterized in that, The determining the operating state of the edge computing node according to the current idle computing power resource information includes: Match the current idle computing power resource information with an idle resource threshold; If the current idle computing power resource information is greater than the idle resource threshold, then determine the operating state of the edge computing node as the non-full-load operating state; If the current idle computing power resource information is less than the idle resource threshold, then determine the operating state of the edge computing node as the full-load operating state.

3. The method according to claim 1, characterized in that, The if the operating state of the edge computing node is a full-load operating state, then determine the target operating frequency of the edge computing node within the target time period according to the occupied computing power resource information and the expected required computing power resource information includes: If the operating state of the edge computing node is a full-load operating state, then compare the expected required computing power resource information with the occupied computing power resource information; If the expected required computing power resource information is greater than the occupied computing power resource information, then obtain a mapping table; the mapping table includes N mapping relationships; one mapping relationship includes a corresponding relationship between a configured operating frequency and a configured maximum computing power resource information; In the mapping table, determine the configured operating frequency corresponding to the configured maximum computing power resource information greater than the expected required computing power resource information as the target operating frequency; Adjust the operating frequency of the edge computing node from the current operating frequency to the target operating frequency.

4. The method according to claim 3, wherein, the method further includes: If the expected required computing power resource information is less than the occupied computing power resource information, then determine the current operating frequency as the target operating frequency.

5. The method according to claim 1, wherein, when the operating state of the edge computing node is an under-loaded operating state, and the expected required computing power resource information is greater than the occupied computing power resource information, determining the target operating frequency of the edge computing node within the target time period according to the current idle computing power resource information, the occupied computing power resource information, and the expected required computing power resource information includes: When the operating state of the edge computing node is an under-loaded operating state, and the expected required computing power resource information is greater than the occupied computing power resource information, determine the absolute value of the resource difference between the expected required computing power resource information and the occupied computing power resource information; If the current idle computing power resource information is less than the absolute value of the resource difference, obtain a mapping table; the mapping table includes N mapping relationships; one mapping relationship includes a corresponding relationship between a configured operating frequency and a configured maximum computing power resource information; In the mapping table, determine the configured operating frequency corresponding to the configured maximum computing power resource information greater than the expected required computing power resource information as the target operating frequency; Adjust the operating frequency of the edge computing node from the current operating frequency to the target operating frequency.

6. The method according to claim 5, wherein, the method further includes: If the current idle computing power resource information is greater than the absolute value of the resource difference, then determine the current operating frequency as the target operating frequency.

7. The method according to claim 1, wherein, the method further includes: When the operating state of the edge computing node is an under-loaded operating state, and the expected required computing power resource information is less than the occupied computing power resource information, obtain a mapping table; the mapping table includes N mapping relationships; one mapping relationship includes a corresponding relationship between a configured operating frequency, a configured maximum computing power resource information, and a configured operating consumption; Determine the configured operating consumption corresponding to the current operating frequency in the mapping table as the to-be-compared configured operating consumption, and determine the target operating frequency of the edge computing node within the target time period according to the to-be-compared configured operating consumption.

8. The method according to claim 7, wherein, determining the target operating frequency of the edge computing node within the target time period according to the to-be-compared configured operating consumption includes: In the mapping table, determine the configured operating consumption corresponding to the configured maximum computing power resource information greater than the expected required computing power resource information as the target operating consumption; Compare the target operating consumption with the to-be-compared configured operating consumption; If the target running consumption is less than the to-be-compared configured running consumption, determine the configured running frequency corresponding to the target running consumption as the target running frequency, and adjust the running frequency of the edge computing node from the current running frequency to the target running frequency; If the target running consumption is greater than the configured running consumption, determine the current running frequency as the target running frequency.

9. The method according to claim 1, wherein, the obtaining the occupied computing power resource information of the target application for the edge computing node includes: obtaining N current running information of the target application; one current running information includes an application activity scenario where an online service object is located, and the resource occupancy information of the service object in the application activity scenario; obtaining the resource occupancy information of the service object corresponding to each current running information to obtain N pieces of resource occupancy information of the service object; determine the sum of the N pieces of resource occupancy information of the service object as the occupied computing power resource information of the target application for the edge computing node.

10. The method according to claim 9, wherein, The number of application activity scenarios included in the N current running information is one or more; the one or more application activity scenarios include application activity scenario M i ; i is a positive integer; the determining the expected required computing power resource information of the target application for the edge computing node during the target time period according to the newly added service object, the offline service object, and the scenario-switching service object includes: Among the N online service objects, the online service objects whose application activity scenario is the application activity scenario M i are determined as the service objects to be counted; Statistically analyze the total resource occupancy information of the business objects to be counted in the application activity scenario M i for the business objects; Obtain the number of objects corresponding to the business object to be counted, and determine the application activity scenario M according to the total resource occupancy information of the business object and the number of objects i corresponding average computing power demand information; when determining the average computing power demand information corresponding to the one or more application activity scenarios respectively, determine the expected required computing power resource information of the target application for the edge computing node during the target time period according to the average computing power demand information corresponding to the one or more application activity scenarios respectively, the newly added service object, the offline service object, and the scenario-switching service object.

11. The method according to claim 10, wherein, the determining the expected required computing power resource information of the target application for the edge computing node during the target time period according to the average computing power demand information corresponding to the one or more application activity scenarios respectively, the newly added service object, the offline service object, and the scenario-switching service object includes: obtaining the newly added quantity corresponding to the newly added service object, the offline quantity corresponding to the offline service object, and the switching quantity corresponding to the scenario-switching object; predicting the initial login scenario corresponding to the newly added service object; the one or more application activity scenarios include the initial login scenario; determine the application activity scenario where the offline service object is located when the offline behavior occurs as the offline application activity scenario; determine the application activity scenario where the scenario-switching object is located before the scenario switching as the initial application activity scenario, and determine the application activity scenario where the scenario-switching object is located after the scenario switching as the target application activity scenario; Determine the expected demand computing power resource information of the target application for the edge computing node within the target time period according to the average computing power demand information corresponding to the initial login scenario, the offline application activity scenario, the initial application activity scenario, and the target application activity scenario, as well as the switching quantity, the offline quantity, and the new addition quantity.

12. The method according to claim 11, wherein, the determining the expected demand computing power resource information of the target application for the edge computing node within the target time period according to the average computing power demand information corresponding to the initial login scenario, the offline application activity scenario, the initial application activity scenario, and the target application activity scenario, as well as the switching quantity, the offline quantity, and the new addition quantity, includes: determining the first changed computing power resource information corresponding to the newly added service object according to the new addition quantity and the average computing power demand information corresponding to the initial login scenario; determining the second changed computing power resource information corresponding to the offline service object according to the offline quantity and the average computing power demand information corresponding to the online application activity scenario; determining the third changed computing power resource information corresponding to the scenario switching object according to the average computing power demand information corresponding to the initial application activity scenario, the average computing power demand information corresponding to the target application activity scenario, and the switching quantity; determining the expected demand computing power resource information of the target application for the edge computing node within the target time period according to the first changed computing power resource information, the second changed computing power resource information, the third changed computing power resource information, and the occupied computing power resource information of the edge computing node.

13. A computer device, wherein, it includes: a processor, a memory, and a network interface; the processor is connected to the memory and the network interface, wherein the network interface is used to provide network communication functions, the memory is used to store program codes, and the processor is used to call the program codes so that the computer device executes the method according to any one of claims 1-12.

14. A computer-readable storage medium, wherein, a computer program is stored in the computer-readable storage medium, and the computer program is adapted to be loaded and executed by a processor to execute the method according to any one of claims 1-12.

Images