CN117440046A

CN117440046A - Data processing method and device for power computing network

Info

Publication number: CN117440046A
Application number: CN202310280401.6A
Authority: CN
Inventors: 蔡国兴
Original assignee: Ultrapower Software Co ltd
Current assignee: Ultrapower Software Co ltd
Priority date: 2023-03-21
Filing date: 2023-03-21
Publication date: 2024-01-23

Abstract

The application provides a data processing method and device for an algorithm power network, which can be used for scheduling algorithm power resources and network resources in a cross-region mode according to the requirements of users and service scenes of the algorithm power network. The method comprises the following steps: acquiring first parameter information and second parameter information of a computing power network, wherein the first parameter information comprises interface parameters of at least one computing power resource service, and the second parameter information comprises interface parameters of at least one network resource service; determining at least one service scheme according to the first parameter information, the second parameter information and at least one service requirement supported by the computing power network; determining a target business scheme from at least one business scheme according to the target business scene and the target business requirement; and invoking a target computing power resource service and a target network resource service indicated by the target business scheme, wherein the target computing power resource service is contained in at least one computing power resource service, and the target network resource service is contained in at least one network resource service.

Description

Data processing method and device for power computing network

Technical Field

The present invention relates to the field of communications networks, and in particular, to a data processing method and apparatus for a computing power network.

Background

With the popularization of new generation information technologies such as big data, artificial intelligence, cloud computing, 5G and the like, the demands of the whole society on data storage, computing, transmission and application are greatly improved, and the demands on computing power with strong industry permeability and wide social applicability are higher and higher. A calculation network with calculation force as a carrier is generated.

A computing network is a "new information infrastructure" that allocates and flexibly schedules computing resources, storage resources, and network resources as needed among clouds, networks, edges, according to business needs. In order to develop the technology of the power network, the country proposes development plans of transregional dispatching power resources and network resources such as Dongxil, dongxil training, dongxil memory, dongxil rendering and the like. However, current computing networks do not enable the scheduling of computing resources and network resources across regions.

Therefore, a data processing method and apparatus for a computing network are needed.

Disclosure of Invention

The application provides a data processing method and device for an algorithm power network, which can be used for scheduling algorithm power resources and network resources in a cross-region mode according to the requirements of users and service scenes of the algorithm power network.

In a first aspect, a data processing method for a computing power network is provided, including:

Acquiring first parameter information and second parameter information of a computing power network, wherein the first parameter information comprises interface parameters of at least one computing power resource service, and the second parameter information comprises interface parameters of at least one network resource service;

determining at least one service scheme according to the first parameter information, the second parameter information and at least one service requirement supported by the computing power network;

determining a target business scheme from the at least one business scheme according to the target business scene and the target business requirement;

and calling a target computing power resource service and a target network resource service indicated by the target business scheme, wherein the target computing power resource service is contained in the at least one computing power resource service, and the target network resource service is contained in the at least one network resource service.

In one example, determining at least one business scenario from the first parameter information, the second parameter information, and at least one business requirement supported by the computing power network comprises:

evaluating the at least one computing power resource service according to the first parameter information, and determining a first evaluation result;

evaluating the at least one network resource service according to the second parameter information, and determining a second evaluation result;

And determining at least one service scheme according to the first parameter information, the second parameter information, the first evaluation result, the second evaluation result and at least one service requirement supported by the computing power network.

In one example, the determining the target business scenario from the at least one business scenario according to the target business scenario and the target business requirement includes:

determining a strategy according to a target business scene, wherein the strategy comprises one or more of cost for calling network resource service, cost for calling computing resource service, service opening time, performance of computing resource service and performance of a network link;

and determining a target business scheme from the at least one business scheme according to the strategy, the target business scene and the target business requirement.

In one example, the determining a policy according to the target traffic scenario includes:

training a neural network model through a service scene and a strategy in a training sample, so that the neural network model learns the corresponding relation between the service scene and the strategy;

and determining a strategy corresponding to the target service scene according to the target service scene based on the neural network model.

In one example, the acquiring the first parameter information and the second parameter information of the computing power network includes:

acquiring interface parameters of multiple groups of computing power resource services acquired from multiple geographic areas, wherein the interface parameters of the multiple groups of computing power resource services form first parameter information, the multiple groups of computing power resource services are in one-to-one correspondence with the multiple geographic areas, and in the process of acquiring the interface parameters of the current group of computing power resource services corresponding to each geographic area:

according to the interface name, transmission protocol, registration position and/or security information of each computing power resource service of the current group of computing power resource services, interfacing with the interface of the corresponding computing power resource service to obtain the interface parameters of each computing power resource service of the current group of computing power resource services;

acquiring interface parameters of a plurality of groups of network resource services acquired from a plurality of geographic areas, wherein the interface parameters of the plurality of groups of network resource services form second parameter information, the plurality of groups of network resource services are in one-to-one correspondence with the plurality of geographic areas, and in the process of acquiring the interface parameters of the current group of network resource services corresponding to each geographic area:

according to the interface name, transmission protocol, registration position and/or security information of each network resource service of the current group of network resource services, interfacing with the interface of the corresponding network resource service to obtain the interface parameters of each network resource service of the current group of network resource services.

