WO2021203975A1 - 服务器调配方法、装置、设备及存储介质 - Google Patents

服务器调配方法、装置、设备及存储介质 Download PDF

Info

Publication number
WO2021203975A1
WO2021203975A1 PCT/CN2021/082702 CN2021082702W WO2021203975A1 WO 2021203975 A1 WO2021203975 A1 WO 2021203975A1 CN 2021082702 W CN2021082702 W CN 2021082702W WO 2021203975 A1 WO2021203975 A1 WO 2021203975A1
Authority
WO
WIPO (PCT)
Prior art keywords
server
reliable
task
servers
cloud computing
Prior art date
Application number
PCT/CN2021/082702
Other languages
English (en)
French (fr)
Inventor
黄丽媛
Original Assignee
平安科技(深圳)有限公司
Priority date (The priority date is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the date listed.)
Filing date
Publication date
Application filed by 平安科技(深圳)有限公司 filed Critical 平安科技(深圳)有限公司
Publication of WO2021203975A1 publication Critical patent/WO2021203975A1/zh

Links

Images

Classifications

    • HELECTRICITY
    • H04ELECTRIC COMMUNICATION TECHNIQUE
    • H04LTRANSMISSION OF DIGITAL INFORMATION, e.g. TELEGRAPHIC COMMUNICATION
    • H04L67/00Network arrangements or protocols for supporting network services or applications
    • H04L67/01Protocols
    • H04L67/10Protocols in which an application is distributed across nodes in the network
    • H04L67/1001Protocols in which an application is distributed across nodes in the network for accessing one among a plurality of replicated servers
    • H04L67/1004Server selection for load balancing
    • H04L67/1008Server selection for load balancing based on parameters of servers, e.g. available memory or workload

