Data processing method, computer device and readable storage medium

Abstract

The present invention discloses a data processing method, a computer device, and a readable storage medium. The data processing method includes: obtaining current operation activity information and one or more associated current operation information of a target application running in an edge computing node; then determining average computing power demand information corresponding to at least two application activity scenarios based on the current operation activity information and the application activity scenario and business user resource occupation information in each current operation information; then determining the changed computing power resource information of the edge computing node within the target time period based on the average computing power demand information and the predicted new business users, offline business users, and scene switching business users of the target application within the target time period; finally, based on the changed computing power resource information, determining the idle computing power resource information of the edge computing node within the target time period. By adopting the present invention, the idle situation of computing power resources within the target time period can be estimated.

Description

Technical Field

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

[0002] Cloud gaming refers to the process of running games on remote servers, compressing and encoding the rendered game screen, and then streaming it to the terminal via the network as an audio and video stream. Cloud gaming eliminates the need to consider terminal configuration, completely solving the technical problem of insufficient terminal performance to run heavy games. However, cloud gaming has very high requirements for network latency. To provide users with more stable network conditions, cloud gaming servers are generally placed closer to users through the large-scale deployment of edge computing nodes.

[0003] Because the number of online game users fluctuates significantly, in order to provide users with a better experience, we generally prepare computing resources based on the peak number of online game users. However, these computing resources will be idle during non-peak hours. During the actual operation of the game, the current idle computing power of the edge computing node is usually obtained by detecting the real-time load of the edge computing node's CPU (central processing unit) or GPU (graphics processing unit). However, if offline computing tasks are assigned to the edge computing node based on the current idle computing power, the offline computing tasks usually take some time to complete. Moreover, since it is impossible to predict whether the computing power resources of the edge computing node will remain idle during this period of time in the future, once the computing power resources of the edge computing node are no longer idle during this period of time, but the offline computing tasks have to continue to be executed, it is easy to cause the edge computing node to be overloaded, resulting in inaccurate resource scheduling. Summary of the Invention

[0004] The embodiments of the present application provide a data processing method, a computer device, and a readable storage medium, which can achieve the quantification of computing resources and estimate the idleness of computing resources in future time periods, thereby performing accurate resource scheduling of edge computing nodes.

[0005] On the one hand, an embodiment of the present application provides a data processing method, including:

[0006] Obtain the current operation activity information of the target application running in the edge computing node, and obtain one or more current operation information associated with the target application; the target application includes at least two application activity scenarios; one current operation information includes the application activity scenario of a business user and the business user's occupied resource information;

[0007] Determine average computing power requirements corresponding to at least two application activity scenarios based on current operation activity information and application activity scenarios and business user resource usage information in each current operation information;

[0008] Predicting new business users, offline business users, and scenario switching business users for target applications on edge computing nodes within a target time period;

[0009] Determine the changed computing power resource information of edge computing nodes within the target time period based on average computing power demand information, new business users, offline business users, and scenario switching business users;

[0010] Based on the total computing power resource information, occupied computing power resource information and changed computing power resource information of the edge computing node, the idle computing power resource information of the edge computing node within the target time period is determined.

[0011] In one aspect, an embodiment of the present application provides a data processing device, including:

[0012] The operation information acquisition module is used to obtain the current operation activity information of the target application running in the edge computing node;

[0013] The operation information acquisition module is used to obtain one or more current operation information associated with the target application; the target application includes at least two application activity scenarios; one current operation information includes the application activity scenario where a business user is located and the business user's occupied resource information;

[0014] A scenario demand determination module is used to determine the average computing power demand information corresponding to at least two application activity scenarios based on the current operation activity information and the application activity scenario and business user resource occupation information in each current operation information;

[0015] The user prediction module is used to predict the number of new business users, offline business users, and scenario switching business users for the target application of the edge computing node within the target time period;

[0016] The change resource determination module is used to determine the change computing power resource information of the edge computing node within the target time period based on the average computing power demand information, new business users, offline business users, and scenario switching business users;

[0017] The idle computing power determination module is used to determine the idle computing power resource information of the edge computing node within the target time period based on the total computing power resource information, occupied computing power resource information and changed computing power resource information of the edge computing node.

[0018] Among them, at least two application activity scenes include application activity scene M i, i is a positive integer less than or equal to the total number of at least two application activity scenarios;

[0019] Scenario requirement determination module, including:

[0020] The first information acquisition unit is used to obtain the application activity scene M in each current operation information. i The corresponding business user's occupied resource information is used as the resource information to be processed;

[0021] The second information acquisition unit is configured to acquire historical operation activity information matching the current operation activity information in the target application, and acquire one or more historical operation information associated with the historical operation activity information;

[0022] The second information acquisition unit is further configured to acquire the application activity scene M from one or more historical operation information. i The corresponding historical business user resource occupation information is used as historical resource information;

[0023] The mean processing unit is used to perform mean processing on historical resource information and pending resource information to obtain the application activity scenario M i The corresponding average computing power demand information.

[0024] Among them, the user prediction module includes:

[0025] A data acquisition unit, used to acquire historical business behavior data of the target application;

[0026] The first user prediction unit is configured to determine new business users of the edge computing node for the target application within a target time period based on the current idle computing resource information, current operation activity information, and historical business behavior data of the edge computing node;

[0027] The second user prediction unit is used to determine the offline business users and scene switching business users of the edge computing node for the target application within the target time period based on the application activity scenario of the business user, current operation activity information, and historical business behavior data.

[0028] Wherein, the first user prediction unit includes:

[0029] A new prediction subunit is added to predict the total number of new business users of the target application in the target time period based on current operational activity information, historical business behavior data, and holiday information;

[0030] The scheduling subunit is used to perform pre-online scheduling processing on the total new business users based on the current idle computing power resource information of the edge computing node, and determine the new business users of the edge computing node for the target application within the target time period.

[0031] The changed resource determination module includes:

[0032] The first resource determination unit is configured to determine new computing resource information corresponding to the new service user based on average computing power demand information corresponding to an initial login scenario in which the new service user resides; the initial login scenario belongs to at least two application activity scenarios;

[0033] The second resource determination unit is configured to determine the released computing power resource information corresponding to the offline service user based on the average computing power demand information corresponding to the application activity scenario where the offline service user is located;

[0034] a third resource determination unit, configured to determine the application activity scene in which the scene switching service user is located before the scene switching as the first application activity scene, and determine the application activity scene in which the scene switching service user is located after the scene switching as the second application activity scene;

[0035] The third resource determination unit is further configured to determine switching computing power resource information corresponding to the scene switching service user based on the average computing power requirement information corresponding to the first application activity scene and the average computing power requirement information corresponding to the second application activity scene;

[0036] The changed resource determination unit is used to determine the changed computing resource information of the edge computing node within the target time period based on the newly added computing resource information, the released computing resource information and the switched computing resource information.

[0037] The one or more current operation information include current operation information N j , j is a positive integer less than or equal to the total number of one or more current operation information; the current operation information N j Including application activity scene H j and business user resource usage information I j ; Business user resource occupancy information I j Including CPU computing power information;

[0038] Run the information acquisition module, including:

[0039] The first computing power acquisition unit is used to determine the central processor of the edge computing node in a unit time for the application activity scenario H j The number of transactions processed is used as the CPU computing power information.

[0040] The one or more current operation information include current operation information N j , j is a positive integer less than or equal to the total number of one or more current operation information; the current operation information N j Including application activity scene H j and business user resource usage information Ij ; Business user resource occupancy information I j Including graphics processor computing power information;

[0041] Run the information acquisition module, including:

[0042] The second computing power acquisition unit is used to determine the graphics processor of the edge computing node in a unit time for the application activity scenario H j The number of floating-point operations performed is used as the computing power information of the graphics processor.

[0043] The one or more current operation information include current operation information N j , j is a positive integer less than or equal to the total number of one or more current operation information; the current operation information N j Including application activity scene H j and business user resource usage information I j ; Business user resource occupancy information I j Including graphics processor computing power information;

[0044] Run the information acquisition module, including:

[0045] The third computing power acquisition unit is used to determine the graphics processor of the edge computing node for the application activity scene H in unit time if the application operation type of the target application is a video processing operation type. j Throughput;

[0046] The third computing power acquisition unit is also used to determine the graphics processor of the edge computing node in unit time for the application activity scenario H j The number of floating-point operations performed;

[0047] The third computing power acquisition unit is further configured to use the throughput and the number of floating-point operations as graphics processor computing power information.

[0048] The one or more current operation information include current operation information N j , j is a positive integer less than or equal to the total number of one or more current operation information; the current operation information N j Including application activity scene H j and business user resource usage information I j ;

[0049] Run the information acquisition module, including:

[0050] The fourth computing power acquisition unit is used to determine the edge computing node running the application activity scenario H per unit time. j Memory usage information is used as the resource usage information of business users. j ;

[0051] The fourth computing power acquisition unit is also used to determine the edge computing node running the application activity scenario H per unit time. j The network bandwidth usage information is used as the resource usage information of the business user. j ;

[0052] The fourth computing power acquisition unit is further used to determine the number of application activity scenarios H that the disk of the edge computing node runs in unit time. j The amount of data read and written is used as disk read and write computing power information, and the disk read and write computing power information is used as business user resource occupation information I j .

[0053] The data processing device further includes:

[0054] The first computing power determination module is used to accumulate the CPU computing power information in the resource occupied information of each business user to obtain the CPU occupied computing power information;

[0055] The second computing power determination module is used to accumulate the graphics processor computing power information in the resource occupation information of each business user to obtain the graphics processor occupied computing power information;

[0056] The third computing power determination module is used to accumulate the memory usage information in the resource usage information of each business user to obtain memory usage computing power information;

[0057] The third computing power determination module is further configured to accumulate the network bandwidth usage information in the resource usage information of each service user to obtain network bandwidth usage computing power information;

[0058] The third computing power determination module is further used to accumulate the disk read and write computing power information in the resource occupied information of each business user to obtain the disk read and write occupied computing power information;

[0059] The occupied computing power determination module is used to use the computing power information occupied by the central processing unit, the computing power information occupied by the graphics processing unit, the computing power information occupied by the memory, the computing power information occupied by the network bandwidth, and the computing power information occupied by the disk read and write as the occupied computing power resource information of the server.

[0060] The floating-point type of the floating-point operation count information includes a half-precision floating-point type, a single-precision floating-point type, and a double-precision floating-point type;

[0061] The second computing power determination module includes:

[0062] A first precision processing unit is configured to sum floating-point operation count information of a half-precision floating-point type from the graphics processor computing power information in the resource occupation information of each service user to obtain half-precision floating-point operation count information;

[0063] The first precision processing unit is further configured to determine a half-precision floating-point computing power percentage based on the half-precision floating-point operation count information and the upper limit of the half-precision floating-point operation count;

[0064] A second precision processing unit is configured to sum floating-point operation count information of a single-precision floating-point type from the graphics processor computing power information in the resource occupation information of each service user to obtain single-precision floating-point operation count information;

[0065] The second precision processing unit is further configured to determine the single-precision floating-point computing power percentage based on the single-precision floating-point operation count information and the total number of single-precision floating-point operations;

[0066] A third precision processing unit is configured to sum floating-point operation count information of a double-precision floating-point type from the graphics processor computing power information in the resource occupation information of each service user to obtain double-precision floating-point operation count information;

[0067] The third precision processing unit is further configured to determine a double-precision floating-point computing power percentage based on the double-precision floating-point operation count information and the total number of double-precision floating-point operations;

[0068] The occupancy information determining unit is used to determine the floating point computing power occupancy percentage based on the half-precision floating point operation count information, the single-precision floating point computing power percentage, and the double-precision floating point computing power percentage, and use the floating point computing power occupancy percentage as the graphics processor occupancy computing power information.

[0069] The data processing device further includes:

[0070] The adjustment module is configured to obtain X business users corresponding to one or more pieces of current running information as business users to be adjusted if the computing power information occupied by the central processing unit is greater than the upper limit threshold of the computing power of the central processing unit, or the computing power information occupied by the graphics processing unit is greater than the upper limit threshold of the computing power of the graphics processing unit, or the computing power information occupied by the memory is greater than the upper limit threshold of the memory computing power, or the computing power information occupied by the network bandwidth is greater than the upper limit threshold of the computing power of the graphics processing unit, or the computing power information occupied by the disk read and write is greater than the upper limit threshold of the disk read and write computing power, where X is a positive integer less than or equal to the total number of the one or more pieces of current running information;

[0071] The migration release module is used to send a scenario migration request for the business user to be adjusted to the edge computing node, so that the edge computing node migrates the application activity scenario of the business user to be adjusted to the idle edge computing node; the idle edge computing node is used to allocate computing resources based on the business user occupied resource information corresponding to the business user to be adjusted; the edge computing node after migration releases the computing resources occupied by the business user to be adjusted.

