EP-4738797-A1 - SERVICE DEPLOYMENT METHOD, CLOUD MANAGEMENT PLATFORM, AND RELATED DEVICE

EP4738797A1EP 4738797 A1EP4738797 A1EP 4738797A1EP-4738797-A1

Abstract

This application belongs to the field of cloud computing technologies, and discloses a service deployment method, a cloud management platform, and a related apparatus. The method includes: providing a configuration interface, obtaining service configuration information from the configuration interface, determining a resource deployment policy of a target service based on the service configuration information and a number of service copies, and deploying the target service on a plurality of service units according to the resource deployment policy, where the resource deployment policy indicates that one service unit bears at least one user group of a tenant, every Rf service units bear service copies of a same user group, and Rf is the number of service copies. For any user in any user group, even if a service unit that bears the user is faulty, the user can alternatively obtain a corresponding service from another service unit. In this way, a redundancy capability of the target service is effectively improved.

Inventors

WU, ZHENHUA
WANG, Baolin
XU, Wenchang
ZHU, Yelong

Assignees

Huawei Cloud Computing Technologies Co., Ltd.

Dates

Publication Date: 20260506
Application Date: 20240218

Claims (20)

A service deployment method, applied to a cloud management platform, wherein the cloud management platform is configured to manage a cloud infrastructure, the cloud infrastructure comprises at least one cloud data center, a plurality of service units are deployed in the at least one cloud data center, and the method comprises: providing a configuration interface, wherein the configuration interface indicates a tenant to enter service configuration information of a to-be-deployed target service, and the service configuration information is used to configure the target service; obtaining the service configuration information from the configuration interface; and determining a resource deployment policy of the target service based on the service configuration information and a number of service copies, and deploying the target service on the plurality of service units according to the resource deployment policy, wherein the resource deployment policy indicates that one service unit bears at least one user group of the tenant, every Rf service units bear service copies of a same user group, Rf is the number of service copies, and Rf ≥2.
The method according to claim 1, wherein the resource deployment policy indicates that Rf service copies of the user group are respectively deployed on independent service units, and every Rf service units bear service copies of at most one same user group.
The method according to claim 1 or 2, wherein the resource deployment policy comprises a number and a deployment specification of the service units, and the determining the resource deployment policy of the target service based on the service configuration information and the number of service copies comprises: determining the number and the deployment specification of the service units based on the service configuration information, the number of service copies, and a resource deployment constraint condition, wherein the resource deployment constraint condition comprises one or more of the following: a minimum number of startup nodes of a microservice, a minimum startup specification of the microservice, a maximum number of resources, a service performance parameter, a maximum explosion radius, the number of service copies, or a performance baseline of the microservice; and the maximum number of resources indicates an upper limit of resources occupiable by the target service, the service performance parameter indicates a lower limit of performance of the target service, the maximum explosion radius indicates a lower limit of a redundancy capability of the target service, and the performance baseline indicates a correspondence between a performance parameter and each of a startup specification and a number of startup nodes of the microservice.
The method according to claim 3, wherein the determining the number and the deployment specification of the service units based on the service configuration information, the number of service copies, and the resource deployment constraint condition comprises: determining a plurality of candidate resource deployment policies based on the service configuration information, the number of service copies, and the resource deployment constraint condition, wherein the candidate resource deployment policy comprises a candidate number and a candidate deployment specification of the service units; and selecting a candidate resource deployment policy from the plurality of candidate resource deployment policies, and using a candidate number and a candidate deployment specification that are of the service units and that are comprised in the selected candidate resource deployment policy as the number and the deployment specification that are of the service units and that are comprised in the resource deployment policy of the target service.
The method according to claim 4, wherein the candidate deployment specification comprises a number of startup nodes and a startup specification that respectively correspond to at least one microservice used to provide the target service, and the resource deployment constraint condition satisfies the following inequality: min num i ≤ num i min s i ≤ s i s i num i ≤ S ∑ i = 1 n s i num i ≤ S N min t j ≥ T N B ≥ 1 C N Rf num i ∈ X i , s i ∈ X i , j ∈ 1 , … , n wherein i indicates an i th microservice, min num i indicates a minimum number of startup nodes of the i th microservice, num i indicates a number of startup nodes of the i th microservice in the candidate deployment specification, min s i indicates a minimum startup specification of the i th microservice, s i indicates a startup specification of the i th microservice in the candidate deployment specification, S indicates the maximum number of resources, n indicates a number of the at least one microservice, N indicates the candidate number, min t ( j ) indicates a minimum performance parameter corresponding to a startup specification and a number of startup nodes of a j th microservice in the candidate deployment specification, T indicates the service performance parameter, B indicates the maximum explosion radius, Rf indicates the number of service copies, and X i indicates a performance baseline of the i th microservice.
The method according to claim 4, wherein the service configuration information comprises an explosion radius weight and a resource weight; and the selecting the candidate resource deployment policy from the plurality of candidate resource deployment policies comprises: selecting the candidate resource deployment policy from the plurality of candidate resource deployment policies based on the explosion radius weight and the resource weight.