Definitions

  • This application relates to the field of cloud technology and blockchain technology, and in particular to a server deployment method, device, equipment, and storage medium.
  • the core of cloud computing is to provide users with services in the form of renting computing resources after virtualizing the computing resources of one or several data centers.
  • the function of the cloud computing resource management system is to receive resource requests from cloud computing users, and to encapsulate specific resources into services and provide them to the resource requester.
  • the prior art selects appropriate computing resources from the perspectives of computing power, reliability, and cost factors of the resources. However, no matter from which point of view the resources are selected, only one or more factors are considered, and comprehensive factors are not considered comprehensively. Therefore, the inventor realizes that the prior art is only randomly selected from the physical resources that meet the task's computing capability request. It does not consider the reliability of the resource or the cost factor of the resource, and at the same time, it does not have sufficient computing power but does not The use of reliable resources will result in waste of resources. Choosing the resource with the best computing capability for a task each time will also cause waste of computing resources, ignoring the operating cost of the cloud service provider. That is, the matching degree of existing computing resource adjustment to computing tasks is not high enough.
  • the main purpose of this application is to solve the technical problem that the matching degree between the existing computing resource adjustment and the computing task is not high enough.
  • the first aspect of this application provides a server deployment method, including:
  • a reliable server or server group with the highest deployment profit is selected to execute the cloud computing task.
  • the second aspect of the present application provides a server deployment device, including:
  • An acquisition module used to acquire cloud computing tasks and corresponding task deadlines, and to determine multiple servers for processing the cloud computing tasks
  • the dividing module is configured to divide the multiple servers into reliable servers and unreliable servers by using a preset task failure prediction method according to the task deadline;
  • the matching module is used to sequentially perform task migration and matching between the unreliable servers and the reliable servers to obtain multiple server groups that can be used to complete the cloud computing tasks, wherein the server group includes an unreliable server and A reliable server;
  • a calculation module for calculating the first allocation income of each of the server groups and the second allocation income of each of the reliable servers
  • the deployment module is configured to select a reliable server or server group with the highest deployment profit to execute the cloud computing task based on the first deployment profit and the second deployment profit.
  • a third aspect of the present application provides a computer device, including: a memory and at least one processor, the memory stores instructions; the at least one processor calls the instructions in the memory to make the computer
  • the device executes the steps of the server provisioning method described below:
  • a reliable server or server group with the highest deployment profit is selected to execute the cloud computing task.
  • the fourth aspect of the present application provides a computer-readable storage medium that stores instructions in the computer-readable storage medium, which when run on a computer, causes the computer to execute the steps of the server deployment method described below:
  • a reliable server or server group with the highest deployment profit is selected to execute the cloud computing task.
  • a preset task failure prediction method is adopted, and multiple servers used to process the cloud computing task are divided into reliable servers and unreliable servers; Servers can be matched with corresponding reliable servers to form a server group, so that unreliable servers can still be used; then calculate the deployment revenue of reliable servers and server groups separately, and select the server or server group with the highest deployment revenue to perform cloud computing tasks , Utilize all currently available servers, whether they are reliable or unreliable servers, and only select the server or server combination with the highest deployment revenue to perform the cloud computing task, which can more fully consider the impact of cloud computing tasks on server selection Factors to improve the matching degree between cloud computing tasks and corresponding processing servers.
  • FIG. 1 is a schematic diagram of an embodiment of a server provisioning method in an embodiment of this application
  • FIG. 2 is a schematic diagram of another embodiment of a server provisioning method in an embodiment of the application.
  • FIG. 3 is a schematic diagram of an embodiment of a server provisioning device in an embodiment of the application
  • FIG. 4 is a schematic diagram of another embodiment of the server deployment device in the embodiment of the application.
  • Fig. 5 is a schematic diagram of an embodiment of a computer device in an embodiment of the application.
  • the embodiments of the application provide a server deployment method, device, equipment, and storage medium.
  • this application uses a preset task failure prediction method to divide multiple servers into reliable servers and unreliable servers ; Select an unreliable server and a reliable server in turn for task migration matching, and obtain multiple server groups that can be used to complete cloud computing tasks; calculate the first deployment revenue of each server group separately, and calculate the second deployment of each reliable server separately Income: Based on the first allocation income and the second allocation income, select a reliable server or server group with the highest allocation income to perform cloud computing tasks.
  • This application more comprehensively considers the various factors influencing the selection of servers by cloud computing tasks, and improves the degree of matching between cloud computing tasks and corresponding processing servers.
  • the first embodiment of the server provisioning method in the embodiment of the present application includes:
  • the execution subject of this application may be a server deployment device, or a terminal or a server, which is not specifically limited here.
  • the embodiment of the present application takes the server as the execution subject as an example for description. It should be emphasized that, in order to further ensure the privacy and security of the cloud computing task, the cloud computing task can also be stored in a node of a blockchain.
  • a request when a user request for cloud computing is received, a request can be divided into multiple tasks for parallel processing, and each task is a cloud computing task, and each cloud computing task can be optimally configured using the method of this application.
  • Servers perform calculations; each server acts as a cloud computing cluster (IaaS (Infrastructure as a Service), PaaS (Platform as a Service, platform as a service), SaaS (Software as a Service, software as a service), etc. ) Physical nodes, and store different types of cloud computing resources for performing cloud computing tasks; the task deadline is the completion period of the cloud computing task, which is manually set according to the business development process, where the same user requests to share The task deadlines for multiple cloud computing tasks that come out are the same.
  • a preset task failure prediction method to divide the multiple servers into reliable servers and unreliable servers
  • the preset task failure prediction method is used to predict whether each server fails during the execution of the cloud computing task, that is, the internal server caused by hardware or software defects, design errors, unstable environment, or operational errors. Computing resources are unavailable.
  • the corresponding server is classified as a reliable server, otherwise it is classified as unreliable server. Fuzzy neural network model, hidden Markov model and cloud theory combination, association rule mining and other technologies can be used to configure preset task failure prediction methods.
  • the steps for dividing each server through the preset task failure prediction model are as follows:
  • an association rule mining technology is used to construct a preset task failure prediction method, which is based on the failure history data based on the correlation of the failures in each server, and uses the probability to share the risk to construct the correlation of the model failure.
  • This method uses the association rule mining method that can be executed in parallel to extract frequent items in the data to describe the association of system failures.
  • Reliable servers are servers that are predicted to complete the cloud computing task before the deadline. Unreliable servers cannot be completed before the deadline because of server hardware or software defects, design errors, unstable environments, or operational errors.
  • the process of predicting the completion time of multiple servers for processing the cloud computing task may include the following steps:
  • the preset task failure prediction method is used to respectively predict the completion time of the multiple servers to process the cloud computing task.
  • an unreliable server is selected in order, and then the corresponding reliable server is matched. For those that cannot be matched to a reliable server, it can be eliminated, and the task migration matching is completed until the unreliable servers are matched to the reliable server.
  • task migration matching refers to the process of cloud computing tasks that can be processed by migrating from an unreliable server to a reliable server to complete the cloud computing task, and then the unreliable server and the reliable server are matched.
  • unreliable servers cannot complete the processing of cloud computing tasks on their own. By matching a reliable server to form a server group, unreliable servers can be used for cloud computing tasks.
  • the task migration matching process of the server group may include the following steps:
  • the first deployment benefit takes into account the cost, computing power and reliability of reliable servers in the server group, as well as the cost and computing power of unreliable servers, and calculates the benefits of all aspects of the use of the server, through the first deployment benefit
  • the matching degree between the server group's processing of cloud computing tasks and the three influencing factors is quantified; the specific calculation method is as follows:
  • the second deployment benefit takes into account the cost, computing power and reliability of reliable servers, and is also used to quantify the benefits of a single server in processing cloud computing; the specific calculation method is as follows:
  • c(t 0 ) is the computing power of the unreliable server at t0
  • c(t 1 ) and c(t 2 ) are the computing power of the reliable server at t 1 and t 2 respectively
  • p(t 0 ) is the price of unreliable server's unit computing power at t 0 and t 1
  • w(t 0 ) is the load at t 0
  • w(t 1 ) is the load at t 1
  • ⁇ , ⁇ , ⁇ , ⁇ , ⁇ , and ⁇ are adjustment factors.
  • the server group and a single reliable server are simultaneously used as alternatives for performing cloud computing tasks.
  • the server group or the single reliable server with the largest deployment revenue is considered server cost and reliability. After the performance and computing power, the choice that best matches the cloud computing task is obtained.
  • a preset task failure prediction method is adopted, and multiple servers used for processing the cloud computing task are divided into reliable servers and unreliable servers; then, for unreliable servers, It can match the corresponding reliable servers to form a server group, so that unreliable servers can still be used; then calculate the deployment revenue of reliable servers and server groups separately, and select the server or server group with the highest revenue to perform cloud computing tasks. All currently available servers, whether reliable or unreliable, are used. Only the server or server combination with the highest deployment revenue is selected to perform the cloud computing task, which can more comprehensively consider the various factors affecting the selection of the server by the cloud computing task. Improve the matching degree between cloud computing tasks and corresponding processing servers.