[0072] The data processing device further includes:

[0073] A neighbor idle module is used to determine the idle computing resource information of the neighbor edge computing nodes of the edge computing node within the target time period; the edge computing node and the neighbor edge computing node belong to the same edge node;

[0074] The edge idle module is used to determine the total idle computing resources of the edge node within the target time period based on the idle computing resources of the edge computing node within the target time period and the idle computing resources of the neighboring edge computing nodes within the target time period.

[0075] The embodiment of the present application can obtain the current operation activity information of the target application running in the edge computing node and one or more current operation information associated with the target application, and then determine the average computing power demand information corresponding to at least two application activity scenarios contained in the target application based on the current operation activity information and the application activity scenario and business user occupied resource information in each current operation information, and then predict the new business users, offline business users and scene switching business users of the edge computing node for the target application within the target time period, and determine the change of computing power resource information of the edge computing node within the target time period based on the average computing power demand information, new business users, offline business users and scene switching business users, and finally determine the idle computing power resource information of the edge computing node within the target time period based on the total computing power resource information, occupied computing power resource information and changed computing power resource information of the edge computing node. Using the method provided by the present application, the computing power resources occupied by the application activity scenario where the business user is located can be quantified by the business user occupied resource information, so as to estimate the idleness of computing power resources within the target time period based on the user changes within the target time period and the average computing power demand information corresponding to the application activity scenario. Since it is possible to accurately predict whether the computing power resources of the edge computing node will be idle in the future time period, this application can better reduce the probability of overload of the edge computing node for tasks that need to occupy the future time period to be completed, that is, it can perform more accurate resource scheduling for the edge computing node. BRIEF DESCRIPTION OF THE DRAWINGS

[0076] In order to more clearly illustrate the embodiments of the present application or the technical solutions in the prior art, the following briefly introduces the drawings required for use in the embodiments or the description of the prior art. Obviously, the drawings described below are only some embodiments of the present application. For ordinary technicians in this field, other drawings can be obtained based on these drawings without any creative work.

[0077] Figure 1 This is a schematic diagram of a network architecture provided by an embodiment of the present application;

[0078] Figure 2a-2b This is a schematic diagram of a scenario for predicting idle computing resources provided by an embodiment of the present application;

[0079] Figure 3 This is a flow chart of a data processing method provided in an embodiment of the present application;

[0080] Figure 4 This is a flow chart of a data processing method provided in an embodiment of the present application;

[0081] Figure 5 This is a schematic diagram of a scenario for determining average computing power requirement information provided by an embodiment of the present application;

[0082] Figure 6 This is a schematic diagram of a system module architecture provided by an embodiment of the present application;

[0083] Figure 7 This is a schematic diagram of a scenario computing power requirement calculation process provided by an embodiment of the present application;

[0084] Figure 8 This is a schematic diagram of an idle computing power prediction process provided by an embodiment of the present application;

[0085] Figure 9 is a structural diagram of a data processing device provided in an embodiment of the present application;

[0086] Figure 10 It is a structural diagram of a computer device provided in an embodiment of the present application. DETAILED DESCRIPTION

[0087] The following will be combined with the drawings in the embodiments of this application to clearly and completely describe the technical solutions in the embodiments of this application. Obviously, the embodiments described are only part of the embodiments of this application, not all of the embodiments. Based on the embodiments in this application, all other embodiments obtained by ordinary technicians in this field without making creative efforts are within the scope of protection of this application.

[0088] Cloud computing refers to the delivery and usage model of IT infrastructure, enabling on-demand, scalable access to required resources over the internet. Broadly speaking, cloud computing refers to the delivery and usage model of services, enabling on-demand, scalable access to required services over the internet. These services can be IT-related, software-related, internet-related, or other services. Cloud computing is the product of the convergence of traditional computer and network technologies, including grid computing, distributed computing, parallel computing, utility computing, network storage technologies, virtualization, and load balancing.

[0089] Cloud computing has rapidly grown, driven by the internet, real-time data streams, the diversification of connected devices, and the growing demand for search services, social networks, mobile commerce, and open collaboration. Unlike previous parallel and distributed computing approaches, the emergence of cloud computing will fundamentally revolutionize the entire internet and enterprise management model.

[0090] Cloud gaming, also known as gaming on demand, is an online gaming technology based on cloud computing. Cloud gaming enables thin clients with relatively limited graphics and data processing capabilities to run high-quality games. In cloud gaming scenarios, the game runs not on the player's terminal but on a cloud server. The cloud server renders the game scene into a video and audio stream, which is then transmitted to the player's terminal over the network. The player's terminal does not need powerful graphics and data processing capabilities; it only needs basic streaming capabilities and the ability to receive player input and send it to the cloud server.

[0091] Computing power, as the name suggests, refers to the computing power of a device. This power exists across a wide range of hardware devices, from mobile phones and computers to supercomputers. Computing resources are the hardware or network resources required by a device to perform computing tasks. These typically include CPU and GPU computing resources, memory resources, network bandwidth resources, and disk resources.

[0092] The solution provided in the embodiments of the present application relates to cloud computing and cloud gaming technology in the field of cloud technology, and the specific process is explained through the following embodiments.

[0093] See Figure 1 , Figure 1This is a network architecture diagram provided by an embodiment of the present application. Figure 1 As shown, the network architecture may include a management server 100 and edge nodes 11, 12, ..., and n, wherein the edge node 11 may include multiple computing servers such as computing servers 11a and 11b, and the edge node 12 may include multiple computing servers such as computing servers 12a and 12b. Figure 1 As shown, computing servers 11a, 11b, and other computing servers in edge node 11 can communicate with each other, and computing servers 12a, 12b, and other computing servers in edge node 12 can communicate with each other. Any computing server in edge node 11, any computing server in edge node 12, ..., and any computing server in edge node 1n can each establish a network connection with the management server 100, so that each computing server can exchange data with the management server 100 via the network connection, and so that each computing server can receive management data from the management server 100. It will be understood that computing servers in edge nodes are typically deployed in the same region, while different edge nodes are typically deployed in different regions.

[0094] like Figure 1 As shown, the computing servers in the above-mentioned edge nodes can all correspond to the terminal device cluster, and each terminal device in the terminal device cluster can be integrated with the target application. When the target application is running in each terminal device, it can exchange data with the computing server assigned to it by the management server 100. Among them, the target application can 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. that have the function of displaying text, images, audio and video data information. The computing server provides corresponding functional services for the target application running in the terminal device, but at the same time consumes corresponding computing resources. The computing resources of a computing server are limited, that is, the number of terminal devices that a computing server can correspond to at the same time is limited. Therefore, when the target application starts running in the terminal device, it will first exchange data with the management server to obtain a computing server that can provide the functional services required by the target application. In order to better select computing servers for newly connected terminal devices and make rational use of computing servers with idle computing resources, the management server 100 will interact with each computing server to obtain the total computing resource information and occupied computing resource information of each computing server, and predict the idle computing resource information of each computing server within the target time period, so as to better reasonably schedule the idle computing resources of the computing server within the target time period.

[0095] For ease of understanding, the idle computing resource information of the computing server 11a within the target time period is estimated by the management server 100 as an example. Figure 1As shown, the management server 100 obtains the current operation activity information of the target application running on the computing server 11a, and then obtains one or more current operation information associated with the target application. The current operation activity information can be the configuration information of the current operation activity of the current application, such as the start time, end time, activity type, activity reward, etc. The target application includes at least two application activity scenarios. The application activity scenario refers to the scene type of the screen displayed on the terminal device when the target application is running on the terminal device. An application activity scenario can implement a business function of the target application. For example, when the target application includes a social application, there are corresponding text scenes, phone scenes, and video scenes. The display screen of the target application belongs to the text scene, and the business user can chat with others in text; the display screen of the target application belongs to the phone scene, and the business user can chat with others by voice; the display screen of the target application belongs to the video scene, and the business user can chat with others by video. A current operation information includes the application activity scene of a business user and the business user's occupied resource information. One business user corresponds to one terminal device. The business user resource usage information is a quantitative indicator of the computing power resources occupied by the computing server 11a when providing functional services to a target application in a terminal device. This quantitative indicator information may include one or more indicators from multiple indicators, such as CPU computing power information, GPU computing power information, memory usage information, network bandwidth usage information, and disk read / write capacity information. The management server 100 then determines the average computing power requirement information corresponding to each application activity scenario of the target application based on the current operational activity information, the application activity scenario in each current operation information, and the business user resource usage information. The average computing power requirement information for an application activity scenario refers to the quantitative indicator information corresponding to the average computing power resources required by multiple business users when the business users are in that application activity scenario. The management server 11a then identifies business users whose computing power resources are subject to change. For example, the computing server 11a may identify business users who provide functional services within a target time period and have only started running the target application within the target time period, as well as business users whose application activity scenarios within the target time period differ from their current application activity scenarios. Furthermore, the management server 100 predicts business users who are currently running the target application and will stop running the target application within the target time period from among the business users currently running the target application. Subsequently, the idle computing resource information of the computing server 11a within the target time period is determined based on the business users whose computing resources will change, the average computing resource demand information, and the total computing resource information and occupied computing resource information of the computing server 11a.Among them, the total computing power resource information is the quantitative index information corresponding to all computing power resources of the computing server 11a, the occupied computing power resource information is the quantitative index information corresponding to the computing power resources occupied by the computing server 11a at the current moment, and the idle computing power resource information is the quantitative index information corresponding to the computing power resources that are not occupied by the computing server 11a within the target time period.

[0096] It is understandable that the above processing process can be executed by the management server alone, by the computing server alone, or by the management server and the computing server together. The specific implementation can be adjusted according to actual needs and is not limited here.

[0097] It is understood that the methods provided in the embodiments of the present application can be executed by computer devices, including but not limited to terminal devices, computing servers, or management servers. The management server can be an independent physical server, 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.

[0098] It is understandable that the above-mentioned devices (such as the above-mentioned management server 100, computing server 11a, computing server 11b, computing server 12a, ..., computing server 12b, etc.) can be a node in a distributed system, wherein the distributed system can be a blockchain system, and the blockchain system can be a distributed system formed by connecting the multiple nodes through 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 the Transmission Control Protocol (TCP) protocol. In a distributed system, any form of computer device, such as a server, terminal device, or other electronic device, can become a node in the blockchain system by joining the peer-to-peer network.

[0099] Among them, the terminal devices in the above-mentioned terminal device cluster may include mobile phones, tablet computers, laptops, PDAs, smart speakers, mobile Internet devices (MID), POS (Point Of Sales) machines, wearable devices (such as smart watches, smart bracelets, etc.), vehicle-mounted equipment, etc.

[0100] For further understanding, the following specific description is given by taking the management server 100 predicting the idle computing resources of the computing server 11b within the target time period as an example.

[0101] Please also see Figure 2a-2b , Figure 2a-2b This is a schematic diagram of a scenario for predicting idle computing resources provided by an embodiment of the present application. Figure 2a As shown, assuming that the target application is a multiplayer shooting cloud game application, business user A, business user B, ... and business user N are all running the cloud game application through terminal devices with which they are bound, and the computing server 11b provides service support for the cloud game application in terminal device 10a, terminal device 10b, ..., terminal device 10n, where computing server 11b can also be called a cloud game server or edge computing node. It is understandable that the computing power resources of the computing server 11b required for different application activity scenarios of the cloud game application are usually different. For example, the application activity scenarios of the cloud game application include lobby scenes, single-player competition scenes, and multiplayer competition scenes. In the lobby scene, usually only simple operations such as character presentation, character costume change, and competition mode selection are required, and the calculations corresponding to the operations are relatively small, so the computing power resource requirements corresponding to the lobby scene are relatively small; in the single-player competition scene, the business user plays a single skydiving game, and the business user's operation instructions correspond to more calculations, and the corresponding computing power resource requirements are relatively large; in the multiplayer competition scene, in addition to meeting the calculations corresponding to the business user's game operation instructions, it is also necessary to provide computing services for operation instructions such as communication exchanges, so the computing power resource requirements corresponding to the multiplayer competition scene are the highest. Figure 2a As shown, the application activity scene 111 of business user A displayed on terminal device 10a can be a single-player competitive scene, the application activity scene 112 of business user B displayed on terminal device 10b can be a single-player competitive scene, ..., and the application activity scene 11n of business user N displayed on terminal device 10n can be a lobby scene. The application activity scene of each business user is one of the lobby scene, the single-player competitive scene, and the multiplayer competitive scene.