In one example, the evaluating the at least one computing power resource service according to the first parameter information, determining a first evaluation result includes:

constructing at least one computational power model for the at least one computational power resource service, wherein each computational power model corresponds to one computational power resource service, the computational power model comprises an attribute of a specific computational power resource service, a name of the specific computational power resource service and a first association relation, the first association relation is used for associating the same attribute of different computational power resource services, and the different computational power resource services comprise the specific computational power resource service;

determining the computing power of each resource pool in at least one resource pool according to the at least one computing power model, and determining a first evaluation result according to the computing power of each resource pool, wherein one or more of the at least one computing power resource services form each resource pool, and the computing power resource services included in different resource pools are not identical.

In one example, the evaluating the at least one network resource service according to the second parameter information, determining a second evaluation result includes:

Constructing at least one network model for the at least one network resource service, wherein each network model corresponds to one network resource service, the network model comprises an attribute of a specific network resource service, a name of the specific network resource service and a second association relation, the second association relation is used for associating the same attribute of different network resource services, and the different network resource services comprise the specific network resource service;

and evaluating the quality of network resource service corresponding to each network model according to the performance of the network link corresponding to each network model so as to acquire a second evaluation result.

In one example, the determining at least one business scenario according to the first parameter information, the second parameter information, the first evaluation result, the second evaluation result, and at least one business requirement supported by the computing power network includes:

determining a first computing resource service having only initial service capabilities from the at least one computing resource service, and determining a first network resource service having only initial service capabilities from the at least one network resource service;

According to the first parameter information, the second parameter information, the first evaluation result and the second evaluation result, the first computing resource service and the first network resource service are organized and organized through an online orchestrator, and at least one intermediate scheme is generated;

and according to the at least one intermediate scheme, at least one business requirement supported by the computing power network, interface parameters of computing power resource services except the first computing power resource service in the at least one computing power resource service, and interface parameters of network resource services except the first network resource service in the at least one network resource service, organizing and arranging the at least one computing power resource service and the at least one network resource service through an online orchestrator, and generating at least one business scheme.

In one example, the invoking the target computing power resource service and the target network resource service indicated by the target business scheme includes:

determining whether a target computing power resource service indicated by the target business scheme is located in a plurality of callable geographic areas;

if the target computing power resource service is located in the plurality of callable geographic areas, determining a target callable geographic area from the plurality of callable geographic areas according to the position of the user;

And calling part of the computing power resource services in the target calling geographic area in the target computing power resource services, and calling the target network resource services indicated by the target business scheme.

In a second aspect, there is provided a data processing apparatus for a computing network, comprising:

the service collection module is used for acquiring first parameter information and second parameter information of the power computing network, wherein the first parameter information comprises interface parameters of at least one power computing resource service, and the second parameter information comprises interface parameters of at least one network resource service;

the service organization arrangement module is used for determining at least one service scheme according to the first parameter information, the second parameter information and at least one service requirement supported by the computing power network;

the service scheduling module is used for determining a target service scheme from the at least one service scheme according to the target service scene and the target service requirement;

the service scheduling module is further configured to invoke a target computing power resource service and a target network resource service indicated by the target business scheme, where the target computing power resource service is included in the at least one computing power resource service, and the target network resource service is included in the at least one network resource service.

According to the data processing method for the power network, interface parameters of at least one power resource service and at least one network resource service are collected uniformly, so that at least one business scheme can be determined by combining business requirements. Then, according to the target service scene and the target service requirement, a target service scheme matched with the target service scene and the target service requirement is determined from at least one service scheme, and then, according to the target service scheme, the target computing power resource service and the target network resource service are called. According to the scheme, the computing power resource service and the network resource service of each area can be collected, so that the computing power resource service and the network resource service which can be called can be conveniently determined according to the target service scene and the target service requirement of the user, the purpose that the user can dispatch the proper computing power resource and network resource in a cross-area mode without setting the computing power resource and the network resource is achieved, and the development cost of the user is greatly saved.

Drawings

In order to more clearly illustrate the technical solutions of the present application, the drawings that are needed in the embodiments will be briefly described below, and it will be obvious to those skilled in the art that other drawings can be obtained from these drawings without inventive effort.

FIG. 1 is a schematic diagram of a computing power network architecture provided in an exemplary embodiment of the present application;

fig. 2 is a flow chart of a data processing method for a computing power network according to an exemplary embodiment of the present application.

FIG. 3 is a schematic diagram of yet another exemplary force network architecture provided in an exemplary embodiment of the present application;

FIG. 4 is a schematic diagram of an exemplary online orchestrator provided in an exemplary embodiment of the present application;

fig. 5 is a schematic diagram of yet another exemplary power network architecture provided in an exemplary embodiment of the present application.

Detailed Description

Embodiments of the present application are described in detail below, examples of which are illustrated in the accompanying drawings, wherein the same or similar reference numerals refer to the same or similar elements or elements having the same or similar functions throughout. The embodiments described below by referring to the drawings are exemplary only for the purpose of illustrating the present application and are not to be construed as limiting the present application. It should be noted that, in the case of no conflict, the embodiments and features in the embodiments may be combined with each other.