  • the second embodiment of the server deployment method in the embodiment of the present application includes:
  • a preset task failure prediction method to divide the multiple servers into reliable servers and unreliable servers.
  • the unreliable server and the reliable server are combined to perform cloud computing tasks.
  • the second cost factor is different due to different calculation stages, and its pricing is also different.
  • the estimation of the first cost factor here refers to the calculation cost of unreliable servers from the start time of cloud computing to the migration time node
  • the estimation of the second cost factor refers to the migration From the time node to the task completion time, the calculation cost of a reliable server.
  • the reliability of the unreliable server is low, it is outside the evaluation range of the cloud computing task, so its reliability is not considered.
  • the corresponding first initial deployment income is based on two factors of its cost and computing capacity. Calculation to evaluate the first initial deployment benefits of unreliable servers when performing cloud computing tasks in the previous period (before the migration time node), and to measure the cost and computing power of unreliable servers; while the second initial deployment benefits are It is calculated based on the cost and computing power of reliable servers to evaluate the second initial deployment benefits of reliable servers when performing cloud computing tasks at a later stage (after the migration time node), and to measure the cost-effectiveness of reliable servers in terms of cost and computing power.
  • the first initial deployment benefit and the second initial deployment benefit respectively quantify the cost performance of an unreliable server performing cloud computing tasks before the migration time node, and the cost performance of a reliable server performing cloud computing tasks after the migration time node. Add them together to get the overall cost performance of the server group when performing cloud computing tasks, that is, the first deployment benefit.
  • the reliability of different reliable servers is different.
  • the reliability factor of each reliable server can be calculated according to the completion time of the cloud computing task predicted by the task failure prediction method by each reliable server. Factor to describe the reliability of a reliable server.
  • the cost performance of a single reliability factor to perform cloud computing tasks is quantified through the second deployment benefit, which is linked to the three factors of cloud computing task cost, computing power, and reliability, and calculated using the aforementioned formula.
  • the first cost factor of the unreliable server, the second cost factor and the computing power of the reliable server in each server group are used to calculate the first deployment benefit of each server group for the execution of cloud computing tasks.
  • Reliability factor, second cost factor and computing power of reliable servers calculate the second deployment benefit of each reliable server, and take into account the reliability, cost, computing power and other factors of all servers, and deploy the most matching server or server group for cloud computing Tasks are processed.
  • An embodiment of the server provisioning apparatus in the embodiment of the present application includes:
  • the obtaining module 301 is configured to obtain cloud computing tasks and corresponding task deadlines, and determine multiple servers for processing the cloud computing tasks;
  • the dividing module 302 is configured to divide the multiple servers into reliable servers and unreliable servers by using a preset task failure prediction method according to the task deadline;
  • the matching module 303 is configured to perform task migration matching between the unreliable servers and the reliable servers in sequence to obtain multiple server groups that can be used to complete the cloud computing tasks, wherein the server group includes an unreliable server With a reliable server;
  • the calculation module 304 is configured to calculate the first allocation income of each of the server groups and the second allocation income of each of the reliable servers respectively;
  • the deployment module 305 is configured to select a reliable server or server group with the highest deployment profit to execute the cloud computing task based on the first deployment profit and the second deployment profit.
  • a preset task failure prediction method is adopted, and multiple servers used for processing the cloud computing task are divided into reliable servers and unreliable servers; then, for unreliable servers, It can match the corresponding reliable servers to form a server group, so that unreliable servers can still be used; then calculate the deployment revenue of reliable servers and server groups separately, and select the server or server group with the highest revenue to perform cloud computing tasks. All currently available servers, whether reliable servers or unreliable servers, are used. Only the server or server combination with the highest deployment revenue is selected to perform the cloud computing task, which can more comprehensively consider the various factors affecting the selection of the server by the cloud computing task. Improve the matching degree between cloud computing tasks and corresponding processing servers.
  • FIG. 4 another embodiment of the server configuration device in the embodiment of the present application includes:
  • the obtaining module 301 is configured to obtain cloud computing tasks and corresponding task deadlines, and determine multiple servers for processing the cloud computing tasks;
  • the dividing module 302 is configured to divide the multiple servers into reliable servers and unreliable servers by using a preset task failure prediction method according to the task deadline;
  • the matching module 303 is configured to perform task migration matching between the unreliable servers and the reliable servers in sequence to obtain multiple server groups that can be used to complete the cloud computing tasks, wherein the server group includes an unreliable server With a reliable server;
  • the calculation module 304 is configured to calculate the first allocation income of each of the server groups and the second allocation income of each of the reliable servers respectively;
  • the deployment module 305 is configured to select a reliable server or server group with the highest deployment profit to execute the cloud computing task based on the first deployment profit and the second deployment profit.
  • the dividing module 302 includes:
  • the prediction unit 3021 is configured to use the preset task failure prediction method to respectively predict the completion time of the multiple servers to process the cloud computing task;
  • the judging unit 3022 is configured to sequentially judge whether each of the completion times is before the deadline of the task;
  • the dividing unit 3023 is configured to classify the corresponding server as a reliable server if the completion time is before the task deadline, otherwise classify the corresponding server as an unreliable server.
  • the prediction unit 3021 includes:
  • the determining subunit 30211 is configured to determine the load of the cloud computing task and determine the computing capability of each server;
  • the prediction sub-unit 30212 is configured to use the preset task failure prediction method based on the load and the computing capability to respectively predict the completion time of the cloud computing tasks processed by the multiple servers.
  • the matching module 303 includes:
  • the selecting unit 3031 is used to select an unreliable server and a reliable server in sequence, and respectively determine the computing capabilities of the two;
  • the first calculation unit 3032 is configured to calculate the migration time node of the cloud computing task from the unreliable server to the reliable server based on the computing capabilities of the two and the load amount;
  • the judging unit 3033 is used to judge whether the migration time node is before the preset time node;
  • the matching unit 3034 is configured to, if the migration time node is before the preset time node, pair the corresponding unreliable server and the reliable server to obtain multiple corresponding server groups.
  • the calculation module 304 includes a second calculation unit 3041, and the second calculation unit 3041 is configured to:
  • the first deployment benefit of each of the server groups is determined.
  • the calculation module 304 further includes a third calculation unit 3042, and the third calculation unit 3042 is configured to:
  • the second deployment benefit of each reliable server is calculated respectively.
  • FIG. 5 is a schematic structural diagram of a computer device provided by an embodiment of the present application.
  • the computer device 500 may have relatively large differences due to different configurations or performance, and may include one or more processors (central processing units, CPU) 510 (For example, one or more processors) and a memory 520, and one or more storage media 530 (for example, one or more storage devices in a large amount) storing application programs 533 or data 532.
  • the memory 520 and the storage medium 530 may be short-term storage or persistent storage.
  • the program stored in the storage medium 530 may include one or more modules (not shown in the figure), and each module may include a series of instruction operations on the computer device 500.
  • the processor 510 may be configured to communicate with the storage medium 530 and execute a series of instruction operations in the storage medium 530 on the computer device 500.
  • the computer device 500 may also include one or more power supplies 540, one or more wired or wireless network interfaces 550, one or more input and output interfaces 560, and/or one or more operating systems 531, such as Windows Serve, Mac OS X, Unix, Linux, FreeBSD, etc.
  • operating systems 531 such as Windows Serve, Mac OS X, Unix, Linux, FreeBSD, etc.
  • the present application also provides a computer device that includes a memory and a processor, and computer-readable instructions are stored in the memory.
  • the processor executes the steps described in the foregoing embodiments. The steps of the server provisioning method.
  • the computer device can be any device that performs server deployment processing.
  • the present application also provides a computer-readable storage medium.
  • the computer-readable storage medium may be a non-volatile computer-readable storage medium, and the computer-readable storage medium may also be a volatile computer-readable storage medium.
  • the computer-readable storage medium stores instructions, and when the instructions run on a computer, the computer executes the steps of the server deployment method.
  • the integrated unit is implemented in the form of a software functional unit and sold or used as an independent product, it can be stored in a computer readable storage medium.
  • the technical solution of the present application essentially or the part that contributes to the existing technology or all or part of the technical solution can be embodied in the form of a software product, and the computer software product is stored in a storage medium , Including several instructions to make a computer device (which may be a personal computer, a server, or a network device, etc.) execute all or part of the steps of the methods described in the various embodiments of the present application.
  • the aforementioned storage media include: U disk, mobile hard disk, read-only memory (read-only memory, ROM), random access memory (random access memory, RAM), magnetic disks or optical disks and other media that can store program codes. .
  • the blockchain referred to in this application is a new application mode of computer technology such as distributed data storage, point-to-point transmission, consensus mechanism, and encryption algorithm.
  • Blockchain essentially a decentralized database, is a series of data blocks associated with cryptographic methods. Each data block contains a batch of network transaction information for verification. The validity of the information (anti-counterfeiting) and the generation of the next block.
  • the blockchain can include the underlying platform of the blockchain, the platform product service layer, and the application service layer.