[0102] It should be noted that the computing power resource requirements corresponding to the three application activity scenarios of the above-mentioned cloud game application are not fixed. For example, they are affected by some operational activities of the cloud game application. For example, during the Dragon Boat Festival event, the cloud game application will release a new batch of character costumes. At this time, the computing power resource requirements corresponding to the lobby scene will be greater than the computing power resource requirements when there is no Dragon Boat Festival event. Therefore, when the management server 100 predicts the idle computing power resources of the calculation server 11b within the target time period (such as the next ten minutes, the next half hour, the next 24 hours, etc., there is no restriction here), it will first determine the average computing power demand information corresponding to the three application activity scenarios of the cloud game application at the current moment. Figure 2bAs shown, the computing server 11b sends the current operation activity information of the cloud gaming application and one or more associated current operation information to the management server 100. Among them, the current operation activity information includes the configuration information corresponding to the current operation activity of the cloud gaming application. Among them, a current operation information includes an application activity scene where a business user is located and the business user occupied resource information. The business user occupied resource information is used to indicate the computing power resources of the computing server 100 consumed by the application activity scene where the business user is located. Figure 2a It can be seen that the current operation information sent by the computing server 11b to the management server 100 includes the current operation information corresponding to business user A, business user B, ..., and business user N. After the management server 100 obtains the multiple current operation information sent by the computing server 11b, it will obtain the business user resource occupation information corresponding to the lobby scene where the business user's application activity scenario is located, such as the business user resource occupation information corresponding to the application activity scenario 11n where the business user N is located, as the lobby resource information to be processed. At the same time, the management server 100 can also query the historical computing power occupancy, obtain the historical operation information associated with the historical operation activity information that matches the current operation activity information, and then obtain the business user resource occupation information where the application activity scenario is the lobby scene as the historical lobby resource information. Subsequently, the management server 100 will perform average processing on the lobby resource information to be processed and the historical lobby resource information to obtain the average computing power demand information corresponding to the lobby scene. Similarly, the management server 100 can determine the average computing power demand information corresponding to the above-mentioned single-player competitive scene and the above-mentioned multi-player competitive scene.

[0103] like Figure 2bAs shown, after determining the average computing power requirement information, the management server 100 predicts business users whose computing power resources for the cloud gaming application will change during the target time period. For example, business users who require computing server 11b to provide functional services during the target time period and who only start running the cloud gaming application during the target time period, when these business users run the cloud gaming application through a terminal device with a binding relationship, the terminal device usually displays a lobby scene. With each new lobby scene, computing server 11b will occupy the computing power resources corresponding to the average computing power requirement information corresponding to the lobby scene. Business users whose application activity scenes during the target time period are different from their current application activity scenes will also see their occupied computing power resources change when they switch their application activity scenes. In addition, among the business users currently running the cloud gaming application, business users predicted to stop running the cloud gaming application during the target time period will have their computing power resources corresponding to the average computing power requirement information corresponding to their current application activity scenes released when they exit the cloud gaming application. Based on the average computing power demand information corresponding to the three application activity scenarios of the above-mentioned cloud gaming application, the business users whose computing power resources will change, and the total computing power resource information of the computing server 11b and the occupied computing power resource information at the current moment, the idle computing power resource information of the computing server 11b within the target time period can be determined.

[0104] It is understandable that the management server 100 generally manages the above Figure 1 The multiple computing servers shown in FIG. 1 , thus the management server 100 can simultaneously predict the above Figure 1 The idle computing resource information of each computing server in the target time period is shown. The prediction process can be seen in the above Figure 2a-2b The described scenario process. After obtaining the idle computing power resource information of all computing servers, the management server 100 can schedule the computing servers according to actual needs and reasonably utilize the computing servers with relatively idle computing power resources, such as arranging some offline computing tasks for them. Optionally, since the management server 100 can simultaneously predict the idle computing power resource information of multiple computing servers, the management server 100 will obtain the current operation information transmitted by multiple computing servers respectively. The management server 100 can summarize the current operation information transmitted by all computing servers, and then obtain the average computing power demand information corresponding to the application activity scenario of the target application.

[0105] Further, see Figure 3 , Figure 3 This is a flow chart of a data processing method provided by an embodiment of the present application. Figure 1 The computer device described in the embodiment can be executed Figure 1The management server 100 in Figure 1 The computing servers in the edge node cluster (including computing servers 11a, 11b, 12a, and 12b) are shown in FIG. Figure 3 As shown, the data processing method may include the following steps S101 to S105.

[0106] Step S101, obtain the current operation activity information of the target application running in the edge computing node, and obtain one or more current operation information associated with the target application; the target application includes at least two application activity scenarios; one current operation information includes the application activity scenario where a business user is located and the business user's occupied resource information.

[0107] Specifically, the edge computing node can be a computer device that can provide computing or application services, such as a server (which can be the above Figure 1 The target application is the application for which the edge computing node needs to complete the relevant computing tasks. For example, the target application can be a cloud gaming application. Based on cloud computing technology, cloud gaming is usually run on a remote server. The terminal device only needs to receive the audio and video streams sent by the remote server, and then decode and play them. In this case, the remote server is the edge computing node.

[0108] Specifically, when the terminal device runs the target application, the edge computing node will occupy the local computing resources when providing the corresponding computing services. Among them, computing power refers to the computing power of the edge computing node. The target application usually contains at least two application activity scenarios, such as the above Figure 2a The cloud gaming application in includes three application activity scenarios: lobby scenario, single-player competitive scenario, and multiplayer competitive scenario. When the application activity scenarios of business users are different, the computing resources occupied by edge computing nodes may be quite different. Therefore, the current operation information corresponding to a business user usually includes the application activity scenario where the business user is located and the computing resources occupied by the edge computing node to provide computing services for the application activity scenario. The business user occupied resource information is a quantitative indicator information used to characterize the computing resources occupied by the business user. It can be understood that an edge computing node can complete the computing services required by multiple business users when running the target application through terminal devices with which it has a binding relationship. Therefore, an edge computing node may correspond to multiple current operation information.

[0109] Specifically, current operational activity information may refer to the current operational activity configuration of the target application. This information may include the launch of limited-time events. For example, during the Dragon Boat Festival, the target application will launch a limited-time Dragon Boat Festival event. In this case, new application activity scenarios may appear for the target application in response to the current operational activity. For example, if the target application is the aforementioned multiplayer shooter cloud gaming application, based on the current operational activity configuration, the cloud gaming application will launch a limited-time single-player challenge mode during the Dragon Boat Festival. This single-player challenge mode is different from the cloud gaming application's usual single-player and multiplayer competitive modes. The scenario corresponding to the single-player challenge mode of the cloud gaming application is called a single-player challenge scenario. The computing resources required for this application activity scenario may differ from those required for other application activity scenarios. The current operational activity information may also add new elements to the target application's existing application activity scenarios. For example, in cloud gaming, these may include adding fun gameplay, multiple character rewards, new costumes, etc. In this case, the computing resources required for each application activity scenario of the target application will generally be greater than when no operational activity occurs.

[0110] Step S102: Determine average computing power requirement information corresponding to the at least two application activity scenarios respectively according to the current operation activity information and the application activity scenario and business user resource occupation information in each current operation information.

[0111] Specifically, an edge computing node may currently provide computing services for multiple business users in the same application activity scenario. In fact, due to the different operations performed by different business users through terminal devices in the application activity scenario, the computing tasks processed by the edge computing node may also be different. Therefore, the computing power resources occupied by the edge computing node for different business users in the same application activity scenario will also be slightly different. According to the above description of the current operation activity information, it can also be understood that when the target application has different operation activities, the computing power resources corresponding to the application activity scenario will also be different. Therefore, the edge computing node usually determines the average computing power requirement information corresponding to at least two application activity scenarios of the target application based on the current operation activity information and the application activity scenario and business user resource occupation information in each current operation information obtained. Among them, the average computing power requirement information of an application activity scenario refers to the quantitative indicator information corresponding to the average computing power resources required when the business user is in the application activity scenario.

[0112] Specifically, assuming that at least two application activity scenarios of the target application include application activity scenario M i, i is a positive integer less than or equal to the total number of at least two application activity scenarios, then according to the current operation activity information and the application activity scenario and business user resource occupation information in each current operation information, determine the average computing power demand information corresponding to at least two application activity scenarios respectively, including: in each current operation information, obtain the application activity scenario M i The corresponding business user occupied resource information is used as the resource information to be processed; the historical operation activity information matching the current operation activity information in the target application is obtained, and one or more historical operation information associated with the historical operation activity information is obtained; in one or more historical operation information, the application activity scenario M is obtained i The corresponding historical business user resource occupation information is used as historical resource information; the historical resource information and the pending resource information are averaged to obtain the application activity scenario M i The corresponding average computing power demand information. Among them, historical operation activity information refers to the configuration of the target application's previous operation activities. For example, if the Spring Festival operation activity of a cloud gaming application is to enable a limited-time single-player challenge mode, and the current Dragon Boat Festival operation activity also enables a limited-time single-player challenge mode, then the Spring Festival operation activity and the Dragon Boat Festival operation activity can be considered to match.

[0113] Step S103 : predicting the number of new business users, offline business users, and scene switching business users for the target application at the edge computing node within a target time period.

[0114] Specifically, a new business user is a business user who launches the target application within the target time period and for whom the edge computing node provides functional services. A offline business user is a business user currently providing functional services to the edge computing node but who will deactivate the target application within the target time period. A context-switching business user is a business user currently serving a business user whose application activity context differs from the application activity context in the target time period. For example, the edge computing node currently provides functional services to business user A in application activity context 1 and business user B in application activity context 2. However, based on predictions, the edge computing node will need to provide functional services to business user C in the initial login context and business user B in application activity context 1 within the target time period. The initial login context is an application activity context, and users who have just launched the target application are in the initial login context. Therefore, newly added business user C is considered a newly added business user, while business user A, who does not require functional services from the edge computing node, is considered an offline business user. Business user B, whose application activity context has changed, is considered a context-switching business user.

[0115] The computer device will obtain the historical business behavior data of the target application, and then determine the new business users of the edge computing node for the target application within the target time period based on the current idle computing resource information, current operation activity information, and historical business behavior data of the edge computing node; and determine the offline business users and scene-switching business users of the target application within the target time period based on the application activity scenario, current operation activity information, and historical business behavior data of the business user. Among them, the historical business behavior data may include the relevant behavior data of the historical online business users, historical online business users, historical offline business users, and historical scene-switching business users of the target application at each time node in the historical time period, etc. Among them, the relevant behavior data may include the application activity scenario, operation behavior, application running time, etc.

[0116] Specifically, the process of determining the number of new business users of the edge computing node for the target application within the target time period based on the current idle computing resource information, current operation activity information, and historical business behavior data of the edge computing node can be as follows: predicting the total number of new business users of the target application within the target time period based on the current operation activity information, historical business behavior data, and holiday information; performing pre-online scheduling processing on the total number of new business users based on the current idle computing resource information of the edge computing node, and determining the number of new business users of the edge computing node for the target application within the target time period. It should be noted that computer equipment (such as the above Figure 1 The management server 100 shown typically manages multiple edge computing nodes, each of which will have corresponding new business users within a target time period. The computer device typically predicts all business users who will launch the target application within the target time period, which will be used as the total new business users for the target application within the target time period. Pre-online scheduling involves the computer device pre-allocating the new business users corresponding to each edge computing node within the target time period based on the currently idle computing resource information of each edge computing node.