As used herein, the singular forms "a", "an", "the" and "the" are intended to include the plural forms as well, unless expressly stated otherwise, as understood by those skilled in the art. It will be further understood that the terms "comprises" and/or "comprising," when used in this specification, specify the presence of stated features, integers, steps, operations, elements, and/or components, but do not preclude the presence or addition of one or more other features, integers, steps, operations, elements, components, and/or groups thereof. It will be understood that when an element is referred to as being "connected" or "coupled" to another element, it can be directly connected or coupled to the other element or intervening elements may also be present. Further, "connected" or "coupled" as used herein may include wirelessly connected or wirelessly coupled. The term "and/or" as used herein includes all or any element and all combination of one or more of the associated listed items.

Fig. 1 is a schematic view of a computing power network architecture provided in an exemplary embodiment of the present application, and as shown in fig. 1, the architecture includes a data processing device 110 for a computing power network, a computing power resource service 120, a network resource service 130, business requirements 140, a business scheme 150, and a user 160.

The computing resource service 120 is configured to provide computing resource services to the data processing device 110 of the computing network, where the computing resource service 120 may include heterogeneous (i.e., different vendors and/or different types of cloud platforms, cloud management platforms, central clouds, edge clouds, and other computing power in society. Specific devices of these computing forces include, but are not limited to, intelligent terminals, edge computing devices (such as edge computing gateways), cloud computing devices, and the like, where the intelligent terminals may be any devices capable of providing computing force resources, such as mobile phones, computers, intelligent voice interaction devices, intelligent home appliances, vehicle terminals, test terminals, computers, and the like, and the present example is not limited thereto. The computational resources include computing resources such as a central processing unit (central processing unit, CPU) or a graphics processor (graphics processing unit, GPU).

The network resource service 130 is used to provide network resource services for the data processing apparatus 110 of the computing network, including networks providing computing connection services for internet protocol (internet protocol, IP) bearer networks, core networks, transport networks, cloud private networks, etc. of heterogeneous operators. Specific devices of these networks include, but are not limited to, gateways, switches, routers, etc., as this example is not limited. Among other things, the network resources provided by these devices include bandwidth, latency, etc.

The business requirements 140 include a demand for an algorithm resource and/or a demand for a network resource, and may further include a regional requirement, a type requirement of an algorithm resource, or a type requirement of a network resource. The computing power resource types comprise a cloud platform, a cloud management platform, a center cloud and an edge cloud. The types of network resources include heterogeneous IP bearer networks, core networks, transport networks, and cloud private networks. For example, a certain traffic demand includes a demand for power storage resources of 50T and a demand for network resources of 500M bandwidth. As another example, business requirements include using XX provincial computing resources.

The processing device 110 of the power network is configured to generate a business scenario 150 according to the power resource provided by the power resource service 120 and the network resource provided by the network resource service 130, in combination with the requirements included in the business requirement 140, to adapt to various business requirements, where the business scenario 150 includes a plurality of business scenarios. The processing device 110 of the computing power network is further configured to select a target business scenario from the business scenarios 150 according to the target business requirements of the user, and schedule computing power resources in the computing power resource service 120 and network resources in the network resource service 130 according to the target business scenario, thereby providing services of the computing power network to the user. Wherein each business scheme comprises schedulable computing power resources and network resources determined according to business requirements.

On the basis of the architecture shown in fig. 1, the embodiment of the present application further provides a data processing method for a computing power network, where the execution subject is a processing device 110 of the computing power network, as shown in fig. 2, and the method includes:

s210, acquiring first parameter information and second parameter information of the computing power network, wherein the first parameter information comprises interface parameters of at least one computing power resource service, and the second parameter information comprises interface parameters of at least one network resource service.

Wherein the interface parameters include one or more of the following: interface name, interface type (e.g., restFul API, etc.), transmission protocol of interface data (e.g., hypertext transfer protocol (hyper text transfer protocol, HTTP), secure hypertext transfer protocol (hypertext transfer protocol secure, HTTPs), transmission control protocol (transmission control protocol, TCP), user datagram protocol (user datagram protocol, UDP), etc.), acquisition information of a service provided by the interface (e.g., acquisition domain of the service, home area of the service, location of server where the service is deployed), security information of the interface (e.g., information for verifying user security, information for verifying session security).

In an example, in order to implement the scheduling of the regional computing power resources and the network resources of the computing power network, an embodiment of the present application further provides an example of an interface parameter collection architecture, as shown in fig. 3, after the interface parameters of the computing power resource service and the interface parameters of the network resource service are collected uniformly in each geographic area, the collected interface parameters and the interface parameters of the network resource service are uniformly transmitted to the processing device 110 of the computing power network, and the processing device 110 of the computing power network performs the scheduling of the regional computing power resource service and the network resource service. In the case of this architecture of the present invention,

based on fig. 3, the manner of acquiring the first parameter information and the second parameter information of the computing power network, which are located in a plurality of geographic areas, includes:

acquiring interface parameters of a plurality of groups of computing power resource services acquired from a plurality of geographic areas, wherein the interface parameters of the plurality of groups of computing power resource services form first parameter information, the plurality of groups of computing power resource services are in one-to-one correspondence with the plurality of geographic areas, and in the process of acquiring the interface parameters of the current group of computing power resource services corresponding to each geographic area: according to the interface name, transmission protocol, registration position and/or safety information of each computing resource service of the current group computing resource service, interfacing with the interface of the corresponding computing resource service to obtain the interface parameter of each computing resource service of the current group computing resource service.