The method according to any one of claims 1 to 6, wherein the method further comprises: obtaining user feature information, wherein the user feature information is feature information of a plurality of users in user groups of the tenant, the feature information comprises risk feature information, and the risk feature information indicates a feature that easily causes a service fault; determining at least one risky user from the plurality of users based on the risk feature information of the plurality of users; grouping the at least one risky user to obtain a risky user group to which the at least one risky user separately belongs; and grouping regular users based on feature information of the regular users in the plurality of users, to obtain regular user groups to which the regular users separately belong, wherein the regular users are users in the plurality of users except the risky user.
The method according to claim 7, wherein the resource deployment policy comprises the number of the service units; the method further comprises: determining a number of user groups based on the number of the service units and the number of service copies; and the grouping the regular users based on the feature information of the regular users in the plurality of users comprises: grouping the regular users based on a number of risky user groups, the number of user groups, and the feature information of the regular users.
The method according to claim 8, wherein the grouping the regular users based on the number of risky user groups, the number of user groups, and the feature information of the regular users comprises: determining feature vectors of the regular users based on the feature information of the regular users; clustering the regular users based on the feature vectors of the regular users, to obtain a plurality of regular user clusters; and determining, based on the plurality of regular user clusters and a number of regular user groups, the regular user groups to which the regular users separately belong, wherein the number of regular user groups is a difference between the number of user groups and the number of risky user groups.
The method according to any one of claims 1 to 9, wherein the number of service copies is entered by the tenant on the configuration interface.
The method according to any one of claims 1 to 10, wherein after determining the resource deployment policy of the target service based on the service configuration information and the number of service copies, the method further comprises: displaying a service deployment suggestion interface, wherein the service deployment suggestion interface is used to display the resource deployment policy of the target service.
The method according to claim 11, wherein the service deployment suggestion interface is further used to display at least one of a resource consumption status and an actual explosion radius, the resource consumption status indicates a number of resources occupied by the resource deployment policy, and the actual explosion radius indicates the redundancy capability of the target service deployed according to the resource deployment policy.
The method according to claim 11 or 12, wherein the service deployment suggestion interface further indicates the tenant to trigger a service deployment request, and before deploying the target service on the plurality of service units according to the resource deployment policy, the method further comprises: receiving the service deployment request, wherein the service deployment request indicates the cloud management platform to deploy the target service according to the resource deployment policy displayed on the service deployment suggestion interface.
A service deployment apparatus, wherein the service deployment apparatus is comprised in a cloud management platform, the cloud management platform is configured to manage a cloud infrastructure, the cloud infrastructure comprises at least one cloud data center, a plurality of service units are deployed in the at least one cloud data center, and the apparatus comprises: a providing module, configured to provide a configuration interface, wherein the configuration interface indicates a tenant to enter service configuration information of a to-be-deployed target service, and the service configuration information is used to configure the target service; a first obtaining module, configured to obtain the service configuration information from the configuration interface; and a first determining module, configured to: determine a resource deployment policy of the target service based on the service configuration information and a number of service copies, and deploy the target service on the plurality of service units according to the resource deployment policy, wherein the resource deployment policy indicates that one service unit bears at least one user group of the tenant, every Rf service units bear service copies of a same user group, Rf is the number of service copies, and Rf ≥2.
The apparatus according to claim 14, wherein the resource deployment policy indicates that Rf service copies of the user group are respectively deployed on independent service units, and every Rf service units bear service copies of at most one same user group.
The apparatus according to claim 14 or 15, wherein the apparatus further comprises: a second obtaining module, configured to obtain user feature information, wherein the user feature information is feature information of a plurality of users in user groups of the tenant, the feature information comprises risk feature information, and the risk feature information indicates a feature that easily causes a service fault; a second determining module, configured to determine at least one risky user from the plurality of users based on the risk feature information of the plurality of users; a first grouping module, configured to group the at least one risky user to obtain a risky user group to which the at least one risky user separately belongs; and a second grouping module, configured to group regular users based on feature information of the regular users in the plurality of users, to obtain regular user groups to which the regular users separately belong, wherein the regular users are users in the plurality of users except the risky user.
The apparatus according to claim 16, wherein the resource deployment policy comprises a number of the service units; the apparatus further comprises: a third determining module, configured to determine a number of user groups based on the number of the service units and the number of service copies; and the second grouping module is specifically configured to: group the regular users based on a number of risky user groups, the number of user groups, and the feature information of the regular users.
The apparatus according to any one of claims 14 to 17, wherein the number of service copies is entered by the tenant on the configuration interface.
The apparatus according to any one of claims 14 to 18, wherein the apparatus further comprises: a display module, configured to display a service deployment suggestion interface, wherein the service deployment suggestion interface is used to display the resource deployment policy of the target service.
A compute device cluster, wherein the compute device cluster comprises at least one compute device, and each compute device comprises a processor and a memory; and the processor of the at least one compute device is configured to execute instructions stored in the memory of the at least one compute device, to enable the compute device cluster to perform the method according to any one of claims 1 to 13.