[0117] Optionally, the solutions provided in the embodiments of the present application may involve machine learning technology for artificial intelligence. Machine learning (ML) is a multidisciplinary interdisciplinary subject that involves probability theory, statistics, approximation theory, convex analysis, algorithmic complexity theory, and other disciplines. It specifically studies how computers simulate or implement human learning behavior to acquire new knowledge or skills and reorganize existing knowledge structures to continuously improve their performance. Machine learning is the core of artificial intelligence and the fundamental way to make computers intelligent. Its applications are spread across all areas of artificial intelligence. Machine learning and deep learning generally include artificial neural networks, belief networks, reinforcement learning, transfer learning, inductive learning, and learning by teaching. This is specifically illustrated by the following embodiment: When a computer device predicts the offline business users of an edge computing node for a target application within a target time period, it can generate offline behavior characteristics corresponding to the application activity scenario and current operational activity information of the business user based on the offline prediction model, and then output a predicted offline label corresponding to the offline behavior feature in the offline prediction model. Then, based on the predicted offline label, the offline business user is determined. The offline prediction model is a machine learning model trained based on historical business behavior data. It simulates the offline behavior of business users over different time periods and infers the application activity scenarios, time points, and other behavioral states that trigger user offline. Similarly, computer equipment can also predict business users who switch scenarios and the total number of new business users using the corresponding machine learning model.

[0118] Step S104: Determine the changed computing power resource information of the edge computing node within the target time period based on the average computing power demand information, the newly added business users, the offline business users, and the scene switching business users.

[0119] Specifically, the computer device can determine the newly added computing power resource information corresponding to the new service user based on the average computing power requirement information corresponding to the initial login scenario in which the new service user resides. The newly added computing power resource information can be quantitative indicators corresponding to the computing power resources required by the edge computing node to provide functional services to the new service user. The newly added computing power resource information can be the product of the total number of new service users and the average computing power requirement information corresponding to the initial login scenario. The initial login scenario can include at least two application activity scenarios. It is understood that when a service user logs in or opens a target application, they typically enter a default interface. The application activity scenario corresponding to this default interface can be referred to as the initial login scenario. The computer device can then determine the released computing power resource information corresponding to the offline service user based on the average computing power requirement information corresponding to the application activity scenario in which the offline service user resides. The released computing power resource information can be quantitative indicators corresponding to the computing power resources released after the edge computing node stops providing functional services to the offline service user. The released computing power resource information is equal to the sum of the average computing power requirement information corresponding to the application activity scenarios corresponding to all offline service users. The application activity scene in which the scene switching service user was located before the scene switching was determined as the first application activity scene, and the application activity scene in which the scene switching service user was located after the scene switching was determined as the second application activity scene. The switching computing power resource information corresponding to the scene switching service user is then determined based on the average computing power requirement information corresponding to the first application activity scene and the average computing power requirement information corresponding to the second application activity scene. The switching computing power resource information may be the computing power resources occupied by the edge computing node when providing functional services to the scene switching service user after the application activity scene in which the scene switching service user is located has changed, minus the computing power resources occupied by the edge computing node when providing functional services to the scene switching service user before the application activity scene in which the scene switching service user is located has changed, resulting in quantitative indicator information corresponding to the changed computing power resources. The switching computing power resource information may be equal to the difference between the sum of the average computing power requirement information corresponding to the application activity scenes in which all scene switching service users are located after the change, minus the sum of the average computing power requirement information corresponding to the application activity scenes in which all scene switching service users are located before the change. Based on the newly added computing power resource information, the released computing power resource information, and the switching computing power resource information, the changed computing power resource information of the edge computing node within the target time period is determined. Among them, the changed computing power resource information refers to the quantitative indicator information corresponding to the computing power resources that change from the current moment to the target time period of the edge computing node.Assume that within the target time period, an edge computing node has new business users A and B in the initial login scenario, offline business user C in application activity scenario 1, offline business user D in application activity scenario 2, and scenario switching business user E switching from application activity scenario 2 to application activity scenario 1. The newly added computing power resource information can be the average computing power requirement information corresponding to the initial login scenario multiplied by 2. The released computing power resource information is equal to the average computing power requirement information corresponding to application activity scenario 1 plus the average computing power requirement information corresponding to application activity scenario 2. The switched computing power resource information is equal to the average computing power requirement information corresponding to application activity scenario 1 minus the average computing power requirement information corresponding to application activity scenario 2. The changed computing power resource information is equal to the newly added computing power resource information minus the released computing power resource information plus the switched computing power resource information. If the changed computing power resource information is a positive number, it indicates that the computing power resources occupied by the edge computing node increased during the target time period. If the changed computing power resource information is a negative number, it indicates that the computing power resources occupied by the edge computing node decreased during the target time period.

[0120] Step S105: Determine the idle computing resource information of the edge computing node within the target time period based on the total computing resource information, the occupied computing resource information, and the changed computing resource information of the edge computing node.

[0121] Specifically, the total computing power resource information is the quantitative index information corresponding to all computing power resources of the edge computing node, the occupied computing power resource information is the quantitative index information corresponding to the computing power resources occupied by the edge computing node at the current moment, and the idle computing power resource information is the quantitative index information corresponding to the computing power resources that are not occupied by the edge computing node during the target time period. The idle computing power resource information is equal to the difference between the total computing power resource information of the edge computing node and the occupied computing power resource information, and then subtracting the changed computing power resource information. Among them, when the changed computing power resource information is positive, it indicates the increase of the computing power resources of the edge computing node during the target time period, and when the changed computing power resource information is negative, it indicates the release of the computing power resources of the edge computing node during the target time period.

[0122] It is understandable that after obtaining the idle computing power resource information of the edge computing node within the target time period, the resource scheduling of the edge computing node can be better carried out accurately according to actual needs. In the embodiment of the present application, the computing power resources required for the application activity scenario where a business user is located can be measured by the average computing power demand information. According to the idle computing power resource information of the edge computing node within the target time period and the average computing power demand information, the business users allocated to the edge computing node within the target time period can be adjusted. In addition, according to the idle computing power resource information of the edge computing node within the target time period, some offline computing tasks that can be completed within the target time period can be allocated to the edge computing node to avoid the situation where the target application has a new business user online while the edge computing node is still using idle computing power resources to perform offline computing tasks, thereby causing the edge computing node to be overloaded or even crash.

[0123] In an embodiment of the present application, by obtaining the business user occupied resource information corresponding to the computing power resources occupied by each business user in the edge computing node and the application activity scenario in which the business user is located, the average computing power demand information corresponding to each application activity scenario of the target application is determined, and then by predicting the new business users, offline business users and scene switching business users of the edge computing node for the target application within the target time period, and then according to the average computing power demand information, the new business users, offline business users and scene switching business users, the change computing power resource information of the edge computing node within the target time period is determined, and finally, according to the total computing power resource information, the occupied computing power resource information and the change computing power resource information of the edge computing node, the idle computing power resource information of the edge computing node within the target time period can be determined. The method provided by the embodiment of the present application can realize the quantification of computing power resources and estimate the idle situation of computing power resources within the target time period, which can better schedule the computing power resources of the edge computing node, reduce the situation where the edge computing node is overloaded or the idle computing power resources are wasted, and improve the reasonable utilization rate of computing power resources.

[0124] Further, see Figure 4 , Figure 4 This is a flow chart of a data processing method provided by an embodiment of the present application. Figure 1 The computer device described in the embodiment can be executed Figure 1 The management server 100 in Figure 1 The computing servers in the edge node cluster (including computing servers 11a, 11b, 12a, and 12b) are shown in FIG. Figure 3 As shown, the data processing method may include the following steps S201 to S205.

[0125] Step S201, obtain the current operation activity information of the target application running in the edge computing node, and obtain one or more current operation information associated with the target application; the target application includes at least two application activity scenarios; one current operation information includes an application activity scenario where a business user is located and the business user's occupied resource information; the business user's occupied resource information may include one or more indicator information.

[0126] Specifically, the computing power resources of edge computing nodes usually involve hardware resources such as CPU, GPU, memory, network bandwidth and disk. Therefore, the resource usage information of business users can include one or more indicator information such as central processing unit computing power information, graphics processing unit computing power information, memory usage information, network bandwidth usage information, disk read and write capability information, etc.

[0127] Specifically, assuming that one or more current operation information includes current operation information N j , j is a positive integer less than or equal to the total number of one or more current operation information; the current operation information N j Including application activity scene H j and business user resource usage information I j At this time, the computer device determines the service user's resource occupation information I j The process of obtaining the CPU computing power information contained in can be as follows: determining the CPU computing power of the edge computing node in unit time for the application activity scenario H j The number of transactions processed is used as the CPU computing power information. Computer equipment determines the resource usage information of business users. j The process of obtaining the GPU computing power information contained in can be as follows: determining the GPU computing power of the edge computing node in unit time for the application activity scenario H j The number of floating-point operations performed is used as the graphics processor computing power information. Optionally, the computer device can also determine the application operation type of the target application. If the application operation type of the target application is a video processing operation type, the image processor of the edge computing node is determined to be able to process the application activity scene H in unit time. j The throughput and the number of floating point operations are then used together as the GPU computing power information. j The process can also be: determining the edge computing node to run the application activity scenario H per unit time. j Memory usage information is used as the resource usage information of business users. j Or, determine the edge computing node running application activity scenario H per unit time. jThe network bandwidth usage information is used as the resource usage information of the business user. j Or, determine the disk of the edge computing node running the application activity scenario H per unit time. j The amount of data read and written is used as disk read and write computing power information, and the disk read and write computing power information is used as business user resource occupation information I j .

[0128] It's understandable that the metric types included in the business user resource usage information can be manually pre-configured. All subsequent resource information describing computing resources should use the same metric types as the business user resource usage information to avoid computational confusion. To better quantify edge computing nodes, business user resource usage information can typically include information on CPU computing power, GPU computing power, memory usage, network bandwidth usage, and disk read / write capacity.

[0129] Step S202: Determine average computing power requirement information corresponding to the at least two application activity scenarios respectively based on the current operation activity information and the application activity scenarios and one or more indicator information in each current operation information.

[0130] Specifically, assuming that at least two application activity scenarios of the target application include application activity scenario M i , i is a positive integer less than or equal to the total number of at least two application activity scenes. At this time, the computer device will first determine the application activity scene M i The specific implementation of the corresponding historical resource information and the to-be-processed resource information can refer to the specific description of the above step S102, which will not be repeated here.

[0131] Specifically, if the business user's resource occupation information only contains one indicator information, the computer device will perform average processing on the indicator information contained in the historical resource information and the to-be-processed resource information, and then obtain the average indicator information as the application activity scenario M. i Corresponding average computing power demand information. If the business user's resource information contains multiple indicator information, when performing average processing on the historical resource information and the pending resource information, the indicator information of the same indicator type in the historical resource information and the pending resource information can be averaged to obtain the average indicator information corresponding to the indicator type, and then the average indicator information corresponding to multiple indicator types is used as the application activity scenario M. i For example, the business user's resource usage information includes the CPU computing power information and the GPU computing power information. iWhen the corresponding historical resource information and the to-be-processed resource information are averaged, the computer device averages the CPU computing power information in the historical resource information and the to-be-processed resource information to obtain the average CPU computing power information, and then averages the GPU computing power information in the historical resource information and the to-be-processed resource information to obtain the average GPU computing power information, and then uses the CPU computing power average information and the GPU computing power average information as the application activity scenario M. i The corresponding average computing power demand information.

[0132] Step S203 : predicting the number of new business users, offline business users, and scene switching business users for the target application by the edge computing node within a target time period.

[0133] Specifically, the implementation process of step S203 can be found in the above Figure 3 The specific description of step S103 in the corresponding embodiment will not be repeated here.

[0134] Step S204: Determine the changed computing power resource information of the edge computing node within the target time period based on the average computing power demand information, the newly added business users, the offline business users, and the scene switching business users.

[0135] Specifically, as can be seen from step S202, the average computing power requirement information may include one or more average index information. When the average computing power requirement information includes multiple average index information, it is assumed that the multiple average index information includes average index information. k , k is a positive integer less than or equal to the total number of multiple average indicator information, then according to the average computing power demand information, new business users, offline business users and scene switching business users, the process of determining the change of computing power resource information of the edge computing node within the target time period can be: according to the average indicator information O k , new business users, offline business users and scene switching business users, determine the change indicator information P of the edge computing node within the target time period k , the change index information P k Add to the edge computing node's changing computing power resource information within the target time period.