Acquiring interface parameters of a plurality of groups of network resource services acquired from a plurality of geographic areas, wherein the interface parameters of the plurality of groups of network resource services form first parameter information, the plurality of groups of network resource services are in one-to-one correspondence with the plurality of geographic areas, and in the process of acquiring the interface parameters of the current group of network resource services corresponding to each geographic area: according to the interface name, transmission protocol, registration position and/or safety information of each network resource service of the current group of network resource services, interfacing with the interface of the corresponding network resource service to obtain the interface parameter of each network resource service of the current group of network resource services.

Illustratively, the manner of obtaining interface parameters for a multi-group computing resource service acquired from a plurality of geographic areas includes:

after the interface parameters of the computing resource service are acquired from each geographic area, the acquired interface parameters of the computing resource service are encrypted, and the interface parameters of the computing resource service are transmitted in a mode of combining the google remote procedure call (google remote procedure call, gRPC) protocol, secure socket layer (secure socket layer, SSL) protocol or secure transport layer (transport layer security, TLS) protocol with Token authentication, so that the security and confidentiality of data transmission are improved.

The method for acquiring the interface parameters of the multiple sets of network resource services acquired from multiple geographic areas is described above, and is not described herein.

In the above manner, by setting the function of interfacing with interfaces of various types of computing resource services, the interface parameters of the computing resource services and the interface parameters of the network resource services can be collected from each geographic area respectively, so that the data can be uniformly transmitted to the processing device 110 of the computing network, the load of the processing device 110 of the computing network is reduced, and the data collection of safety isolation under the cross-regional and cross-network conditions is realized.

Illustratively, when the interface parameters of the computing resource service are collected from each geographic area, a springboot architecture (a brand new framework provided by a Pivotal team) is adopted, and a swagger (interface document generating tool) is utilized to provide a restful API (REST-style API) interface to the outside in a unified way, so that loose coupling between the collection and the computing network service is realized. For distributed acquisition tasks, an Elastic joob and Zookeeper distributed design is adopted, wherein the Elastic Job relies on the Zookeeper to finish storage of execution task information (such as task names, task participation examples, task execution strategies and the like), so that Elastic adjustment of clusters is realized, the problem of failure transfer of acquisition task scheduling is solved, and the continuous integrity of acquired data is ensured. The Dubbo registration is utilized to externally expose a remote method call (remote method invocation, RMI) mode interface, and data access of various data sources is supported, wherein Dubbo is a high-performance and lightweight open source Java RPC framework.

Before the interface parameters of the current group of the computing power resource services corresponding to each geographic area are collected, the computing power resource services in each geographic area are registered based on the set service gateway according to the information such as the interface name, the transmission protocol, the registration position and/or the security information of the computing power resource services in each geographic area, so that the unified collection of the information of various different manufacturers and different types of computing power resource services is realized, and the subsequent efficient interfacing with the interfaces of the computing power resource services in each geographic area is facilitated. The types of the computing power resource service comprise a computing power resource opening service, a computing power resource changing service, a computing power resource canceling service and the like.

The network resource service in each geographic area is registered according to the information such as the interface name, the transmission protocol, the registration position and/or the security information of the network resource service in each geographic area before the interface parameters of the current group of network resource services corresponding to each geographic area are collected, so that the unified collection of the information of various different vendors and different types of network resource services is realized, and the subsequent efficient interfacing with the interfaces of the network resource services in each geographic area is facilitated. The types of the network resource services comprise a network resource opening service, a network resource changing service, a network resource canceling service and the like.

In one example, a label for each of the power resource services is determined based on each of the power resource services, the registration location, the performance index, and/or the statistical information (e.g., total CPU core number, total memory size, etc.), and interface parameters for the power resource services are stored based on the labels for each of the power resource services to facilitate subsequent selection of the power resource services when determining the business scenario based on the business requirements. The performance indexes of the computing power resource service comprise the utilization rate or distribution rate of the CPU, the utilization rate or distribution rate of the GPU, the size of storage resources and the like.

In one example, a label for each network resource service is determined based on each network resource service, registration location, performance metrics, and/or statistics (e.g., network delay, jitter, packet loss), etc., and interface parameters for each network resource service are stored based on the label for that network resource service to facilitate subsequent selection of the network resource service in determining a business scenario based on business needs. The performance indexes of the network resource service comprise network bandwidth, time delay, jitter, packet loss and the like.

S220, determining at least one service scheme according to the first parameter information, the second parameter information and at least one service requirement supported by the computing power network.

In one example, determining at least one business scenario from the first parameter information, the second parameter information, and at least one business requirement supported by the computing network comprises:

first, a first computing resource service having only initial service capabilities is determined based on the at least one computing resource service, and a first network resource service having only initial service capabilities is determined based on the at least one network resource service. And then organizing and arranging the first computing resource service and the first network resource service through an online scheduler according to the first parameter information and the second parameter information to generate at least one intermediate scheme. And then organizing and arranging the at least one power resource service and the at least one network resource service through an online orchestrator according to the at least one intermediate scheme, the at least one business requirement supported by the power network, the interface parameters of the power resource services except the first power resource service in the at least one power resource service, and the interface parameters of the network resource services except the first network resource service in the at least one network resource service.

The initial service capability of the first computing resource service (may be called as a computing atomic service) includes computing resource opening, computing resource changing, computing resource deleting and the like.

In one example, in organizing and orchestrating the first computing resource service and the first network resource service by the online orchestrator according to the first parameter information and the second parameter information, during the generating of the at least one intermediate solution:

and according to at least one business requirement, the interface of the first computing power resource service is packaged and/or modified through an online orchestrator.