Description

This application claims priorities to Chinese Patent Application No. 202310911636.0, filed on July 24, 2023 and entitled "ARCHITECTURE DIVISION METHOD, APPARATUS, AND SYSTEM", and to Chinese Patent Application No. 202311541864.X, filed on November 17, 2023 and entitled "SERVICE DEPLOYMENT METHOD, CLOUD MANAGEMENT PLATFORM, AND RELATED APPARATUS", both of which are incorporated herein by reference in their entireties. TECHNICAL FIELD This application relates to the field of cloud computing technologies, and in particular, to a service deployment method, a cloud management platform, and a related apparatus. BACKGROUND With the continuous development of information technologies, especially the internet technologies, large-scale governments, enterprises, and financial institutions start to build own data centers step by step to satisfy service requirements of the large-scale governments, enterprises, and financial institutions for resources. The cost of building a data center by simply adding physical devices is extremely high. Therefore, cloud computing emerges. A cloud service provider can construct a cloud platform by using a cloud computing technology, where the cloud platform includes a cloud management platform and a cloud infrastructure, the cloud management platform is configured to manage the cloud infrastructure, the cloud infrastructure includes at least one cloud data center, and a tenant can deploy a corresponding service in the at least one cloud data center. In a related technology, to ensure service availability, the tenant needs to deploy, in the at least one cloud data center, a plurality of service units that are independent of each other and that have a same service function, and each service unit has a complete service function. In this way, after a service request is received, the service request is routed, according to a specified rule, to a corresponding service unit for processing. For example, a service request of a user is routed, based on a correspondence between the user served by a service and a service unit, to the corresponding service unit for processing. However, one service unit may correspond to a plurality of users. If the service unit is faulty due to reasons such as overload or a distributed denial of service (Distributed Denial of Service, DDoS) attack, the plurality of users corresponding to the service unit cannot use a service, and a redundancy capability of the service is poor. SUMMARY This application provides a service deployment method, a cloud management platform, and a related apparatus, to resolve a problem of a poor redundancy capability of a service in a related technology. The technical solutions are as follows. According to a first aspect, a service deployment method is provided, applied to a cloud management platform, where the cloud management platform is configured to manage a cloud infrastructure, the cloud infrastructure includes at least one cloud data center, a plurality of service units are deployed in the at least one cloud data center, and the method includes: providing a configuration interface, where the configuration interface indicates a tenant to enter service configuration information of a to-be-deployed target service, and the service configuration information is used to configure the target service; obtaining the service configuration information from the configuration interface; and determining a resource deployment policy of the target service based on the service configuration information and a number of service copies, and deploying the target service on the plurality of service units according to the resource deployment policy, where the resource deployment policy indicates that one service unit bears at least one user group of the tenant, every Rf service units bear service copies of a same user group, Rf is the number of service copies, and Rf ≥2. In this application, the resource deployment policy of the target service can be determined based on the service configuration information and the number of service copies that are of the target service and that are entered by the tenant, and then the target service is deployed on the plurality of service units according to the resource deployment policy. The number of service copies is greater than or equal to 2, the resource deployment policy indicates that the service unit bears the at least one user group of the tenant, every Rf service units bear the service copies of the same user group, and Rf is the number of service copies. In other words, the service unit can provide a service for the at least one user group of the tenant, and each user group can be borne by at least two service units. In this case, for any user in any user group, even if a service unit that bears the user is faulty, the user can alternatively obtain a corresponding service from another service unit. In comparison with a method in which users served by a service unit cannot normally obtain a service because the se