[0136] Step S205: Determine the idle computing resource information of the edge computing node within the target time period based on the total computing resource information, the occupied computing resource information, and the changed computing resource information of the edge computing node.

[0137] Specifically, computing power resources are characterized by one or more indicators, including CPU computing power information, GPU computing power information, memory usage information, network bandwidth usage information, and disk read / write capacity information. Each indicator information does not interfere with each other and is calculated separately. For example, total computing power resource information includes total CPU computing power information, total GPU computing power information, total memory usage information, total network bandwidth usage information, and total disk read / write capacity information. Occupied computing power resource information includes CPU computing power information, GPU computing power information, memory usage information, network bandwidth usage information, and disk read / write capacity information. Changing computing power resource information includes CPU computing power information, GPU computing power information, memory usage information, network bandwidth usage information, and disk read / write capacity information. The predicted idle computing power resource information will include CPU idle computing power information, GPU idle computing power information, memory usage information, network bandwidth information, and disk read / write capacity information. Among them, the idle computing power information of the central processing unit is equal to the difference between the total computing power information of the central processing unit and the occupied computing power information of the central processing unit, and then subtracted the changed computing power information of the central processing unit; the idle computing power information of the graphics processing unit is equal to the difference between the total computing power information of the graphics processing unit and the occupied computing power information of the graphics processing unit, and then subtracted the changed computing power information of the graphics processing unit; the idle computing power information of memory usage is equal to the difference between the total computing power information of memory usage and the occupied computing power information of memory usage, and then subtracted the changed computing power information of memory usage; the idle computing power information of network bandwidth usage is equal to the difference between the total computing power information of network bandwidth usage and the occupied computing power information of network bandwidth usage, and then subtracted the changed computing power information of network bandwidth usage; the idle computing power information of disk read and write capacity is equal to the difference between the total computing power information of disk read and write capacity and the occupied computing power information of disk read and write capacity, and then subtracted the changed computing power information of disk read and write capacity.

[0138] Optionally, when the occupied computing power resource information includes the computing power information of the central processing unit (CPU), the computing power information of the graphics processing unit (GPU), the computing power information of the memory, the computing power information of the network bandwidth, and the computing power information of the disk read and write, the computer device accumulates the computing power information of the CPU in the resource information occupied by each business user to obtain the computing power information of the CPU; accumulates the computing power information of the graphics processing unit (GPU) in the resource information occupied by each business user to obtain the computing power information of the graphics processing unit (GPU); accumulates the memory usage information in the resource information occupied by each business user to obtain the computing power information of the memory; accumulates the network bandwidth usage information in the resource information occupied by each business user to obtain the computing power information of the network bandwidth; accumulates the computing power information of the disk read and write in the resource information occupied by each business user to obtain the computing power information of the disk read and write. Because the floating-point types of the floating-point operation count information include half-precision floating-point type, single-precision floating-point type, and double-precision floating-point type, the process of accumulating the graphics processor computing power information in the resource occupied information of each business user to obtain the graphics processor computing power information can be as follows: from the graphics processor computing power information in the resource occupied information of each business user, the floating-point operation count information of the half-precision floating-point type is summed to obtain the half-precision floating-point operation count information, and then the half-precision floating-point operation count information and the upper limit of the half-precision floating-point operation are used to determine the half-precision floating-point computing power percentage; from the graphics processor computing power information in the resource occupied information of each business user, the floating-point type of the single-precision floating-point type is summed to obtain the half-precision floating-point operation count information. The floating-point operation count information is summed to obtain single-precision floating-point operation count information; the single-precision floating-point computing power percentage is determined based on the single-precision floating-point operation count information and the total number of single-precision floating-point operations; the floating-point operation count information of the double-precision floating-point type is summed from the graphics processor computing power information in the resource occupied information of each business user to obtain double-precision floating-point operation count information; the double-precision floating-point computing power percentage is determined based on the double-precision floating-point operation count information and the total number of double-precision floating-point operations; the floating-point computing power occupancy percentage is determined based on the half-precision floating-point operation count information, the single-precision floating-point computing power percentage, and the double-precision floating-point computing power percentage, and the floating-point computing power occupancy percentage is used as the graphics processor computing power occupancy information.

[0139] Optionally, if the computing power information occupied by the central processing unit is greater than the upper limit threshold of the computing power of the central processing unit, or the computing power information occupied by the graphics processing unit is greater than the upper limit threshold of the computing power of the graphics processing unit, or the computing power information occupied by the memory is greater than the upper limit threshold of the computing power of the memory, or the computing power information occupied by the network bandwidth is greater than the upper limit threshold of the computing power of the graphics processing unit, or the computing power information occupied by the disk read and write is greater than the upper limit threshold of the computing power of the disk read and write, then X business users are obtained from the business users corresponding to one or more current running information as the business users to be adjusted; X is a positive integer less than or equal to the total number of one or more current running information; then a scenario migration request for the business user to be adjusted is sent to the edge computing node, so that the edge computing node migrates the application activity scenario of the business user to be adjusted to an idle edge computing node; the idle edge computing node is used to allocate computing power resources based on the business user resource occupation information corresponding to the business user to be adjusted; the migrated edge computing node releases the computing power resources occupied by the business user to be adjusted.

[0140] Optionally, multiple edge computing nodes can form an edge node, and usually different edge nodes will be deployed in different areas. The computer device will also determine the idle computing power resource information of the neighboring edge computing nodes of the edge computing node within the target time period, wherein the edge computing node and the neighboring edge computing node belong to the same edge node. Then, based on the idle computing power resources of the edge computing node within the target time period and the idle computing power resources of the neighboring edge computing nodes within the target time period, the total idle computing power resources of the edge node within the target time period are determined. Among them, to determine the idle computing power resource information corresponding to the idle computing power resources of the neighboring edge computing nodes within the target time period, you can refer to the above-mentioned implementation process of the idle computing power resources corresponding to the edge computing nodes, which will not be repeated here.

[0141] Optionally, after obtaining the idle computing resource information of all edge computing nodes within the target time period, the computer device can treat the edge computing nodes whose idle computing resource information exceeds the idle threshold as idle edge computing nodes, and assign some offline computing tasks to them within the target time period, which can improve the utilization rate of the computing resources of the edge computing nodes. The computer device can also treat the edge computing nodes whose idle computing resource information is less than the busy threshold as busy edge computing nodes, adjust the new business users that need to be assigned to the busy edge computing nodes within the target time period, or select a portion of the business users who are providing functional services to the busy edge computing nodes and assign them to the idle edge computing nodes, which can reduce the operation and maintenance losses caused by overload of the busy edge computing nodes.

[0142] By adopting the method provided in the embodiment of the present application, the computing power resources of the edge computing node occupied by the application activity scenario of a business user can be quantified through indicator information such as central processing unit computing power information, graphics processing unit computing power information, memory usage information, network bandwidth usage information, and disk read and write capability information. Therefore, based on the application activity scenario of each business user in the edge computing node and the corresponding multiple indicator information, multiple average indicator information corresponding to an application activity scenario is determined as the average computing power demand information required for the application activity scenario. Further, the idle computing power resource information of the edge computing node is determined based on the user changes of the target application and the average computing power demand information, so that the scheduling of computing power resources of each edge computing node can be more reasonable and effective.

[0143] To facilitate understanding of the above Figure 4 In the corresponding embodiment, when the user resource information is multiple indicator information, the process of the computer device determining the average computing power requirement information of the application activity scenario of the target application is as follows: Figure 5 , Figure 5 This is a schematic diagram of a scenario for determining average computing power demand information provided by an embodiment of the present application. Assuming that the application activity scenarios of the target application include application activity scenario A1, application activity scenario A2, and application activity scenario A3, the business user resource occupation information includes indicator information x and indicator information y, wherein indicator information x and indicator information y can be any two indicator information among the above-mentioned central processing unit computing power information, graphics processing unit computing power information, memory usage information, network bandwidth usage information, and disk read and write capability information. Figure 5 As shown, the management server (which can be the above Figure 1 The management server 100 shown in FIG. 1 obtains m pieces of current operation information, including current operation information 51, current operation information 52, current operation information 53, current operation information 54, ..., current operation information m. One piece of current operation information includes an application activity scenario, indicator information x, and indicator information y of a business user, such as Figure 5 As shown, the current operation information 51 indicates that the application activity scenario of the business user is application activity scenario A1, and the indicator information x is indicator information x. 51 , indicator information y is indicator information y 51 .

[0144] like Figure 5As shown, the management server will divide the current operation information containing the same application activity scenario from the m current operation information to obtain the current operation information set 510, the current operation information set 520 and the current operation information set 530, wherein the application activity scenarios contained in the current operation information in the current operation information set 510 are all application activity scenario A1, the application activity scenarios contained in the current operation information in the current operation information set 520 are all application activity scenario A2, and the application activity scenarios contained in the current operation information in the current operation information set 530 are all application activity scenario A3. Then, the management server will obtain the indicator information of the same indicator type in each current operation information set, and then perform mean processing on it to obtain the average indicator information corresponding to the indicator type, and then use the average indicator information corresponding to all indicator types as the average computing power requirement information corresponding to the application activity scenario. Take the current operation information set 510 corresponding to the application activity scenario A1 as an example for explanation, as shown in Figure 5 As shown, all indicator information corresponding to indicator information x is obtained from the current operation information set 510 to obtain indicator information x set 511, and all indicator information corresponding to indicator information y is obtained from the current operation information set 510 to obtain indicator information y set 512. The management server will perform the operations on the indicator information x in the indicator information x set 511. 51 , indicator information x 52 Perform mean processing and obtain the average index information corresponding to the index information x The management server will process the indicator information y in the indicator information y set 512. 51 , indicator information 52 Perform mean processing and obtain the average indicator information corresponding to the indicator information y The management server then sends the average indicator information and average indicator information As the average computing power requirement information corresponding to the application activity scenario A1. It should be noted that Figure 5 In the scenario diagram shown, when calculating the average indicator information, only the indicator information in the current operation information is used as an example for explanation. To more accurately calculate the average indicator information corresponding to application activity scenario A1, the indicator information corresponding to application activity scenario A1 can also be obtained from the historical operation information associated with the historical operation activity information that matches the current operation activity information of the target application. Based on the indicator type of this indicator information, the indicator information is added to the above-mentioned indicator information x set 511 or indicator information y set 512, and then the indicator information in each set is averaged. This will not be further described here.

[0145] Further, see Figure 6 , Figure 6This is a schematic diagram of a system module architecture provided by an embodiment of the present application. This system module architecture diagram is suitable for cloud gaming scenarios, such as Figure 6 As shown, the system includes a cloud game client 60 (which can be integrated and installed in the above Figure 1 The terminal device corresponding to the computing server), the cloud game server 61 (that is, the above Figure 1 The computing server 11a, computing server 11b, computing server 12a or computing server 12b) and the edge node management server 62 (ie, the above Figure 1 An edge node management server can manage multiple cloud game servers, and a cloud game server can provide services for multiple cloud game clients.

[0146] like Figure 6 As shown, the cloud gaming client 60 may include a node selection module 601 and a core function module 602. The node selection module 601 is used to obtain a suitable cloud gaming server (i.e., the edge computing node mentioned above) by interacting with the edge node management server 62 when a game user (i.e., the business user corresponding to the above-mentioned target application) newly enters the game. The core function module 602 is used to implement the core functions of the cloud gaming, including receiving rendering data from the cloud gaming server, generating corresponding operation instructions in response to the game user's game operation, and uploading the operation instructions to the cloud gaming server.

[0147] like Figure 6As shown, the cloud game server 61 may include a cloud game rendering module 611, a computing power resource management module 612, and a cloud game instance management module 613. The cloud game rendering module 611 is used to implement the core functions of cloud games, including cloud game rendering, cloud game logic calculation and other functions. The computing power resource management module 612 is used to manage the computing power resources of the local machine to ensure that the demand for all real-time running tasks of the local machine does not exceed the physical upper limit. The main functions are: first, the local real-time computing power data (i.e., the total computing power resource information of the edge computing node mentioned above, the occupied computing power resource information, and the business user occupied resource information) is collected and reported. After the computing power data is reported to the edge node management server 62, it will be used to determine the idle computing power resource information of the cloud game server in the next time period (i.e., the above-mentioned target time period); second, local computing power management, that is, when the real-time computing power data of the local machine exceeds the local physical upper limit, it can notify the edge node management server 62 to dispatch some cloud game instances to other idle cloud game servers. It should be noted that computing power resources, including hardware resources such as CPU, GPU, memory network bandwidth, and disk I / O (input / output), are most critical in cloud gaming applications. Generally, computing power evaluations are based on GPUs. However, different cloud games, and even different application scenarios of the same game, have different computing power requirements. Therefore, when predicting real-time peak computing power, the actual impact of these factors must be considered, with the weakest link determining the upper limit. The computing power of these hardware resources is generally measured using the following indicators:

[0148] CPU computing power (i.e. the central processing unit computing power information mentioned above): generally measured by OPS (Operations Per Second). The CPU computing power requirements of different cloud games are also approximately converted into OPS.