Landscapes

  • Engineering & Computer Science (AREA)
  • Computer Hardware Design (AREA)
  • General Engineering & Computer Science (AREA)
  • Computer Networks & Wireless Communication (AREA)
  • Signal Processing (AREA)
  • Hardware Redundancy (AREA)

Abstract

本申请涉及云技术领域和区块链技术领域,公开了一种服务器调配方法、装置、设备及存储介质。该方法包括:获取云计算任务及对应的任务截止时间,并确定用于处理云计算任务的多个服务器;根据云计算任务的任务截止时间,采用预置任务失效预测方法,将多个服务器划分为可靠服务器和不可靠服务器;对各不可靠服务器与各可靠服务器进行任务迁移匹配,得到可用于完成云计算任务的多个服务器组;分别计算各服务器组的第一调配收益和各可靠服务器的第二调配收益;基于第一调配收益和第二调配收益,选取一个调配收益最高的可靠服务器或服务器组执行云计算任务。本申请考虑了云计算任务对服务器选择的各影响因素,提升了云计算任务与对应处理服务器的匹配程度。

Description

服务器调配方法、装置、设备及存储介质
本申请要求于2020年11月11日提交中国专利局、申请号为202011251978.7、发明名称为“服务器调配方法、装置、设备及存储介质”的中国专利申请的优先权,其全部内容通过引用结合在申请中。
技术领域
本申请涉及云技术领域和区块链技术领域,尤其涉及一种服务器调配方法、装置、设备及存储介质。
背景技术
云计算的核心是将某一个或某几个数据中心的计算资源虚拟化之后,向用户提供以租用计算资源为形式的服务。云计算资源管理系统的功能是接收来自云计算用户的资源请求,并且把特定的资源封装成服务提供给资源请求者。
现有技术从资源的计算能力、可靠性、成本因素等角度选择适当的计算资源。但是不论从哪个角度去选择资源,都只考虑了某一个或多个因素,并没有综合考虑全面因素。因此,发明人意识到,现有技术只是在满足任务计算能力请求的物理资源中随机选择的,既没有考虑资源的可靠性,也没有考虑资源的成本因素,同时没有将计算能力充足但并不可靠的资源加以利用,会造成资源的浪费,每次为任务选择计算能力最好的资源也会造成计算资源的浪费,忽略了云服务提供者的运营成本因素。即现有计算资源调配对计算任务的匹配度不够高。
发明内容
本申请的主要目的在于解决现有计算资源调配对计算任务的匹配度不够高的技术问题。
本申请第一方面提供了一种服务器调配方法,包括:
获取云计算任务及对应的任务截止时间,并确定用于处理所述云计算任务的多个服务器;
根据所述任务截止时间,采用预置任务失效预测方法,将所述多个服务器划分为可靠服务器和不可靠服务器;
依次对所述各不可靠服务器与所述各可靠服务器进行任务迁移匹配,得到可用于完成所述云计算任务的多个服务器组,其中所述服务器组包含一个不可靠服务器与一个可靠服务器;
分别计算各所述服务器组的第一调配收益,以及分别计算各所述可靠服务器的第二调配收益;
基于所述第一调配收益和所述第二调配收益,选取一个调配收益最高的可靠服务器或 服务器组执行所述云计算任务。
本申请第二方面提供了一种服务器调配装置,包括:
获取模块,用于获取云计算任务及对应的任务截止时间,并确定用于处理所述云计算任务的多个服务器;
划分模块,用于根据所述任务截止时间,采用预置任务失效预测方法,将所述多个服务器划分为可靠服务器和不可靠服务器;
匹配模块,用于依次对所述各不可靠服务器与所述各可靠服务器进行任务迁移匹配,得到可用于完成所述云计算任务的多个服务器组,其中所述服务器组包含一个不可靠服务器与一个可靠服务器;
计算模块,用于分别计算各所述服务器组的第一调配收益,以及分别计算各所述可靠服务器的第二调配收益;
调配模块,用于基于所述第一调配收益和所述第二调配收益,选取一个调配收益最高的可靠服务器或服务器组执行所述云计算任务。
本申请第三方面提供了一种计算机设备,包括:存储器和至少一个处理器,所述存储器中存储有指令;所述至少一个处理器调用所述存储器中的所述指令,以使得所述计算机设备执行如下所述的服务器调配方法的步骤:
获取云计算任务及对应的任务截止时间,并确定用于处理所述云计算任务的多个服务器;
根据所述任务截止时间,采用预置任务失效预测方法,将所述多个服务器划分为可靠服务器和不可靠服务器;
依次对所述各不可靠服务器与所述各可靠服务器进行任务迁移匹配,得到可用于完成所述云计算任务的多个服务器组,其中所述服务器组包含一个不可靠服务器与一个可靠服务器;
分别计算各所述服务器组的第一调配收益,以及分别计算各所述可靠服务器的第二调配收益;
基于所述第一调配收益和所述第二调配收益,选取一个调配收益最高的可靠服务器或服务器组执行所述云计算任务。
本申请的第四方面提供了一种计算机可读存储介质,所述计算机可读存储介质中存储有指令,当其在计算机上运行时,使得计算机执行如下所述的服务器调配方法的步骤:
获取云计算任务及对应的任务截止时间,并确定用于处理所述云计算任务的多个服务器;
根据所述任务截止时间,采用预置任务失效预测方法,将所述多个服务器划分为可靠服务器和不可靠服务器;
依次对所述各不可靠服务器与所述各可靠服务器进行任务迁移匹配,得到可用于完成所述云计算任务的多个服务器组,其中所述服务器组包含一个不可靠服务器与一个可靠服 务器;
分别计算各所述服务器组的第一调配收益,以及分别计算各所述可靠服务器的第二调配收益;
基于所述第一调配收益和所述第二调配收益,选取一个调配收益最高的可靠服务器或服务器组执行所述云计算任务。
本申请提供的技术方案中,通过云计算任务的任务截止时间,采用预置任务失效预测方法,将用于处理该云计算任务的多个服务器分为可靠服务器和不可靠服务器;然后对于不可靠服务器,可为其匹配相应的可靠服务器,组成服务器组,使得不可靠服务器依然可以被利用;接着分别计算可靠服务器和服务器组的调配收益,并选择调配收益最高的服务器或服务器组执行云计算任务,对当前所有的可用服务器,无论是可靠服务器还是不可靠服务器均进行利用,只选取调配收益最高的服务器或服务器组合执行该云计算任务,可更全面地考虑云计算任务对服务器选择的各影响因素,提升云计算任务与对应处理服务器的匹配程度。
附图说明
图1为本申请实施例中服务器调配方法的一个实施例示意图;
图2为本申请实施例中服务器调配方法的另一个实施例示意图;
图3为本申请实施例中服务器调配装置的一个实施例示意图;
图4为本申请实施例中服务器调配装置的另一个实施例示意图;
图5为本申请实施例中计算机设备的一个实施例示意图。
具体实施方式
本申请实施例提供了一种服务器调配方法、装置、设备及存储介质,本申请根据云计算任务的任务截止时间,采用预置任务失效预测方法,将多个服务器划分为可靠服务器和不可靠服务器;依次选取一个不可靠服务器与一个可靠服务器进行任务迁移匹配,得到可用于完成云计算任务的多个服务器组;分别计算各服务器组的第一调配收益,以及分别计算各可靠服务器的第二调配收益;基于第一调配收益和第二调配收益,选取一个调配收益最高的可靠服务器或服务器组执行云计算任务。本申请更全面地考虑了云计算任务对服务器选择的各影响因素,提升了云计算任务与对应处理服务器的匹配程度。
本申请的说明书和权利要求书及上述附图中的术语“第一”、“第二”、“第三”、“第四”等(如果存在)是用于区别类似的对象,而不必用于描述特定的顺序或先后次序。应该理解这样使用的数据在适当情况下可以互换,以便这里描述的实施例能够以除了在这里图示或描述的内容以外的顺序实施。此外,术语“包括”或“具有”及其任何变形,意图在于覆盖不排他的包含,例如,包含了一系列步骤或单元的过程、方法、系统、产品或设备不必限于清楚地列出的那些步骤或单元,而是可包括没有清楚地列出的或对于这些过 程、方法、产品或设备固有的其它步骤或单元。
为便于理解,下面对本申请实施例的具体流程进行描述,请参阅图1,本申请实施例中服务器调配方法的第一个实施例包括:
101、获取云计算任务及对应的任务截止时间,并确定用于处理所述云计算任务的多个服务器;
可以理解的是,本申请的执行主体可以为服务器调配装置,还可以是终端或者服务器,具体此处不做限定。本申请实施例以服务器为执行主体为例进行说明。需要强调的是,为进一步保证上述云计算任务的私密和安全性,上述云计算任务还可以存储于一区块链的节点中。
本实施例中,当收到用户请求进行云计算时,一个请求可分为多个任务并行处理,每一个任务即为一个云计算任务,各个云计算任务都可应用本申请方法调配最适的服务器执行计算;每个服务器作为云计算集群中(IaaS(Infrastructure as a Service,基础设施即服务)、PaaS(Platform as a Service,平台即服务)、SaaS(Software as a Service,软件即服务)等)的物理节点,并存储有不同类型的云计算资源,以用于执行云计算任务;任务截止时间即该云计算任务的完成期限,根据业务开发流程进行人工设置,其中,同一个用户请求分出来的多个云计算任务的任务截止时间相同。