The initial service capabilities that the first network resource service (which may be referred to as a network atomic service) has include network resource provisioning, network resource querying, network resource changing, and the like.

and according to at least one service requirement, the interface of the first network resource service is packaged and/or modified through an online orchestrator.

In the above example, the interface of the atomic service and the interface of the atomic service are interfaces that have not been encapsulated and modified. Therefore, in the process of generating the intermediate scheme by arranging the computing power atomic service and the network atomic service, the interfaces of the computing power atomic service or the interfaces of the network atomic service can be packaged and/or modified (can be called secondary packaging) in advance based on service requirements, so that the intermediate scheme after secondary packaging can be directly used in the subsequent service scheme arranging process, the complexity of arranging the service scheme is reduced, and the service scheme arranging efficiency is improved.

On-line orchestrator as shown in fig. 4, the operation actions of the on-line orchestrator include: start, end, loop node, wait node, decision node, branch merge node, service node, etc.

Illustratively, a first computing resource service is selected from the at least one computing resource service, and a first network resource service is selected from the at least one network resource service. Then, on the basis of the operation action in the online orchestrator, performing online organization and orchestration, organizing and orchestrating corresponding service nodes according to the initial service capacity of the first computing resource service, and configuring the corresponding service nodes from the interface parameters related to the first computing resource service in the first parameter information; and organizing corresponding service nodes according to the initial service capacity organization of the first network resource service, and configuring the corresponding service nodes from the interface parameters related to the first network resource service in the second parameter information. If all the service nodes from the start node to the end node can execute successfully, an intermediate scheme is generated. Since the initial service capability of the first computing resource service includes a plurality of, and the initial service capability of the first network resource service includes a plurality of, a plurality of intermediate schemes may be generated by selecting and arranging the initial service capability of the first computing resource service and the initial service capability of the first network resource service.

The plurality of intermediate schemes are generated by organizing the computing atomic service and the network atomic service, and are suitable for adding the computing general service or the network general service on the basis of the organization and the organization to generate a final business scheme. The organized intermediate scheme can be stored, and the intermediate scheme can be directly used when the business scheme is organized later, so that the organization efficiency of the business scheme can be improved. The computing power general service and the network general service comprise parameter verification of a service interface, connectivity detection of the service interface, authentication of the service and the like.

The manner in which the business scenario is generated by the intermediate scenario illustratively includes:

the method comprises the steps of determining a computing power general service except a computing power atomic service from at least one computing power resource service, further determining interface parameters of the computing power general service, determining a network general service except a network atomic service from at least one network resource service, and further determining interface parameters of the network general service. And selecting an addable intermediate business scheme according to at least one business requirement supported by the computing power network, and selecting an addable computing power general service and a network general service according to the at least one business requirement on the basis of the intermediate business scheme to generate at least one business scheme. Each business scenario is adapted to one or more business requirements, each business requirement is adapted to one or more business scenarios.

In the above example, according to the interface parameters of at least one computing power resource service and the interface parameters of at least one network resource service collected uniformly across regions, the computing power atomic service and the network atomic service are organized and organized based on the online composer to generate at least one universal intermediate scheme, and then the at least one business scheme can be generated efficiently based on at least one business requirement in combination with the at least one intermediate scheme.

In one example, S220 further includes:

at least one computing power resource service is firstly evaluated according to the first parameter information, and a first evaluation result is determined. And then, evaluating at least one network resource service according to the second parameter information, and determining a second evaluation result. And then determining at least one service scheme according to the first parameter information, the second parameter information, the first evaluation result, the second evaluation result and at least one service requirement supported by the computing power network.

In the above example, after the computing power resource service and the network resource service are evaluated, the service requirements of the computing power resource service and the network resource service adaptation can be accurately determined, so that the computing power resource service and the network resource service meeting the corresponding service requirements can be selected when the service scheme is determined, and the adaptation degree of the service scheme and the service requirements is improved.

In one example, the manner of determining the first evaluation result includes:

at least one computing power model is firstly constructed for at least one computing power resource service, each computing power model corresponds to one computing power resource service, the computing power model comprises an attribute of the aimed computing power resource service, a name of the aimed computing power resource service and a first association relation, the first association relation is used for associating the same attribute of different computing power resource services, and the different computing power resource services comprise aimed computing power resource services. And then determining the computing power of each resource pool in at least one resource pool according to at least one computing power model, and determining a first evaluation result according to the computing power of each resource pool, wherein one or more of at least one computing power resource service forms each resource pool, and the computing power resource services included in different resource pools are not identical. The computational power of each resource pool is the sum of the computational power resource services that the resource pool includes.

The attributes of the computing resource service include service identification, service name, service category, service protocol, service input parameters, service output parameters and the like.

The first association relationship is used for associating the same attribute of different computing power resource services, and comprises the following steps:

The first association is used for associating an association attribute between a service identifier of the specific computing power resource service and service identifiers of other computing power resource services, for example, the service identifier of the specific computing power resource service is a computing service, and the service identifiers of other computing power resource services are authentication services.

It should be appreciated that the role of the computational model is to define what information a computational resource service needs to record, and initially define an abstract computational model that itself is not specific to the computational resource service. After the attribute, the name and the first association relation of one computing resource service are acquired, the information is filled into a computing model to generate a corresponding apparent computing model of the computing resource service.