[0149] GPU computing power (i.e. the above-mentioned graphics processor computing power information): FLOPS (Floating-point Operations Per Second) or OPS are used to distinguish according to different computing scenarios, where FLOPS represents the number of floating-point operations performed by the GPU per second. At the same time, FLOPS is divided into different types of half-precision, single-precision and double-precision, and the computing power converted from these different types of operations is very different. Therefore, the GPU's carrying capacity needs to be calculated separately based on these different types of measurement indicators. For video processing operations, throughput also needs to be considered. Generally, the lower limit of computing power and throughput is taken as the upper limit of carrying capacity.

[0150] Memory (i.e. the memory usage information mentioned above): the memory space required during operation, generally in MB (megabytes).

[0151] Network bandwidth (i.e., the aforementioned network bandwidth usage information): The bandwidth of both individual machines and edge nodes affects the final load-carrying capacity. Therefore, it is necessary to calculate the bandwidth requirements for both the intranet and external egress, and use the lower limit as the basis for load calculation.

[0152] Disk I / O (i.e., the aforementioned disk read and write computing power information) is divided into throughput in both the read and write directions, generally measured in kilobytes per second (KB / S), which represents the amount of data read or written from the disk per second.

[0153] Based on the above indicators, the cloud gaming server 61 can obtain the local real-time computing power data through the computing power resource management module 612.

[0154] The cloud game instance management module 613 is used to manage the local cloud game instance, wherein a cloud game instance provides computing services for a cloud game running in a cloud game client. The main functions include: instance lifecycle management, that is, creating and destroying cloud game instances according to the request of the edge node management server 62, and managing the running status of the cloud game instance. When an abnormality occurs, such as no response, crash or abnormal resource consumption, the current cloud game instance is recycled and the edge node management server is notified to recreate the cloud game instance. Optionally, the cloud game instance management module 613 can also estimate the growth of computing resources for the game in the next time period based on the game scene running on each cloud game instance and the operation activity arrangement of the cloud game, and report it to the edge computing node management server.

[0155] like Figure 6As shown, the edge node management server 62 includes a real-time capacity calculation module 621, a new game user estimation module 622, an offline game user estimation module 623, a switching game user estimation module 624, and an idle prediction module 625. The real-time capacity calculation module 621 can calculate the real-time computing power usage of each cloud game server (i.e., the occupied computing power resource information of the above-mentioned edge computing node) based on the real-time computing power usage reported by the cloud game server (i.e., the business user occupied resource information corresponding to each business user mentioned above), and calculate the computing power usage of each edge node by summarizing. The new game user estimation module 622 can predict the number of new game users in the next stage and the additional computing power requirements of these new game users based on historical data trends, and preliminarily estimate the allocation on each node and server. The offline game user estimation module 623 can estimate the number of offline game users on each cloud game server in the next period based on the current scene of the game users and historical data trends, thereby calculating the computing power resources that can be released by each cloud game server and edge node. The game user switching estimation module 624 can estimate the number of game users who will switch scenes in each cloud game server in the next period based on the scene in which the game users are currently located and the historical data trend, thereby calculating the changes in computing power resources of each cloud game server and edge node caused by the scene switching of the game users. The idle computing power prediction module 625 can determine the idle computing power of each cloud game server and edge node based on the calculation data of each of the above modules. The calculation process can refer to the following formula (1):

[0156]

[0157] Here, SL is the idle computing power of the cloud gaming server, Smax is the maximum computing power currently available to the cloud gaming server, Scur is the computing power currently occupied by the cloud gaming server, S1 is the computing power that can be released in the next phase, primarily the resources released when offline game users log off, S2 is the computing power occupied by new users in the next phase, and Sgi is the computing power resources that may increase or decrease for a single cloud game in the next phase. Because cloud gaming has periodic activities, the computing power resources will also change when game users enter different scenarios. By recording these changes, the increase in performance resources occupied by game users can also be estimated. It should be noted that the same cloud gaming server can provide computing services for multiple cloud games, and different cloud games have different characteristics. Therefore, when calculating the computing power resource changes caused by switching game users, it is necessary to calculate the cloud gaming instances of different cloud games separately and then aggregate them. Then, n is the number of cloud games for which the cloud gaming server provides computing services. It should be noted that the above mentioned measurement of computing power can include multiple indicators such as CPU computing power, CPU computing power (measured from multiple indicators such as FLOPS, OPS, throughput, etc.), memory space, network bandwidth (intranet bandwidth and edge node bandwidth), etc. When calculating the idle computing power, it is necessary to calculate each indicator through the above formula (1) to obtain the idle computing power of the corresponding indicator, thereby determining the idle computing power of the cloud game server. For the functional implementation of each module based on the edge node management server 62, please refer to the above Figure 3 The description of steps S101-S105 in the corresponding embodiment will not be repeated here.

[0158] The system module architecture diagram provided in the embodiment of the present application solves the technical problem of being unable to accurately measure the idle computing power of edge computing nodes in cloud gaming scenarios. By accurately estimating the idle computing power of each edge computing node, it lays the foundation for the scheduling and idle computing power utilization of cloud gaming on different edge computing nodes.

[0159] Further, see Figure 7 , Figure 7 This is a schematic diagram of a scenario computing power requirement calculation process provided by the embodiment of this application. The scenario computing power requirement is the above Figure 3 The average computing power requirement information corresponding to the application activity scenario. Figure 7 As shown in the figure, the real-time computing power calculation process includes:

[0160] Step S71: The cloud gaming server collects operational activity information.

[0161] Specifically, the operational activity information is the above Figure 3Regarding the current operational activity information, a cloud gaming server can simultaneously create multiple cloud gaming instances corresponding to cloud games. The cloud gaming server will collect operational activity information for all cloud games. Because operational activities can significantly impact game user behavior, it is important to clearly understand the current operational activity configuration. This allows for more accurate predictions of the scenarios game users may enter in the next phase and the impact of operational activities on computing power requirements.

[0162] Step S72: The cloud gaming server collects scene information.

[0163] Specifically, the computing power requirements for different application activity scenarios of the same cloud game vary greatly. Therefore, the cloud game server will identify the current application activity scenario of the cloud game in the cloud game client corresponding to each cloud game instance and report it to the edge node management server.

[0164] Step S73: The cloud gaming server collects computing resource information.

[0165] Specifically, the cloud gaming server collects the computing resources currently occupied by each cloud gaming instance, including CPU, GPU, memory, bandwidth, and disk I / O. The cloud gaming server can send the application activity scenario corresponding to a cloud gaming instance and the currently occupied computing resources as the current operation information of the cloud gaming instance to the edge node management server.

[0166] Step S74: The edge node management server calculates the computing power requirements of the scenario.

[0167] Specifically, the scene computing power requirement is determined based on the average situation of computing power resources occupied by the current application activity scene. The calculation process of scene computing power requirement can be found in the above Figure 3 The description of step S102 in the corresponding embodiment will not be repeated here.

[0168] For further information, see Figure 8 , Figure 8 This is a schematic diagram of an idle computing power prediction process provided by an embodiment of the present application. Figure 8 As shown in the figure, the idle computing power prediction process includes:

[0169] Step S81: The edge node management server determines the occupied computing power of the cloud gaming server.

[0170] Specifically, the edge node management server can summarize the computing power resources occupied by each cloud game instance sent by the cloud game server during the computing power demand calculation process of the above scenario, and determine the computing power occupied by the cloud game server.

[0171] Step S82: The edge node management server predicts new game users.

[0172] Specifically, the edge node management server can estimate the number of new game users in the next stage based on historical big data, current operational activity configuration, holiday factors, etc., and calculate the computing power resources that new game users may occupy based on the game categories they may enter.

[0173] Step S83: The edge node management server predicts offline game users.

[0174] Specifically, based on historical big data, current operational activity configuration, holiday factors, etc., the number of game users who may be offline for each game in the next stage is estimated, and the computing power resources that can be released are calculated based on the current occupied computing power resources.

[0175] Step S84: The edge node management server calculates the change in computing power when the scene is switched.

[0176] Specifically, a game user who is already in the game may switch scenes. The edge node management server can predict whether the game user's game scene will be switched based on the current game user's scene, current operation activity configuration and historical big data, and calculate the increase or decrease in computing power resources caused by this switch.

[0177] Step S85: The edge node management server predicts idle computing power.

[0178] Specifically, based on the above calculation results and formula (1), the idle computing resources aggregated by each cloud gaming server and edge node in the next stage can be predicted.

[0179] By adopting the method provided in the embodiment of the present application, the current idle computing power resources can be accurately calculated according to the actual needs of the game, and the growth or decrease trend of idle computing power can be accurately predicted according to the behavior of game users, thereby predicting the idle computing power resources of the cloud game server and edge node in the next stage, thereby performing accurate resource scheduling between the cloud game server and edge nodes.

[0180] Further, see Figure 9 , Figure 9 : is a structural diagram of a data processing device provided in an embodiment of the present application. The above-mentioned data processing device can be a computer program (including program code) running on a computer device, for example, the data processing device is an application software; the device can be used to execute the corresponding steps of the method provided in an embodiment of the present application. Figure 9 As shown, the data processing device 9 may include: an operation information acquisition module 901, an operation information acquisition module 902, a scenario demand determination module 903, a user prediction module 904, a change resource determination module 905 and an idle computing power determination module 906.

[0181] Operation information acquisition module 901, used to obtain current operation activity information of the target application running in the edge computing node;

[0182] The operation information acquisition module 902 is used to obtain one or more current operation information associated with the target application; the target application includes at least two application activity scenarios; one current operation information includes the application activity scenario of a business user and the business user's occupied resource information;

[0183] Scenario demand determination module 903, configured to determine average computing power demand information corresponding to at least two application activity scenarios based on current operation activity information and application activity scenarios and business user resource usage information in each current operation information;

[0184] The user prediction module 904 is used to predict the number of new business users, offline business users, and scene switching business users for the target application of the edge computing node within the target time period;

[0185] The changed resource determination module 905 is used to determine the changed computing power resource information of the edge computing node within the target time period based on the average computing power demand information, new business users, offline business users, and scenario switching business users;

[0186] The idle computing power determination module 906 is used to determine the idle computing power resource information of the edge computing node within the target time period based on the total computing power resource information, occupied computing power resource information and changed computing power resource information of the edge computing node.

[0187] The specific functional implementation of the operation information acquisition module 901, the operation information acquisition module 902, the scenario demand determination module 903, the user prediction module 904, the change resource determination module 905 and the idle computing power determination module 906 can be found in Figure 3 The detailed description of steps S101 to S105 in the corresponding embodiment will not be repeated here.

[0188] Among them, at least two application activity scenes include application activity scene M i , i is a positive integer less than or equal to the total number of at least two application activity scenarios;

[0189] Please see again Figure 9 The scenario requirement determination module 903 may include: a first information acquisition unit 9031 , a second information acquisition unit 9032 , and a mean processing unit 9033 .

[0190] The first information acquisition unit 9031 is used to obtain the application activity scene M in each current operation information. i The corresponding business user's occupied resource information is used as the resource information to be processed;

[0191] The second information acquisition unit 9032 is used to acquire historical operation activity information matching the current operation activity information in the target application, and acquire one or more historical operation information associated with the historical operation activity information;

[0192] The second information acquisition unit 9032 is further configured to acquire the application activity scene M from one or more historical operation information. i The corresponding historical business user resource occupation information is used as historical resource information;

[0193] The mean processing unit 9033 is used to perform mean processing on the historical resource information and the resource information to be processed to obtain the application activity scene M i The corresponding average computing power demand information.