102、根据所述任务截止时间,采用预置任务失效预测方法,将所述多个服务器划分为可靠服务器和不可靠服务器;
本实施例中,预置任务失效预测方法用于预测各服务器在执行该云计算任务时是否发生失效的情况,即由硬件或软件缺陷、设计错误、环境不稳定或操作失误而引起的服务器内计算资源不可用。此处可通过预先设置条件,比如服务器处理云计算任务的预估完成时间在任务截止时间之前,即可在任务截止时间之前完成云计算任务,对应的服务器归为可靠服务器,否则归为不可靠服务器。可采用模糊神经网络模型、隐马尔可夫模型和云理论结合、关联规则挖掘等技术进行预置任务失效预测方法的配置。
具体的,通过预置任务失效预测模型对各服务器进行划分的步骤如下所示:
(1)采用所述预置任务失效预测方法,分别预测所述多个服务器处理所述云计算任务的完成时间;
(2)依次判断各所述完成时间是否在所述任务截止时间之前;
(3)若是,则将对应服务器划分为可靠服务器,否则将对应服务器划分为不可靠服务器。
本实施例中,比如采用关联规则挖掘技术构建预置任务失败预测方法,其针对各服务器中失效呈现出的关联性,基于失效历史数据,利用概率共享风险组建模型失效的关联性。该方法利用可并行化执行的关联规则挖掘方法,提取数据中的频繁项以描述系统失效的关联性。而可靠服务器即在截止时间前预测可完成该云计算任务的服务器,不可靠服务器因为服务器硬件或软件缺陷、设计错误、环境不稳定或操作失误,故不可在截止时间前完成。
其中,多个服务器处理所述云计算任务的完成时间的预测过程可以包括以下步骤:
(1)确定所述云计算任务的负载量,以及确定各所述服务器的计算能力;
(2)基于所述负载量和所述计算能力,采用所述预置任务失效预测方法,分别预测所述多个服务器处理所述云计算任务的完成时间。
本实施例中,各服务器处理云计算任务的完成时间的计算公式如下所示:C(t 1)*(t 0-t 1)=w(t 1),其中,C(t 1)为当前时刻t 1服务器的计算能力,t 0为任务截止时间,C(t 1)为当前时刻t 1的负载量;
103、依次对所述各不可靠服务器与所述各可靠服务器进行任务迁移匹配,得到可用于完成所述云计算任务的多个服务器组,其中所述服务器组包含一个不可靠服务器与一个可靠服务器;
本实施例中,顺序选取一个不可靠服务器,然后为其匹配对应的可靠服务器,对于无法匹配到可靠服务器的,将其剔除即可,并直到不可靠服务器均匹配到可靠服务器,完成任务迁移匹配,其中,任务迁移匹配即云计算任务的处理可以通过从不可靠服务器迁移至可靠服务器,以完成其云计算任务,则对该不可靠服务器和可靠服务器进行匹配的过程。本来不可靠服务器无法独自完成云计算任务的处理,通过匹配一个可靠服务器,组成服务器组,使不可靠服务器得以被利用于云计算任务。
具体的,服务器组的任务迁移匹配过程可以包括以下步骤:
(1)依次选取一个不可靠服务器和一个可靠服务器,并分别确定两者的计算能力;
(2)基于所述两者的计算能力和所述负载量,计算所述云计算任务从所述不可靠服务器迁移到所述可靠服务器的迁移时间节点;
(3)判断所述迁移时间节点是否在预置时间节点之前;
(4)若是,则将对应的不可靠服务器和可靠服务器进行配对,得到对应的多个服务器组。
本实施例中,
104、分别计算各所述服务器组的第一调配收益,以及分别计算各所述可靠服务器的第二调配收益;
本实施例中,第一调配收益考虑了服务器组中可靠服务器的成本、计算能力和可靠性,以及不可靠服务器的成本和计算能力,对服务器的各方面使用效益进行计算,通过第一调配收益量化了服务器组处理云计算任务与该三个影响因素的匹配程度;具体计算方式如下所示:
(1)不可靠服务器调配收益:crp1=λ+(1-λ)×c(t 1)×p(t 0);
(2)可靠服务器调配收益:
Figure PCTCN2021082702-appb-000001
(3)第一调配收益:rb 1=(crp 1+crp 2);
而第二调配收益考虑了可靠服务器的成本、计算能力和可靠性,亦用于量化单个服务器处理云计算的效益;具体计算方式如下所示:
(1)
Figure PCTCN2021082702-appb-000002
(2)rb 2=1/(crp 3);
其中,c(t 0)为不可靠服务器t0时的计算能力,c(t 1)、c(t 2)分别为可靠服务器t 1、t 2时的计算能力,p(t 0)、p(t 1)为不可靠服务器在t 0、t 1时的单位计算能力的价格,w(t 0)为t 0时的负载量,w(t 1)为t 1时的负载量,λ、σ、α、β、γ、δ为调节因子。
105、基于所述第一调配收益和所述第二调配收益,选取一个调配收益最高的可靠服务器或服务器组执行所述云计算任务。
本实施例中,将服务器组和单个可靠服务器同时作为执行云计算任务的备选,第一调配收益和第二调配收益中,最大调配收益的服务器组或单个可靠服务器即为考虑服务器成本、可靠性、计算能力后,得到的与云计算任务最匹配的选择。
本申请实施例中,通过云计算任务的任务截止时间,采用预置任务失效预测方法,将用于处理该云计算任务的多个服务器分为可靠服务器和不可靠服务器;然后对于不可靠服务器,可为其匹配相应的可靠服务器,组成服务器组,使得不可靠服务器依然可以被利用;接着分别计算可靠服务器和服务器组的调配收益,并选择调配收益最高的服务器或服务器组执行云计算任务,对当前所有的可用服务器,无论是可靠服务器还是不可靠服务器均进行利用,只选取调配收益最高的服务器或服务器组合执行该云计算任务,可更全面地考虑云计算任务对服务器选择的各影响因素,提升云计算任务与对应处理服务器的匹配程度。
请参阅图2,本申请实施例中服务器调配方法的第二个实施例包括:
201、获取云计算任务及对应的任务截止时间,并确定用于处理所述云计算任务的多个服务器;
202、根据所述任务截止时间,采用预置任务失效预测方法,将所述多个服务器划分为可靠服务器和不可靠服务器;
203、依次对所述各不可靠服务器与所述各可靠服务器进行任务迁移匹配,得到可用于完成所述云计算任务的多个服务器组,其中所述服务器组包含一个不可靠服务器与一个可靠服务器;
204、分别确定各所述服务器组中不可靠服务器的第一成本因子及可靠服务器的第二成本因子;
本实施例中,对于不可靠服务器,虽然因为前面说到的由硬件或软件缺陷、设计错误、环境不稳定或操作失误而引起的服务器内计算资源不可用,但可能其计算能力和成本优势依然存在,比如说计算能力,只是在某一环节出现了问题,而在前期计算中计算能力仍然 很高,如果不加以利用,一方面白白浪费资源,另一方面使其他的可靠服务器承受更多的计算压力,最后可能无法使云计算任务顺利完成。
故此处将不可靠服务器与可靠服务器进行组合用于执行云计算任务,而为选择性价比最高的服务器或服务器组执行云计算任务,先从数据库获取当前不可靠服务器的第一成本因子和可靠服务器的第二成本因子。成本因子因计算阶段不同,其定价亦不同,此处的第一成本因子的预估指从云计算开始时间到迁移时间节点,不可靠服务器的计算成本,第二成本因子的预估指从迁移时间节点到任务完成时间,可靠服务器的计算成本。
205、基于所述第一成本因子与不可靠服务器的计算能力,分别计算各所述服务器组在所述迁移时间节点之前的第一初始调配收益;
206、基于所述第二成本因子与可靠服务器的计算能力,分别计算各所述服务器组在所述迁移时间节点之后的第二初始调配收益;
本实施例中,不可靠服务器因为其可靠性能较低,在云计算任务的评估范围之外,故不考虑其可靠性,其对应的第一初始调配收益依据其成本与计算能力两个因素进行计算,以评估不可靠服务器在执行前段(迁移时间节点之前的)云计算任务时的第一初始调配收益,衡量不可靠服务器在成本与计算能力上的优劣程度;而第二初始调配收益则是根据可靠服务器的成本与计算能力进行计算,以评估可靠服务器在执行后段(迁移时间节点之后的)云计算任务时的第二初始调配收益,衡量可靠服务器在成本和计算能力上的性价比。
207、基于所述第一初始调配收益和所述第二初始调配收益,确定各所述服务器组的第一调配收益;
本实施例中,第一初始调配收益和第二初始调配收益分别量化了在迁移时间节点之前不可靠服务器执行云计算任务的性价比,以及在迁移时间节点之后可靠服务器执行云计算任务的性价比,两者相加,即可得到服务器组执行云计算任务时的整体性价比,即第一调配收益。
208、基于各所述可靠服务器的完成时间,分别计算各所述可靠服务器的可靠性因子;
209、基于所述可靠性因子与所述第二成本因子、可靠服务器的计算能力,分别计算各所述可靠服务器的第二调配收益;
本实施例中,不同的可靠服务器的可靠程度均不同,此处可根据各可靠服务器通过任务失效预测方法预测到的云计算任务的完成时间,来计算各可靠服务器的可靠性因子,通过可靠性因子来描述可靠服务器的可靠程度。
然后对于单个可靠性因子执行云计算任务的性价比,通过第二调配收益进行量化,将其与云计算任务的成本、计算能力、可靠程度三个因素挂钩,并采用前面提到公式进行计算。
210、基于所述第一调配收益和所述第二调配收益,选取一个调配收益最高的可靠服务器或服务器组执行所述云计算任务。