For example, a resource pool includes three different types of parallel settlement chips b1, b2, b3 manufactured by different manufacturers, and the calculation performance of the resource pool can be represented by the following formula (1):

where f (x) is a mapping function for evaluating the performance of the chip, x=b ₁ ，b ₂ ，b ₃ ，β _n Is a mapping coefficient determined according to the performance of the chip, n=1, 2,3, wherein the performance of the chip includes the memory size and/or the operation speed of the chip, etc.

In the mode, by defining the computational power model, the unified model is created for each computational power resource service, and the unified evaluation of the computational power of each computational power resource service is facilitated.

In one example, the manner of determining the second evaluation result includes:

at least one network model is firstly constructed for at least one network resource service, each network model corresponds to one network resource service, the network model comprises an attribute of the aimed network resource service, a name of the aimed network resource service and a second association relation, the second association relation is used for associating the same attribute of different network resource services, and the different network resource services comprise the aimed network resource service. And then, according to the performance of the network link corresponding to each network model, evaluating the quality of the network resource service corresponding to each network model to obtain a second evaluation result.

Wherein, the network resource service attribute service name, service identification, service category, service protocol, service input parameter, service output parameter, etc.

The second association relationship is used for associating the same attribute of different network resource services, and comprises the following steps:

The second association is used to associate a primary backup attribute between the network link of the targeted network resource service and the network link of the other network resource service, e.g., the network link of the targeted network resource service is the primary link and the network links of the other network resource service are backup links.

It should be appreciated that the role of the network model is to define which of the information a network resource service needs to record, and initially define an abstract network model that itself is free of specific information of the network resource service. After the attribute, the name and the second association relation of a network resource service are acquired, the information is filled into the network model to generate an apparent network model corresponding to the network resource service.

For example, a network model corresponding to a network resource service includes: and the network bandwidth, time delay, jitter, packet loss rate and other information from the source end to the destination end are respectively corresponding to a scaling factor, and each scaling factor represents the proportion of the corresponding information in network quality evaluation.

In the above manner, by defining the network model, a unified model is created for each network resource service, which is beneficial to efficiently and uniformly evaluating the network quality of each network resource service.

In one example, determining the at least one business scenario based on the first parameter information, the second parameter information, the first evaluation result, the second evaluation result, and the at least one business requirement supported by the computing power network comprises:

first, a first computing resource service having only initial service capabilities is determined based on the at least one computing resource service, and a first network resource service having only initial service capabilities is determined based on the at least one network resource service. And then organizing and arranging the first computing power resource service and the first network resource service through an online scheduler according to the first parameter information, the second parameter information, the first evaluation result and the second evaluation result to generate at least one intermediate scheme. And then organizing and arranging the at least one power resource service and the at least one network resource service through an online orchestrator according to the at least one intermediate scheme, the at least one business requirement supported by the power network, the interface parameters of the power resource services except the first power resource service in the at least one power resource service, and the interface parameters of the network resource services except the first network resource service in the at least one network resource service.

The above example is illustrated below on the basis of the online orchestrator shown in fig. 4:

illustratively, a first computing resource service is selected from the at least one computing resource service, and a first network resource service is selected from the at least one network resource service. Then, on the basis of the operation action in the online orchestrator, performing online organization and orchestration, organizing and orchestrating corresponding service nodes according to the initial service capacity of the first computing resource service, and configuring the corresponding service nodes from the interface parameters related to the first computing resource service in the first parameter information; and organizing corresponding service nodes according to the initial service capacity organization of the first network resource service, and configuring the corresponding service nodes from the interface parameters related to the first network resource service in the second parameter information. According to the preset time interval, a first evaluation result of each computing power resource service is obtained, so that the computing power resource service to be arranged in the intermediate scheme is determined according to the computing power information and the like of each computing power resource service. And obtaining a second evaluation result of each network resource service according to the preset time interval so as to determine the network resource service to be arranged in the intermediate scheme according to the information such as the network bandwidth, time delay, jitter, packet loss rate and the like of each network resource service. And generating an intermediate scheme if all service nodes from the starting node to the ending node can be successfully executed in the process of arranging the intermediate scheme according to the to-be-arranged computing resource service and the to-be-arranged network resource service. According to the scheme, the calculation power and the quality of the network resource service which can be provided by each calculation power resource service are monitored in real time, so that the overall execution quality of the intermediate scheme is ensured. Since the initial service capability of the first computing resource service includes a plurality of, and the initial service capability of the first network resource service includes a plurality of, a plurality of intermediate schemes may be generated by selecting and arranging the initial service capability of the first computing resource service and the initial service capability of the first network resource service.

the method comprises the steps of determining a computing power general service except a computing power atomic service from at least one computing power resource service, further determining interface parameters of the computing power general service, determining a network general service except a network atomic service from at least one network resource service, and further determining interface parameters of the network general service. A preliminary intermediate business scheme is then selected that meets the at least one business requirement based on the at least one business requirement supported by the power network. And selecting an intermediate business scheme which can be added and has the network quality and the calculation power meeting the business requirements from the prepared intermediate business schemes according to the first evaluation result and the second evaluation result. And selecting the addable computing power general service and the network general service according to at least one business requirement on the basis of the addable intermediate business scheme to generate at least one business scheme. Each business scenario is adapted to one or more business requirements, each business requirement is adapted to one or more business scenarios.