[0194] The specific functional implementation of the first information acquisition unit 9031, the second information acquisition unit 9032 and the mean value processing unit 9033 can be found in Figure 3 The specific description of step S102 in the corresponding embodiment will not be repeated here.

[0195] Please see again Figure 9 The user prediction module 904 may include: a data acquisition unit 9041 , a first user prediction unit 9042 , and a second user prediction unit 9043 .

[0196] Data acquisition unit 9041, used to acquire historical business behavior data of the target application;

[0197] The first user prediction unit 9042 is used to determine new business users of the edge computing node for the target application within a target time period based on the current idle computing resource information, current operation activity information, and historical business behavior data of the edge computing node;

[0198] The second user prediction unit 9043 is used to determine the offline business users and scene switching business users of the edge computing node for the target application within the target time period based on the application activity scenario of the business user, current operation activity information, and historical business behavior data.

[0199] The specific functional implementation of the data acquisition unit 9041, the first user prediction unit 9042 and the second user prediction unit 9043 can be found in Figure 3 The specific description of step S103 in the corresponding embodiment will not be repeated here.

[0200] Please see again Figure 9 The first user prediction unit 9042 may include: a newly added prediction subunit 90421 and a scheduling subunit 90422.

[0201] A new prediction subunit 90421 is added to predict the total number of new business users of the target application in the target time period based on current operation activity information, historical business behavior data, and holiday information;

[0202] The scheduling subunit 90422 is used to perform pre-online scheduling processing on the total new business users based on the current idle computing power resource information of the edge computing node, and determine the new business users of the edge computing node for the target application within the target time period.

[0203] The specific functional implementation of the newly added prediction subunit 90421 and scheduling subunit 90422 can be found in Figure 3 The specific description of step S103 in the corresponding embodiment will not be repeated here.

[0204] Please see again Figure 9 The changed resource determination module 905 may include: a first resource determination unit 9051 , a second resource determination unit 9052 , a third resource determination unit 9053 and a changed resource determination unit 9054 .

[0205] The first resource determination unit 9051 is configured to determine new computing resource information corresponding to the new service user based on average computing power demand information corresponding to the initial login scenario of the new service user; the initial login scenario belongs to at least two application activity scenarios;

[0206] The second resource determination unit 9052 is configured to determine the released computing power resource information corresponding to the offline service user based on the average computing power demand information corresponding to the application activity scenario where the offline service user is located;

[0207] The third resource determination unit 9053 is configured to determine the application activity scene of the scene switching service user before the scene switching as the first application activity scene, and determine the application activity scene of the scene switching service user after the scene switching as the second application activity scene;

[0208] The third resource determination unit 9053 is further configured to determine switching computing resource information corresponding to the scene switching service user based on the average computing power requirement information corresponding to the first application activity scene and the average computing power requirement information corresponding to the second application activity scene;

[0209] The changed resource determination unit 9054 is used to determine the changed computing resource information of the edge computing node within the target time period based on the newly added computing resource information, the released computing resource information and the switched computing resource information.

[0210] The specific functional implementation of the first resource determination unit 9051, the second resource determination unit 9052, the third resource determination unit 9053 and the changed resource determination unit 9054 can be found in Figure 3 The specific description of step S104 in the corresponding embodiment will not be repeated here.

[0211] The one or more current operation information include current operation information N j , j is a positive integer less than or equal to the total number of one or more current operation information; the current operation information N j Including application activity scene H j and business user resource usage information I j ; Business user resource occupancy information I j Including CPU computing power information;

[0212] Please see again Figure 9 The operation information acquisition module 902 may include: a first computing power acquisition unit 9021.

[0213] The first computing power acquisition unit 9021 is used to determine the CPU of the edge computing node in a unit time for the application activity scenario H j The number of transactions processed is used as the CPU computing power information.

[0214] The specific function implementation of the first computing power acquisition unit 9021 can be found in Figure 4 The specific description of step S201 in the corresponding embodiment will not be repeated here.

[0215] The one or more current operation information include current operation information N j , j is a positive integer less than or equal to the total number of one or more current operation information; the current operation information N j Including application activity scene H j and business user resource usage information I j ; Business user resource occupancy information I j Including graphics processor computing power information;

[0216] Please see again Figure 9 The operation information acquisition module 902 may include: a second computing power acquisition unit 9022.

[0217] The second computing power acquisition unit 9022 is used to determine the GPU of the edge computing node for the application activity scenario H in unit time. j The number of floating-point operations performed is used as the computing power information of the graphics processor.

[0218] The specific function implementation of the second computing power acquisition unit 9022 can be found in Figure 4 The specific description of step S201 in the corresponding embodiment will not be repeated here.

[0219] The one or more current operation information include current operation information N j , j is a positive integer less than or equal to the total number of one or more current operation information; the current operation information N j Including application activity scene H j and business user resource usage information I j ; Business user resource occupancy information I j Including graphics processor computing power information;

[0220] Please see again Figure 9 The operation information acquisition module 902 may include: a third computing power acquisition unit 9023.

[0221] The third computing power acquisition unit 9023 is used to determine the GPU of the edge computing node for the application activity scene H in unit time if the application operation type of the target application is a video processing operation type. j Throughput;

[0222] The third computing power acquisition unit 9023 is also used to determine the graphics processor of the edge computing node in a unit time for the application activity scenario H j The number of floating-point operations performed;

[0223] The third computing power acquisition unit 9023 is further configured to use the throughput and the number of floating-point operations as the graphics processor computing power information.

[0224] The specific function implementation of the third computing power acquisition unit 9023 can be found in Figure 4 The specific description of step S201 in the corresponding embodiment will not be repeated here.

[0225] The one or more current operation information include current operation information N j , j is a positive integer less than or equal to the total number of one or more current operation information; the current operation information N j Including application activity scene H j and business user resource usage information I j ;

[0226] Please see again Figure 9 The operation information acquisition module 902 may include: a fourth computing power acquisition unit 9024.

[0227] The fourth computing power acquisition unit 9024 is used to determine the edge computing node running the application activity scenario H per unit time.j Memory usage information is used as the resource usage information of business users. j ;

[0228] The fourth computing power acquisition unit 9024 is further used to determine the edge computing node running the application activity scenario H per unit time. j The network bandwidth usage information is used as the resource usage information of the business user. j ;

[0229] The fourth computing power acquisition unit 9024 is further configured to determine the number of times the disk of the edge computing node runs the application activity scenario H per unit time. j The amount of data read and written is used as disk read and write computing power information, and the disk read and write computing power information is used as business user resource occupation information I j .

[0230] The specific function implementation of the fourth computing power acquisition unit 9024 can be found in Figure 4 The specific description of step S201 in the corresponding embodiment will not be repeated here.

[0231] Please see again Figure 9 The data processing device 9 may further include: a first computing power determination module 906 , a second computing power determination module 907 , a third computing power determination module 908 and an occupied computing power determination module 909 .

[0232] The first computing power determination module 906 is configured to accumulate the CPU computing power information in the resource usage information of each service user to obtain the CPU computing power usage information;

[0233] The second computing power determination module 907 is configured to accumulate the GPU computing power information in the resource usage information of each service user to obtain GPU computing power information;

[0234] The third computing power determination module 908 is used to accumulate the memory usage information in the resource usage information of each business user to obtain memory usage computing power information;

[0235] The third computing power determination module 908 is further configured to accumulate the network bandwidth usage information in the resource usage information of each service user to obtain network bandwidth usage computing power information;

[0236] The third computing power determination module 908 is further configured to accumulate the disk read and write computing power information in the resource occupied information of each service user to obtain the disk read and write computing power information;

[0237] The occupied computing power determination module 909 is used to use the computing power information occupied by the central processing unit, the computing power information occupied by the graphics processing unit, the computing power information occupied by the memory, the computing power information occupied by the network bandwidth, and the computing power information occupied by the disk read and write as the occupied computing power resource information of the server.

[0238] The specific functional implementation of the first computing power determination module 906, the second computing power determination module 907, the third computing power determination module 908 and the occupied computing power determination module 909 can be found in Figure 4 The optional description of step S205 in the corresponding embodiment will not be repeated here.

[0239] The floating-point type of the floating-point operation count information includes a half-precision floating-point type, a single-precision floating-point type, and a double-precision floating-point type;

[0240] Please see again Figure 9 The second computing power determination module 907 may include: a first precision processing unit 9071, a second precision processing unit 9072, a third precision processing unit 9073 and an occupancy information determination unit 9074.

[0241] The first precision processing unit 9071 is configured to sum floating-point operation count information of half-precision floating-point type from the GPU computing power information in the resource usage information of each service user to obtain half-precision floating-point operation count information;

[0242] The first precision processing unit 9071 is further configured to determine a half-precision floating-point computing power percentage based on the half-precision floating-point operation count information and the upper limit of the half-precision floating-point operation count;

[0243] The second precision processing unit 9072 is configured to sum the floating-point operation count information of the single-precision floating-point type from the GPU computing power information in the resource usage information of each service user to obtain the single-precision floating-point operation count information;

[0244] The second precision processing unit 9072 is further configured to determine the single-precision floating-point computing power percentage based on the single-precision floating-point operation count information and the total number of single-precision floating-point operations;

[0245] The third precision processing unit 9073 is configured to sum the floating-point operation count information of double-precision floating-point type from the GPU computing power information in the resource usage information of each service user to obtain double-precision floating-point operation count information;

[0246] The third precision processing unit 9073 is further configured to determine the double-precision floating-point computing power percentage based on the double-precision floating-point operation count information and the total number of double-precision floating-point operations;

[0247] The occupancy information determining unit 9074 is configured to determine the floating point computing power occupancy percentage based on the half-precision floating point operation count information, the single-precision floating point computing power percentage, and the double-precision floating point computing power percentage, and use the floating point computing power occupancy percentage as the graphics processor occupancy computing power information.

[0248] The specific functional implementation of the first precision processing unit 9071, the second precision processing unit 9072, the third precision processing unit 9073 and the occupancy information determination unit 9074 can be found in Figure 4 The optional description of step S205 in the corresponding embodiment will not be repeated here.

[0249] Please see again Figure 9 The data processing device 9 may further include: an adjustment module 910 and a migration release module 911 .

[0250] Adjustment module 910 is configured to obtain X business users corresponding to one or more pieces of current running information as business users to be adjusted if the CPU computing power usage information is greater than the CPU computing power upper limit threshold, or the GPU computing power usage information is greater than the GPU computing power upper limit threshold, or the memory computing power usage information is greater than the memory computing power upper limit threshold, or the network bandwidth computing power usage information is greater than the GPU computing power upper limit threshold, or the disk read / write computing power usage information is greater than the disk read / write computing power upper limit threshold. X business users are respectively obtained from one or more pieces of current running information as business users to be adjusted, where X is a positive integer less than or equal to the total number of the one or more pieces of current running information.

[0251] The migration release module 911 is used to send a scenario migration request for the business user to be adjusted to the edge computing node, so that the edge computing node migrates the application activity scenario of the business user to be adjusted to the idle edge computing node; the idle edge computing node is used to allocate computing resources based on the business user occupied resource information corresponding to the business user to be adjusted; the migrated edge computing node releases the computing resources occupied by the business user to be adjusted.

[0252] The specific functional implementation of the adjustment module 910 and the migration release module 911 can be found in Figure 4 The optional description of step S205 in the corresponding embodiment will not be repeated here.

[0253] Please see again Figure 9 The data processing device 9 may include: a neighbor idle module 912 and an edge idle module 913.

[0254] A neighbor idle module 912 is configured to determine idle computing resource information of a neighbor edge computing node of the edge computing node within a target time period; the edge computing node and the neighbor edge computing node belong to the same edge node;

[0255] The edge idle module 913 is used to determine the total idle computing resources of the edge node in the target time period based on the idle computing resources of the edge computing node in the target time period and the idle computing resources of the neighboring edge computing nodes in the target time period.

[0256] The specific functional implementation of the neighbor idle module 912 and the edge idle module 913 can be found in Figure 4 The optional description of step S205 in the corresponding embodiment will not be repeated here.