本申请实施例中,介绍了通过各服务器组中不可靠服务器的第一成本因子、可靠服务 器的第二成本因子和计算能力,计算各服务器组对执行云计算任务的第一调配收益,通过各可靠服务器的可靠因子、第二成本因子和计算能力,计算各可靠服务器的第二调配收益,同时兼顾全部服务器的可靠性、成本、计算能力等因素,调配最匹配的服务器或服务器组对云计算任务进行处理。
上面对本申请实施例中服务器调配方法进行了描述,下面对本申请实施例中服务器调配装置进行描述,请参阅图3,本申请实施例中服务器调配装置一个实施例包括:
获取模块301,用于获取云计算任务及对应的任务截止时间,并确定用于处理所述云计算任务的多个服务器;
划分模块302,用于根据所述任务截止时间,采用预置任务失效预测方法,将所述多个服务器划分为可靠服务器和不可靠服务器;
匹配模块303,用于依次对所述各不可靠服务器与所述各可靠服务器进行任务迁移匹配,得到可用于完成所述云计算任务的多个服务器组,其中所述服务器组包含一个不可靠服务器与一个可靠服务器;
计算模块304,用于分别计算各所述服务器组的第一调配收益,以及分别计算各所述可靠服务器的第二调配收益;
调配模块305,用于基于所述第一调配收益和所述第二调配收益,选取一个调配收益最高的可靠服务器或服务器组执行所述云计算任务。
本申请实施例中,通过云计算任务的任务截止时间,采用预置任务失效预测方法,将用于处理该云计算任务的多个服务器分为可靠服务器和不可靠服务器;然后对于不可靠服务器,可为其匹配相应的可靠服务器,组成服务器组,使得不可靠服务器依然可以被利用;接着分别计算可靠服务器和服务器组的调配收益,并选择调配收益最高的服务器或服务器组执行云计算任务,对当前所有的可用服务器,无论是可靠服务器还是不可靠服务器均进行利用,只选取调配收益最高的服务器或服务器组合执行该云计算任务,可更全面地考虑云计算任务对服务器选择的各影响因素,提升云计算任务与对应处理服务器的匹配程度。
请参阅图4,本申请实施例中服务器调配装置的另一个实施例包括:
获取模块301,用于获取云计算任务及对应的任务截止时间,并确定用于处理所述云计算任务的多个服务器;
划分模块302,用于根据所述任务截止时间,采用预置任务失效预测方法,将所述多个服务器划分为可靠服务器和不可靠服务器;
匹配模块303,用于依次对所述各不可靠服务器与所述各可靠服务器进行任务迁移匹配,得到可用于完成所述云计算任务的多个服务器组,其中所述服务器组包含一个不可靠服务器与一个可靠服务器;
计算模块304,用于分别计算各所述服务器组的第一调配收益,以及分别计算各所述 可靠服务器的第二调配收益;
调配模块305,用于基于所述第一调配收益和所述第二调配收益,选取一个调配收益最高的可靠服务器或服务器组执行所述云计算任务。
具体的,所述划分模块302包括:
预测单元3021,用于采用所述预置任务失效预测方法,分别预测所述多个服务器处理所述云计算任务的完成时间;
判别单元3022,用于依次判断各所述完成时间是否在所述任务截止时间之前;
划分单元3023,用于若所述完成时间在所述任务截止时间之前,则将对应服务器划分为可靠服务器,否则将对应服务器划分为不可靠服务器。
具体的,所述预测单元3021包括:
确定子单元30211,用于确定所述云计算任务的负载量,以及确定各所述服务器的计算能力;
预测子单元30212,用于基于所述负载量和所述计算能力,采用所述预置任务失效预测方法,分别预测所述多个服务器处理所述云计算任务的完成时间。
具体的,所述匹配模块303包括:
选取单元3031,用于依次选取一个不可靠服务器和一个可靠服务器,并分别确定两者的计算能力;
第一计算单元3032,用于基于所述两者的计算能力和所述负载量,计算所述云计算任务从所述不可靠服务器迁移到所述可靠服务器的迁移时间节点;
判别单元3033,用于判断所述迁移时间节点是否在预置时间节点之前;
匹配单元3034,用于若所述迁移时间节点在预置时间节点之前,则将对应的不可靠服务器和可靠服务器进行配对,得到对应的多个服务器组。
具体的,所述计算模块304包括第二计算单元3041,所述第二计算单元3041用于:
分别确定各所述服务器组中不可靠服务器的第一成本因子及可靠服务器的第二成本因子;
基于所述第一成本因子与不可靠服务器的计算能力,分别计算各所述服务器组在所述迁移时间节点之前的第一初始调配收益;
基于所述第二成本因子与可靠服务器的计算能力,分别计算各所述服务器组在所述迁移时间节点之后的第二初始调配收益;
基于所述第一初始调配收益和所述第二初始调配收益,确定各所述服务器组的第一调配收益。
具体的,所述计算模块304还包括第三计算单元3042,所述第三计算单元3042用于:
基于各所述可靠服务器的完成时间,分别计算各所述可靠服务器的可靠性因子;
基于所述可靠性因子与所述第二成本因子、可靠服务器的计算能力,分别计算各所述可靠服务器的第二调配收益。
本申请实施例中,不仅介绍了采用预置任务失效预测方法,将用于处理该云计算任务的多个服务器分为可靠服务器和不可靠服务器,对于不可靠服务器,可为其匹配相应的可靠服务器,组成服务器组,使得不可靠服务器依然可以被利用;接着通过调配收益量化服务器与云计算任务的匹配程度,对当前所有的可用服务器进行利用,可调配更适合的服务器或服务器组合执行该云计算任务,可更全面地考虑云计算任务对服务器选择的各影响因素,提升云计算任务与对应处理服务器的匹配程度;还介绍了通过各服务器组对执行云计算任务的第一调配收益,以及各可靠服务器的第二调配收益,同时兼顾全部服务器的可靠性、成本、计算能力等因素,调配最匹配的服务器或服务器组对云计算任务进行处理。
上面图3和图4从模块化功能实体的角度对本申请实施例中的服务器调配装置进行详细描述,下面从硬件处理的角度对本申请实施例中计算机设备进行详细描述。
图5是本申请实施例提供的一种计算机设备的结构示意图,该计算机设备500可因配置或性能不同而产生比较大的差异,可以包括一个或一个以上处理器(central processing units,CPU)510(例如,一个或一个以上处理器)和存储器520,一个或一个以上存储应用程序533或数据532的存储介质530(例如一个或一个以上海量存储设备)。其中,存储器520和存储介质530可以是短暂存储或持久存储。存储在存储介质530的程序可以包括一个或一个以上模块(图示没标出),每个模块可以包括对计算机设备500中的一系列指令操作。更进一步地,处理器510可以设置为与存储介质530通信,在计算机设备500上执行存储介质530中的一系列指令操作。
计算机设备500还可以包括一个或一个以上电源540,一个或一个以上有线或无线网络接口550,一个或一个以上输入输出接口560,和/或,一个或一个以上操作系统531,例如Windows Serve,Mac OS X,Unix,Linux,FreeBSD等等。本领域技术人员可以理解,图5示出的计算机设备结构并不构成对计算机设备的限定,可以包括比图示更多或更少的部件,或者组合某些部件,或者不同的部件布置。
本申请还提供一种计算机设备,所述计算机设备包括存储器和处理器,存储器中存储有计算机可读指令,计算机可读指令被处理器执行时,使得处理器执行上述各实施例中的所述服务器调配方法的步骤。
在实际应用中,该计算机设备可以是执行服务器调配处理的任何一种设备。
本申请还提供一种计算机可读存储介质,该计算机可读存储介质可以为非易失性计算机可读存储介质,该计算机可读存储介质也可以为易失性计算机可读存储介质,所述计算机可读存储介质中存储有指令,当所述指令在计算机上运行时,使得计算机执行所述服务器调配方法的步骤。
所属领域的技术人员可以清楚地了解到,为描述的方便和简洁,上述描述的系统,装置和单元的具体工作过程,可以参考前述方法实施例中的对应过程,在此不再赘述。
所述集成的单元如果以软件功能单元的形式实现并作为独立的产品销售或使用时,可以存储在一个计算机可读取存储介质中。基于这样的理解,本申请的技术方案本质上或者说对现有技术做出贡献的部分或者该技术方案的全部或部分可以以软件产品的形式体现出来,该计算机软件产品存储在一个存储介质中,包括若干指令用以使得一台计算机设备(可以是个人计算机,服务器,或者网络设备等)执行本申请各个实施例所述方法的全部或部分步骤。而前述的存储介质包括:U盘、移动硬盘、只读存储器(read-only memory,ROM)、随机存取存储器(random access memory,RAM)、磁碟或者光盘等各种可以存储程序代码的介质。
本申请所指区块链是分布式数据存储、点对点传输、共识机制、加密算法等计算机技术的新型应用模式。区块链(Blockchain),本质上是一个去中心化的数据库,是一串使用密码学方法相关联产生的数据块,每一个数据块中包含了一批次网络交易的信息,用于验证其信息的有效性(防伪)和生成下一个区块。区块链可以包括区块链底层平台、平台产品服务层以及应用服务层等。
以上所述,以上实施例仅用以说明本申请的技术方案,而非对其限制;尽管参照前述实施例对本申请进行了详细的说明,本领域的普通技术人员应当理解:其依然可以对前述各实施例所记载的技术方案进行修改,或者对其中部分技术特征进行等同替换;而这些修改或者替换,并不使相应技术方案的本质脱离本申请各实施例技术方案的精神和范围。