In the above example, according to the interface parameters of at least one computing power resource service and the interface parameters of at least one network resource service collected in a unified manner across regions, the computing power atomic service and the network atomic service are organized and organized based on the online composer to generate at least one general intermediate scheme, and then the intermediate scheme with high adaptation degree is selected from the at least one intermediate scheme based on the at least one service requirement in combination with the first evaluation result and the second evaluation result, so that the at least one service scheme with high quality can be generated efficiently.

S230, determining a target business scheme from at least one business scheme according to the target business scene and the target business requirement.

The target business scene is a scene of target business scheme application, such as medical scene, financial calculation, industrial calculation, game making scene and the like.

In one example, as illustrated in the architecture of fig. 5, determining a target business scenario from at least one business scenario and target business requirements, comprises:

the policy 170 is first determined based on the target traffic scenario, the policy 170 including one or more of a cost to invoke the network resource service, a cost to invoke the computing resource service, a time to provision the service, a performance of the computing resource service, and a performance of the network link. The data processing device 110 for the power network then determines a target business scenario from the at least one business scenario according to the policy 170, the target business scenario, and the target business requirement.

For example, if the target business scenario is a medical scenario, determining the policy 170 includes: the storage resource size of the computing resource service is not less than 50T, the time delay of the network link is less than 50ms, and the like.

Illustratively, determining a target business scenario from at least one business scenario according to a policy, a target business scenario, and a target business requirement comprises:

and determining a target business scheme from at least one business scheme according to the target business scene and the target business requirement. The target business scheme is correspondingly adjusted according to the strategy, so that the target business scheme meets the requirements in the strategy; or,

and selecting a target service scheme which simultaneously meets the requirements of the strategy, the target service scene and the target service requirement from at least one service scheme.

In the above example, on the basis of determining the target service scheme according to the target service scene and the target service requirement, the conditions specified by the policy are added, so that the target service scheme can further adapt to the requirements of different details, the service quality of the computing network is improved, and the satisfaction degree of the user is improved.

In one example, determining a target business scenario from at least one business scenario according to a target business scenario and a target business requirement comprises:

and acquiring the supplementary appeal of the user, carrying out semantic recognition according to the supplementary appeal of the user, and determining the tendency requirement of the user on the related cost of the business scheme and/or the tendency requirement on the related data processing and network response speed. The policy to be executed is determined according to the trend requirement regarding cost and/or the trend requirement regarding data processing and network response speed, and the target business scheme is determined from the at least one business scheme according to the policy to be executed, the target business scenario and the target business requirement. Wherein, the supplementary appeal is the requirement on the detail that the user further puts forward to the target business demand.

The target traffic scenario is illustratively determined from at least one traffic scenario by means of OR-Tools (which is open source software for combinatorial optimization) according to the policies to be performed, the target traffic scenario and the target traffic requirements. The OR-Tools can search the shortest path and the minimum flow according to the position of the user, the position of the target computing power resource service and the computing power of the target computing power resource service, so that a target service scheme meeting the supplementary requirements of the user can be determined.

In one example, the manner in which policies are determined from a target business scenario includes:

firstly, training a neural network model through a service scene and a strategy in a training sample, so that the neural network model learns the corresponding relation between the service scene and the strategy. And then determining a strategy corresponding to the target service scene according to the target service scene based on the neural network model. The strategy can be determined efficiently and accurately by using the neural network model, and labor is saved.

Illustratively, the neural network model employs an artificial intelligence data mining based prior (Apriori) algorithm, or a frequent pattern growth (FP-growth) algorithm.

S240, calling a target computing power resource service and a target network resource service indicated by the target business scheme, wherein the target computing power resource service is contained in at least one computing power resource service, and the target network resource service is contained in at least one network resource service.

In one example, invoking the target computing power resource service and the target network resource service indicated by the target business plan includes:

it is first determined whether a target computing power resource service indicated by a target business plan is located in a plurality of callable geographic areas. If the target computing power resource service is located in the plurality of callable geographic areas, then determining a target callable geographic area from the plurality of callable geographic areas according to the location of the user. And then, calling part of the computing power resource service in the target calling geographic area and the target network resource service indicated by the target business scheme.

Illustratively, the target computing power resource service indicated by the target business scheme is invoked, and is realized by adopting a simple linux utility program (simple linux utility for resource management, slurm) technology for resource management, wherein the Slurm is a highly scalable and fault-tolerant cluster manager and job scheduling system which can be used for large-scale computing node clusters, and resource and job management is performed by utilizing the Slurm so as to avoid mutual interference among invoked servers and improve the operation efficiency.

In the above example, the present application considers that in the process of computing network scheduling, if one computing resource service is deployed across regions, there may be a problem of mutual interference in invoking the computing resource service. To avoid mutual interference, the present application determines a target calling geographic area from a plurality of callable geographic areas according to the location of the user. The manner in which the portion of the computing power resource service located in the target call geographic area in the target computing power resource service is then scheduled to avoid the above-described problems.