[0257] Further, see Figure 10 , Figure 10 This is a schematic diagram of the structure of a computer device provided in an embodiment of the present application. Figure 10 As shown above Figure 9 The data processing device 9 in the corresponding embodiment can be applied to the above-mentioned computer device 1000, and the above-mentioned computer device 1000 may include: a processor 1001, a network interface 1004 and a memory 1005. In addition, the above-mentioned computer device 1000 also includes: a user 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 user interface 1003 may include a display screen (Display), a keyboard (Keyboard), and the user interface 1003 may optionally 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 (non-volatile memory), such as at least one disk memory. The memory 1005 may optionally also be at least one storage device located away from the aforementioned processor 1001. As Figure 10 As shown, the memory 1005 as a computer-readable storage medium may include an operating system, a network communication module, a user interface module, and a device control application.

[0258] exist Figure 10 In the computer device 1000 shown, the network interface 1004 can provide network communication functions; the user interface 1003 is mainly used to provide an interface for user input; and the processor 1001 can be used to call the device control application stored in the memory 1005 to achieve:

[0259] Obtain the current operation activity information of the target application running in the edge computing node, and obtain one or more current operation information associated with the target application; the target application includes at least two application activity scenarios; one current operation information includes the application activity scenario of a business user and the business user's occupied resource information;

[0260] Determine average computing power requirements corresponding to at least two application activity scenarios based on current operation activity information and application activity scenarios and business user resource usage information in each current operation information;

[0261] Predicting new business users, offline business users, and scenario switching business users for target applications on edge computing nodes within a target time period;

[0262] Determine the changed computing power resource information of edge computing nodes within the target time period based on average computing power demand information, new business users, offline business users, and scenario switching business users;

[0263] Based on the total computing power resource information, occupied computing power resource information and changed computing power resource information of the edge computing node, the idle computing power resource information of the edge computing node within the target time period is determined.

[0264] It should be understood that the computer device 1000 described in the embodiment of the present application can execute the above Figure 3 The description of the data processing method in the corresponding embodiment can also be performed as described above. Figure 9 The description of the data processing device 9 in the corresponding embodiment will not be repeated here. In addition, the description of the beneficial effects of adopting the same method will not be repeated here either.

[0265] In addition, it should be noted that: the embodiment of the present application also provides a computer-readable storage medium, and the computer-readable storage medium stores the computer program executed by the data processing device 9 mentioned above. When the processor executes the computer program, it can execute the above-mentioned Figure 3 The description of the data processing method in the corresponding embodiment will not be repeated here. In addition, the description of the beneficial effects of using the same method will not be repeated here. For technical details not disclosed in the computer storage medium embodiment involved in this application, please refer to the description of the method embodiment of this application.

[0266] The computer-readable storage medium may be the data processing device provided in any of the aforementioned embodiments or the internal storage unit of the computer device, 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. Furthermore, the computer-readable storage medium may also include both the internal storage unit of the computer device and an external storage 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.

[0267] The above disclosure is only a preferred embodiment of the present application, and certainly cannot be used to limit the scope of rights of the present application. Therefore, equivalent changes made according to the claims of the present application are still within the scope covered by the present application.

Claims

1. A data processing method, characterized in that: include: Obtain current operation activity information of a target application running in an edge computing node, and obtain one or more current operation information associated with the target application; The target application includes at least two application activity scenarios; Current operation information includes the application activity scenario of a business user and the resource usage information of the business user; Acquire one or more historical operation information associated with the historical operation activity information matching the current operation activity information in the target application; Determining average computing power requirements corresponding to the at least two application activity scenarios, respectively, based on the one or more historical operation information and the application activity scenarios and the business user resource usage information in each current operation information; Predicting new business users, offline business users, and scenario switching business users for the target application at the edge computing node within a target time period; Determine the changed computing power resource information of the edge computing node within the target time period based on the average computing power demand information, the newly added business users, the offline business users, and the scene switching business users; Determine the idle computing resource information of the edge computing node within the target time period based on the total computing resource information, the occupied computing resource information, and the changed computing resource information of the edge computing node.

2. The method according to claim 1, characterized in that The at least two application activity scenes include application activity scene M i , i is a positive integer less than or equal to the total number of the at least two application activity scenarios; The determining, based on the one or more historical operation information and the application activity scenarios and the business user resource occupation information in each current operation information, the average computing power requirement information corresponding to the at least two application activity scenarios respectively includes: In each current running information, get the application activity scene M i The corresponding business user's occupied resource information is used as the resource information to be processed; In the one or more historical operation information, obtain the application activity scene M i The corresponding historical business user resource occupation information is used as historical resource information; Perform mean processing on the historical resource information and the resource information to be processed to obtain the application activity scenario M i The corresponding average computing power demand information.

3. The method according to claim 1, characterized in that The predicting of new business users, offline business users, and scene switching business users for the target application by the edge computing node within a target time period includes: Obtaining historical business behavior data of the target application; Determining, based on the current idle computing resource information of the edge computing node, the current operation activity information, and the historical business behavior data, a new business user of the edge computing node for the target application within the target time period; According to the application activity scenario of the business user, the current operation activity information, and the historical business behavior data, the edge computing node determines the offline business users and scenario switching business users for the target application within the target time period.

4. The method according to claim 3, characterized in that The determining, based on the current idle computing resource information of the edge computing node, the current operation activity information, and the historical business behavior data, a new business user of the edge computing node for the target application within the target time period includes: Predicting the total number of new business users of the target application within a target time period based on the current operating activity information, the historical business behavior data, and holiday information; Pre-online scheduling is performed on the total new business users according to the current idle computing power resource information of the edge computing node, and the new business users of the edge computing node for the target application within the target time period are determined.

5. The method according to claim 1, wherein The determining, based on the average computing power demand information, the newly added service users, the offline service users, and the scene switching service users, the changed computing power resource information of the edge computing node within the target time period includes: Determining new computing resource information corresponding to the new business user based on average computing power demand information corresponding to an initial login scenario in which the new business user resides; the initial login scenario belongs to the at least two application activity scenarios; Determine the released computing power resource information corresponding to the offline service user based on the average computing power demand information corresponding to the application activity scenario where the offline service user is located; Determine the application activity scene in which the scene switching service user is located before the scene switching as the first application activity scene, and determine the application activity scene in which the scene switching service user is located after the scene switching as the second application activity scene; Determining switching computing resource information corresponding to the scenario switching service user based on the average computing power requirement information corresponding to the first application activity scenario and the average computing power requirement information corresponding to the second application activity scenario; According to the newly added computing power resource information, the released computing power resource information and the switched computing power resource information, the changed computing power resource information of the edge computing node within the target time period is determined.

6. The method according to claim 1, characterized in that The one or more current operation information include current operation information N j , j is a positive integer less than or equal to the total number of the one or more current operation information; The current running information N j Including application activity scene H j and business user resource usage information I j The business user occupies resource information I j Including CPU computing power information; The obtaining one or more pieces of current running information associated with the target application includes: Determine the central processor of the edge computing node in unit time for the application activity scenario H j The number of transactions processed is used as the computing power information of the central processing unit.

7. The method according to claim 1, characterized in that The one or more current operation information include current operation information N j , j is a positive integer less than or equal to the total number of the one or more current operation information; The current running information N j Including application activity scene H j and business user resource usage information I j The business user occupies resource information I j Including graphics processor computing power information; The obtaining one or more pieces of current running information associated with the target application includes: Determine the graphics processor of the edge computing node in unit time for the application activity scenario H j The number of floating-point operations performed is used as the computing power information of the graphics processor.

8. The method according to claim 1, characterized in that The one or more current operation information include current operation information N j , j is a positive integer less than or equal to the total number of the one or more current operation information; The current running information N j Including application activity scene H j and business user resource usage information I j The business user occupies resource information I j Including graphics processor computing power information; The obtaining one or more pieces of current running information associated with the target application includes: If the target application's application operation type is a video processing operation type, then determine the graphics processor of the edge computing node for the application activity scene H in unit time. j Throughput; Determine the graphics processor of the edge computing node in unit time for the application activity scenario H j The number of floating-point operations performed; The throughput and the number of floating-point operations are used together as the graphics processor computing power information.

9. The method according to claim 1, characterized in that The one or more current operation information include current operation information N j , j is a positive integer less than or equal to the total number of the one or more current operation information; The current running information N j Including application activity scene H j and business user resource usage information I j ; The obtaining one or more pieces of current running information associated with the target application includes: Determine that the edge computing node runs the application activity scenario H within a unit time j The memory usage information is used as the resource usage information of the business user. j ; or, Determine that the edge computing node runs the application activity scenario H within a unit time j The network bandwidth usage information is used as the resource occupation information of the service user. j ; or, Determine the disk of the edge computing node running the application activity scenario H in unit time j The data read amount and data write amount are used as disk read and write computing power information, and the disk read and write computing power information is used as the resource occupied information of the business user. j .

10. The method according to claim 1, characterized in that Also includes: Accumulate the CPU computing power information in the resource usage information of each business user to obtain the CPU computing power usage information; Accumulate the GPU computing power information in the resource usage information of each business user to obtain GPU computing power usage information; Accumulate the memory usage information in the resource usage information of each business user to obtain the memory usage computing power information; The network bandwidth usage information in the resource usage information of each business user is accumulated to obtain the network bandwidth usage computing power information; The disk read and write computing power information in the resource usage information of each business user is accumulated to obtain the disk read and write computing power information; The computing power information occupied by the central processing unit, the computing power information occupied by the graphics processing unit, the computing power information occupied by the memory, the computing power information occupied by the network bandwidth, and the computing power information occupied by the disk read and write are respectively used as the occupied computing power resource information of the edge computing node.

11. The method according to claim 10, characterized in that The GPU computing power information in the resource occupied information of each service user is accumulated to obtain the GPU computing power information, including: From the GPU computing power information in the resource usage information of each business user, summing the floating-point operation count information of half-precision floating-point type to obtain half-precision floating-point operation count information; Determining a half-precision floating-point computing power percentage based on the half-precision floating-point operation count information and the upper limit of the half-precision floating-point operation count; From the graphics processor computing power information in the resource usage information of each business user, summing the floating-point operation count information of the single-precision floating-point type to obtain the single-precision floating-point operation count information; Determining a single-precision floating-point computing power percentage based on the single-precision floating-point operation count information and the total number of single-precision floating-point operations; From the GPU computing power information in the resource usage information of each business user, summing the floating-point operation count information of double-precision floating-point type to obtain double-precision floating-point operation count information; Determining a double-precision floating-point computing power percentage based on the double-precision floating-point operation count information and the total number of double-precision floating-point operations; Determine a floating-point computing power occupancy percentage based on the half-precision floating-point operation count information, the single-precision floating-point computing power percentage, and the double-precision floating-point computing power percentage, and use the floating-point computing power occupancy percentage as graphics processor computing power occupancy information.

12. The method according to claim 10, characterized in that The method further comprises: If the CPU computing power information is greater than the CPU computing power upper limit threshold, or the GPU computing power information is greater than the GPU computing power upper limit threshold, or the memory computing power information is greater than the memory computing power upper limit threshold, or the network bandwidth computing power information is greater than the GPU computing power upper limit threshold, or the disk read / write computing power information is greater than the disk read / write computing power upper limit threshold, then X business users are obtained from the business users corresponding to the one or more current running information as the business users to be adjusted; X is a positive integer less than or equal to the total number of the one or more current running information; A scenario migration request for the business user to be adjusted is sent to the edge computing node, so that the edge computing node migrates the application activity scenario of the business user to be adjusted to an idle edge computing node; the idle edge computing node is used to allocate computing resources based on the business user occupied resource information corresponding to the business user to be adjusted; after migration, the edge computing node releases the computing resources occupied by the business user to be adjusted.

13. The method according to claim 1, wherein Also includes: Determine idle computing resource information of neighboring edge computing nodes of the edge computing node within the target time period; The edge computing node and the neighboring edge computing node belong to the same edge node; Determine the total idle computing resources of the edge node in the target time period based on the idle computing resources of the edge computing node in the target time period and the idle computing resources of the neighboring edge computing nodes in the target time period.

14. A computer device, characterized in that: include: processor, memory, and network interface; The processor is connected to the memory and the network interface, wherein the network interface is used to provide a network communication function, the memory is used to store program code, and the processor is used to call the program code to execute the method described in any one of claims 1-13.

15. A computer-readable storage medium, characterized in that The computer-readable storage medium stores a computer program, which is suitable for being loaded by a processor and executing the method according to any one of claims 1 to 13.

Images