Claims (20)

  1. 一种服务器调配方法,其中,所述服务器调配方法包括:
    获取云计算任务及对应的任务截止时间,并确定用于处理所述云计算任务的多个服务器;
    根据所述任务截止时间,采用预置任务失效预测方法,将所述多个服务器划分为可靠服务器和不可靠服务器;
    依次对所述各不可靠服务器与所述各可靠服务器进行任务迁移匹配,得到可用于完成所述云计算任务的多个服务器组,其中所述服务器组包含一个不可靠服务器与一个可靠服务器;
    分别计算各所述服务器组的第一调配收益,以及分别计算各所述可靠服务器的第二调配收益;
    基于所述第一调配收益和所述第二调配收益,选取一个调配收益最高的可靠服务器或服务器组执行所述云计算任务。
  2. 根据权利要求1所述的服务器调配方法,其中,所述根据所述任务截止时间,采用预置任务失效预测方法,将所述多个服务器划分为可靠服务器和不可靠服务器包括:
    采用所述预置任务失效预测方法,分别预测所述多个服务器处理所述云计算任务的完成时间;
    依次判断各所述完成时间是否在所述任务截止时间之前;
    若是,则将对应服务器划分为可靠服务器,否则将对应服务器划分为不可靠服务器。
  3. 根据权利要求2所述的服务器调配方法,其中,所述采用所述预置任务失效预测方法,分别预测所述多个服务器处理所述云计算任务的完成时间包括:
    确定所述云计算任务的负载量,以及确定各所述服务器的计算能力;
    基于所述负载量和所述计算能力,采用所述预置任务失效预测方法,分别预测所述多个服务器处理所述云计算任务的完成时间。
  4. 根据权利要求3所述的服务器调配方法,其中,所述依次对所述各不可靠服务器与所述各可靠服务器进行任务迁移匹配,得到可用于完成所述云计算任务的多个服务器组包括:
    依次选取一个不可靠服务器和一个可靠服务器,并分别确定两者的计算能力;
    基于所述两者的计算能力和所述负载量,计算所述云计算任务从所述不可靠服务器迁移到所述可靠服务器的迁移时间节点;
    判断所述迁移时间节点是否在预置时间节点之前;
    若是,则将对应的不可靠服务器和可靠服务器进行配对,得到对应的多个服务器组。
  5. 根据权利要求4所述的服务器调配方法,其中,所述分别计算各所述服务器组的第一调配收益包括:
    分别确定各所述服务器组中不可靠服务器的第一成本因子及可靠服务器的第二成本因子;
    基于所述第一成本因子与不可靠服务器的计算能力,分别计算各所述服务器组在所述迁移时间节点之前的第一初始调配收益;
    基于所述第二成本因子与可靠服务器的计算能力,分别计算各所述服务器组在所述迁 移时间节点之后的第二初始调配收益;
    基于所述第一初始调配收益和所述第二初始调配收益,确定各所述服务器组的第一调配收益。
  6. 根据权利要求5所述的服务器调配方法,其中,所述分别计算各所述可靠服务器的第二调配收益包括:
    基于各所述可靠服务器的完成时间,分别计算各所述可靠服务器的可靠性因子;
    基于所述可靠性因子与所述第二成本因子、可靠服务器的计算能力,分别计算各所述可靠服务器的第二调配收益。
  7. 一种计算机设备,其中,所述计算机设备包括:存储器和至少一个处理器,所述存储器中存储有指令;
    所述至少一个处理器调用所述存储器中的所述指令,以使得所述计算机设备执行如下所述的服务器调配方法的步骤:
    获取云计算任务及对应的任务截止时间,并确定用于处理所述云计算任务的多个服务器;
    根据所述任务截止时间,采用预置任务失效预测方法,将所述多个服务器划分为可靠服务器和不可靠服务器;
    依次对所述各不可靠服务器与所述各可靠服务器进行任务迁移匹配,得到可用于完成所述云计算任务的多个服务器组,其中所述服务器组包含一个不可靠服务器与一个可靠服务器;
    分别计算各所述服务器组的第一调配收益,以及分别计算各所述可靠服务器的第二调配收益;
    基于所述第一调配收益和所述第二调配收益,选取一个调配收益最高的可靠服务器或服务器组执行所述云计算任务。
  8. 根据权利要求7所述的计算机设备,其中,所述计算机设备执行所述根据所述任务截止时间,采用预置任务失效预测方法,将所述多个服务器划分为可靠服务器和不可靠服务器的步骤时,包括:
    采用所述预置任务失效预测方法,分别预测所述多个服务器处理所述云计算任务的完成时间;
    依次判断各所述完成时间是否在所述任务截止时间之前;
    若是,则将对应服务器划分为可靠服务器,否则将对应服务器划分为不可靠服务器。
  9. 根据权利要求8所述的计算机设备,其中,所述计算机设备执行所述采用所述预置任务失效预测方法,分别预测所述多个服务器处理所述云计算任务的完成时间的步骤时,包括:
    确定所述云计算任务的负载量,以及确定各所述服务器的计算能力;
    基于所述负载量和所述计算能力,采用所述预置任务失效预测方法,分别预测所述多个服务器处理所述云计算任务的完成时间。
  10. 根据权利要求9所述的计算机设备,其中,所述计算机设备执行所述依次对所述各不可靠服务器与所述各可靠服务器进行任务迁移匹配,得到可用于完成所述云计算任务的多个服务器组的步骤时,包括:
    依次选取一个不可靠服务器和一个可靠服务器,并分别确定两者的计算能力;
    基于所述两者的计算能力和所述负载量,计算所述云计算任务从所述不可靠服务器迁移到所述可靠服务器的迁移时间节点;
    判断所述迁移时间节点是否在预置时间节点之前;
    若是,则将对应的不可靠服务器和可靠服务器进行配对,得到对应的多个服务器组。
  11. 根据权利要求10所述的计算机设备,其中,所述计算机设备执行所述分别计算各所述服务器组的第一调配收益的步骤时,包括:
    分别确定各所述服务器组中不可靠服务器的第一成本因子及可靠服务器的第二成本因子;
    基于所述第一成本因子与不可靠服务器的计算能力,分别计算各所述服务器组在所述迁移时间节点之前的第一初始调配收益;
    基于所述第二成本因子与可靠服务器的计算能力,分别计算各所述服务器组在所述迁移时间节点之后的第二初始调配收益;
    基于所述第一初始调配收益和所述第二初始调配收益,确定各所述服务器组的第一调配收益。
  12. 根据权利要求10所述的计算机设备,其中,所述计算机设备执行所述分别计算各所述可靠服务器的第二调配收益的步骤时,包括:
    基于各所述可靠服务器的完成时间,分别计算各所述可靠服务器的可靠性因子;
    基于所述可靠性因子与所述第二成本因子、可靠服务器的计算能力,分别计算各所述可靠服务器的第二调配收益。
  13. 一种计算机可读存储介质,所述计算机可读存储介质上存储有计算机程序,其中,所述计算机程序被处理器执行时实现如下所述的服务器调配方法的步骤:
    获取云计算任务及对应的任务截止时间,并确定用于处理所述云计算任务的多个服务器;
    根据所述任务截止时间,采用预置任务失效预测方法,将所述多个服务器划分为可靠服务器和不可靠服务器;
    依次对所述各不可靠服务器与所述各可靠服务器进行任务迁移匹配,得到可用于完成所述云计算任务的多个服务器组,其中所述服务器组包含一个不可靠服务器与一个可靠服务器;
    分别计算各所述服务器组的第一调配收益,以及分别计算各所述可靠服务器的第二调配收益;
    基于所述第一调配收益和所述第二调配收益,选取一个调配收益最高的可靠服务器或服务器组执行所述云计算任务。
  14. 根据权利要求13所述的计算机可读存储介质,其中,所述计算机程序被处理器执行时实现所述根据所述任务截止时间,采用预置任务失效预测方法,将所述多个服务器划分为可靠服务器和不可靠服务器的步骤时,包括:
    采用所述预置任务失效预测方法,分别预测所述多个服务器处理所述云计算任务的完成时间;
    依次判断各所述完成时间是否在所述任务截止时间之前;
    若是,则将对应服务器划分为可靠服务器,否则将对应服务器划分为不可靠服务器。
  15. 根据权利要求14所述的计算机可读存储介质,其中,所述计算机程序被处理器执行时实现所述采用所述预置任务失效预测方法,分别预测所述多个服务器处理所述云计算任务的完成时间的步骤时,包括:
    确定所述云计算任务的负载量,以及确定各所述服务器的计算能力;
    基于所述负载量和所述计算能力,采用所述预置任务失效预测方法,分别预测所述多个服务器处理所述云计算任务的完成时间。
  16. 根据权利要求15所述的计算机可读存储介质,其中,所述计算机程序被处理器执行时实现所述依次对所述各不可靠服务器与所述各可靠服务器进行任务迁移匹配,得到可用于完成所述云计算任务的多个服务器组的步骤时,包括:
    依次选取一个不可靠服务器和一个可靠服务器,并分别确定两者的计算能力;
    基于所述两者的计算能力和所述负载量,计算所述云计算任务从所述不可靠服务器迁移到所述可靠服务器的迁移时间节点;
    判断所述迁移时间节点是否在预置时间节点之前;
    若是,则将对应的不可靠服务器和可靠服务器进行配对,得到对应的多个服务器组。
  17. 根据权利要求16所述的计算机可读存储介质,其中,所述计算机程序被处理器执行时实现所述分别计算各所述服务器组的第一调配收益的步骤时,包括:
    分别确定各所述服务器组中不可靠服务器的第一成本因子及可靠服务器的第二成本因子;
    基于所述第一成本因子与不可靠服务器的计算能力,分别计算各所述服务器组在所述迁移时间节点之前的第一初始调配收益;
    基于所述第二成本因子与可靠服务器的计算能力,分别计算各所述服务器组在所述迁移时间节点之后的第二初始调配收益;
    基于所述第一初始调配收益和所述第二初始调配收益,确定各所述服务器组的第一调配收益。
  18. 根据权利要求17所述的计算机可读存储介质,其中,所述计算机程序被处理器执行时实现所述分别计算各所述可靠服务器的第二调配收益的步骤时,包括:
    基于各所述可靠服务器的完成时间,分别计算各所述可靠服务器的可靠性因子;
    基于所述可靠性因子与所述第二成本因子、可靠服务器的计算能力,分别计算各所述可靠服务器的第二调配收益。
  19. 一种服务器调配装置,其中,所述服务器调配装置包括:
    获取模块,用于获取云计算任务及对应的任务截止时间,并确定用于处理所述云计算任务的多个服务器;
    划分模块,用于根据所述任务截止时间,采用预置任务失效预测方法,将所述多个服务器划分为可靠服务器和不可靠服务器;
    匹配模块,用于依次对所述各不可靠服务器与所述各可靠服务器进行任务迁移匹配,得到可用于完成所述云计算任务的多个服务器组,其中所述服务器组包含一个不可靠服务器与一个可靠服务器;
    计算模块,用于分别计算各所述服务器组的第一调配收益,以及分别计算各所述可靠 服务器的第二调配收益;
    调配模块,用于基于所述第一调配收益和所述第二调配收益,选取一个调配收益最高的可靠服务器或服务器组执行所述云计算任务。
  20. 根据权利要求19所述的服务器调配装置,其中,所述划分模块包括:
    预测单元,用于采用所述预置任务失效预测方法,分别预测所述多个服务器处理所述云计算任务的完成时间;
    判别单元,用于依次判断各所述完成时间是否在所述任务截止时间之前;
    划分单元,用于若所述完成时间在所述任务截止时间之前,则将对应服务器划分为可靠服务器,否则将对应服务器划分为不可靠服务器。
PCT/CN2021/082702 2020-11-11 2021-03-24 服务器调配方法、装置、设备及存储介质 WO2021203975A1 (zh)

Applications Claiming Priority (2)

Application Number Priority Date Filing Date Title
CN202011251978.7 2020-11-11
CN202011251978.7A CN112333279A (zh) 2020-11-11 2020-11-11 服务器调配方法、装置、设备及存储介质

Publications (1)

Publication Number Publication Date
WO2021203975A1 true WO2021203975A1 (zh) 2021-10-14

Family

ID=74318888

Family Applications (1)

Application Number Title Priority Date Filing Date
PCT/CN2021/082702 WO2021203975A1 (zh) 2020-11-11 2021-03-24 服务器调配方法、装置、设备及存储介质

Country Status (2)

Country Link
CN (1) CN112333279A (zh)
WO (1) WO2021203975A1 (zh)

Families Citing this family (1)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN112333279A (zh) * 2020-11-11 2021-02-05 平安科技(深圳)有限公司 服务器调配方法、装置、设备及存储介质

Citations (5)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN201532634U (zh) * 2009-09-21 2010-07-21 王鹏 基于不可信服务器节点的云计算基础架构系统
CN103546313A (zh) * 2013-09-26 2014-01-29 福州宙斯盾信息技术有限公司 基于云计算的it运维管理系统
US20170093642A1 (en) * 2015-09-25 2017-03-30 Singapore Telecommunications Limited Resource planning system for cloud computing
CN111164597A (zh) * 2017-09-29 2020-05-15 国际商业机器公司 使用流聚类的补丁组的动态重组
CN112333279A (zh) * 2020-11-11 2021-02-05 平安科技(深圳)有限公司 服务器调配方法、装置、设备及存储介质

Patent Citations (5)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN201532634U (zh) * 2009-09-21 2010-07-21 王鹏 基于不可信服务器节点的云计算基础架构系统
CN103546313A (zh) * 2013-09-26 2014-01-29 福州宙斯盾信息技术有限公司 基于云计算的it运维管理系统
US20170093642A1 (en) * 2015-09-25 2017-03-30 Singapore Telecommunications Limited Resource planning system for cloud computing
CN111164597A (zh) * 2017-09-29 2020-05-15 国际商业机器公司 使用流聚类的补丁组的动态重组
CN112333279A (zh) * 2020-11-11 2021-02-05 平安科技(深圳)有限公司 服务器调配方法、装置、设备及存储介质

Also Published As

Publication number Publication date
CN112333279A (zh) 2021-02-05

Similar Documents

Publication Publication Date Title
WO2022037337A1 (zh) 机器学习模型的分布式训练方法、装置以及计算机设备
US10924535B2 (en) Resource load balancing control method and cluster scheduler
Peng et al. Optimus: an efficient dynamic resource scheduler for deep learning clusters
US8799916B2 (en) Determining an allocation of resources for a job
US8650298B2 (en) Resource allocation system
US8117611B2 (en) Method, system, and program product for deploying a platform dependent application in a grid environment
US9483393B1 (en) Discovering optimized experience configurations for a software application
US9807152B2 (en) Distributed processing device and distributed processing system as well as distributed processing method
EP2700008A1 (en) Scheduling map and reduce tasks of jobs for execution according to performance goals
US10782949B2 (en) Risk aware application placement modeling and optimization in high turnover DevOps environments
US20130318538A1 (en) Estimating a performance characteristic of a job using a performance model
CN109614227A (zh) 任务资源调配方法、装置、电子设备及计算机可读介质
Rizvandi et al. A study on using uncertain time series matching algorithms for MapReduce applications
EP2977898B1 (en) Task allocation in a computing environment
Manikandan et al. Virtualized load balancer for hybrid cloud using genetic algorithm
WO2021203975A1 (zh) 服务器调配方法、装置、设备及存储介质
Pande et al. A customer-oriented task scheduling for heterogeneous multi-cloud environment
CN114020469A (zh) 基于边缘节点的多任务学习方法、装置、介质与设备
WO2020143851A2 (en) Method and system for optimizing resource redistribution
Ibrahim et al. Improving mapreduce performance with progress and feedback based speculative execution
US11080092B1 (en) Correlated volume placement in a distributed block storage service
US10387578B1 (en) Utilization limiting for nested object queries
Sindhu et al. Workload characterization and synthesis for cloud using generative stochastic processes
US11755433B2 (en) Method and system for health rank based virtual machine restoration using a conformal framework
US11275529B2 (en) Maintenance management on backup storage systems

Legal Events

Date Code Title Description
121 Ep: the epo has been informed by wipo that ep was designated in this application

Ref document number: 21785496

Country of ref document: EP

Kind code of ref document: A1

NENP Non-entry into the national phase

Ref country code: DE

122 Ep: pct application non-entry in european phase

Ref document number: 21785496

Country of ref document: EP

Kind code of ref document: A1