According to the data processing method for the power network, which is provided by the embodiment, the interface parameters of at least one power resource service and at least one network resource service are collected uniformly, so that at least one business scheme is determined by combining business requirements. Then, according to the target service scene and the target service requirement, a target service scheme matched with the target service scene and the target service requirement is determined from at least one service scheme, and then, according to the target service scheme, the target computing power resource service and the target network resource service are called. According to the scheme, the computing power resource service and the network resource service of each area can be collected, so that the computing power resource service and the network resource service which can be called can be conveniently determined according to the target service scene and the target service requirement of the user, the purpose that the user can dispatch the proper computing power resource and network resource in a cross-area mode without setting the computing power resource and the network resource is achieved, and the development cost of the user is greatly saved.

In combination with the data processing method of the power computing network, the application also provides a data processing device of the power computing network, which comprises the following steps:

the service scheduling module is used for determining a target service scheme from at least one service scheme according to the target service scene and the target service requirement;

In one example, the apparatus further comprises a resource awareness module:

the resource perception module is used for evaluating the at least one computing power resource service according to the first parameter information and determining a first evaluation result;

the resource perception module is further used for evaluating the at least one network resource service according to the second parameter information and determining a second evaluation result;

the service scheduling module is configured to determine a target service scheme from at least one service scheme according to a target service scenario and a target service requirement, and includes:

And the service scheduling module is used for determining at least one service scheme according to the first parameter information, the second parameter information, the first evaluation result, the second evaluation result and at least one service requirement supported by the power computing network.

In one example, the apparatus further comprises a policy generation module:

the system comprises a policy generation module, a service management module and a network link module, wherein the policy generation module is used for determining a policy according to a target business scene, and the policy comprises one or more of cost for calling network resource service, cost for calling computing resource service, service opening time, computing resource service performance and network link performance;

and the service scheduling module is used for determining a target service scheme from the at least one service scheme according to the strategy, the target service scene and the target service requirement.

Other implementations and effects of the apparatus are described in the data processing method for the power network, and are not described herein.

The basic principles of the present application have been described above in connection with specific embodiments, however, it should be noted that the advantages, benefits, effects, etc. mentioned in the present application are merely examples and not limiting, and these advantages, benefits, effects, etc. are not to be considered as necessarily possessed by the various embodiments of the present application. Furthermore, the specific details disclosed herein are for purposes of illustration and understanding only, and are not intended to be limiting, as the application is not intended to be limited to the details disclosed herein as such.

It should be understood that, although the steps in the flowcharts of the figures are shown in order as indicated by the arrows, these steps are not necessarily performed in order as indicated by the arrows. The steps are not strictly limited in order and may be performed in other orders, unless explicitly stated herein. Moreover, at least some of the steps in the flowcharts of the figures may include a plurality of sub-steps or stages that are not necessarily performed at the same time, but may be performed at different times, the order of their execution not necessarily being sequential, but may be performed in turn or alternately with other steps or at least a portion of the other steps or stages.

The block diagrams of the devices, apparatuses, devices, systems referred to in this application are only illustrative examples and are not intended to require or imply that the connections, arrangements, configurations must be made in the manner shown in the block diagrams. As will be appreciated by one of skill in the art, the devices, apparatuses, devices, systems may be connected, arranged, configured in any manner. Words such as "including," "comprising," "having," and the like are words of openness and mean "including but not limited to," and are used interchangeably therewith. The terms "or" and "as used herein refer to and are used interchangeably with the term" and/or "unless the context clearly indicates otherwise. The term "such as" as used herein refers to, and is used interchangeably with, the phrase "such as, but not limited to.

It is also noted that in the apparatus, devices and methods of the present application, the components or steps may be disassembled and/or assembled. Such decomposition and/or recombination should be considered as equivalent to the present application.

The previous description of the disclosed aspects is provided to enable any person skilled in the art to make or use the present application. Various modifications to these aspects will be readily apparent to those skilled in the art, and the generic principles defined herein may be applied to other aspects without departing from the scope of the application. Thus, the present application is not intended to be limited to the aspects shown herein but is to be accorded the widest scope consistent with the principles and novel features disclosed herein.

The foregoing description has been presented for purposes of illustration and description. Furthermore, this description is not intended to limit the embodiments of the application to the form disclosed herein. Although a number of example aspects and embodiments have been discussed above, a person of ordinary skill in the art will recognize certain variations, modifications, alterations, additions, and subcombinations thereof.

Claims

1. A data processing method for a computing power network, comprising:

2. The method of claim 1, wherein determining at least one business scenario based on the first parameter information, the second parameter information, and at least one business requirement supported by the computing power network comprises:

3. The method according to claim 1 or 2, wherein determining a target business scenario from the at least one business scenario according to the target business scenario and the target business requirement comprises:

4. A method according to claim 3, wherein said determining a policy from a target traffic scenario comprises:

5. The method of claim 1 or 2, wherein the obtaining the first parameter information and the second parameter information of the computing power network comprises:

6. The method of claim 2, wherein evaluating the at least one computing power resource service based on the first parameter information, determining a first evaluation result, comprises:

7. The method of claim 6, wherein evaluating the at least one network resource service based on the second parameter information, determining a second evaluation result, comprises:

8. The method of claim 1, wherein the determining at least one business scenario according to the first parameter information, the second parameter information, the first evaluation result, the second evaluation result, and at least one business requirement supported by the computing power network comprises:

9. The method according to claim 1 or 2, wherein said invoking the target computing power resource service and the target network resource service indicated by the target business plan comprises:

10. A data processing apparatus for a computing network, comprising: