WO2020024443A1 - 资源调度方法、装置、计算机设备及计算机可读存储介质 - Google Patents

资源调度方法、装置、计算机设备及计算机可读存储介质 Download PDF

Info

Publication number
WO2020024443A1
WO2020024443A1 PCT/CN2018/111117 CN2018111117W WO2020024443A1 WO 2020024443 A1 WO2020024443 A1 WO 2020024443A1 CN 2018111117 W CN2018111117 W CN 2018111117W WO 2020024443 A1 WO2020024443 A1 WO 2020024443A1
Authority
WO
WIPO (PCT)
Prior art keywords
sample
service
data
model
detected
Prior art date
Application number
PCT/CN2018/111117
Other languages
English (en)
French (fr)
Inventor
易仁杰
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 WO2020024443A1 publication Critical patent/WO2020024443A1/zh

Links

Images

Classifications

    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06F—ELECTRIC DIGITAL DATA PROCESSING
    • G06F9/00—Arrangements for program control, e.g. control units
    • G06F9/06—Arrangements for program control, e.g. control units using stored programs, i.e. using an internal store of processing equipment to receive or retain programs
    • G06F9/44—Arrangements for executing specific programs
    • G06F9/455—Emulation; Interpretation; Software simulation, e.g. virtualisation or emulation of application or operating system execution engines
    • G06F9/45533—Hypervisors; Virtual machine monitors
    • G06F9/45558—Hypervisor-specific management and integration aspects
    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06F—ELECTRIC DIGITAL DATA PROCESSING
    • G06F9/00—Arrangements for program control, e.g. control units
    • G06F9/06—Arrangements for program control, e.g. control units using stored programs, i.e. using an internal store of processing equipment to receive or retain programs
    • G06F9/46—Multiprogramming arrangements
    • G06F9/48—Program initiating; Program switching, e.g. by interrupt
    • G06F9/4806—Task transfer initiation or dispatching
    • G06F9/4843—Task transfer initiation or dispatching by program, e.g. task dispatcher, supervisor, operating system
    • G06F9/4881—Scheduling strategies for dispatcher, e.g. round robin, multi-level priority queues
    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06F—ELECTRIC DIGITAL DATA PROCESSING
    • G06F9/00—Arrangements for program control, e.g. control units
    • G06F9/06—Arrangements for program control, e.g. control units using stored programs, i.e. using an internal store of processing equipment to receive or retain programs
    • G06F9/44—Arrangements for executing specific programs
    • G06F9/455—Emulation; Interpretation; Software simulation, e.g. virtualisation or emulation of application or operating system execution engines
    • G06F9/45533—Hypervisors; Virtual machine monitors
    • G06F9/45558—Hypervisor-specific management and integration aspects
    • G06F2009/4557—Distribution of virtual machine instances; Migration and load balancing
    • Y—GENERAL TAGGING OF NEW TECHNOLOGICAL DEVELOPMENTS; GENERAL TAGGING OF CROSS-SECTIONAL TECHNOLOGIES SPANNING OVER SEVERAL SECTIONS OF THE IPC; TECHNICAL SUBJECTS COVERED BY FORMER USPC CROSS-REFERENCE ART COLLECTIONS [XRACs] AND DIGESTS
    • Y02—TECHNOLOGIES OR APPLICATIONS FOR MITIGATION OR ADAPTATION AGAINST CLIMATE CHANGE
    • Y02D—CLIMATE CHANGE MITIGATION TECHNOLOGIES IN INFORMATION AND COMMUNICATION TECHNOLOGIES [ICT], I.E. INFORMATION AND COMMUNICATION TECHNOLOGIES AIMING AT THE REDUCTION OF THEIR OWN ENERGY USE
    • Y02D10/00—Energy efficient computing, e.g. low power processors, power management or thermal management

Definitions

  • the present application relates to the field of Internet technologies, and in particular, to a method, an apparatus, a computer device, and a computer-readable storage medium for resource scheduling.
  • virtual machines provide users with various services in the form of services. Users on the public network need to access the virtual machine through the public IP (Internet Protocol). Because the virtual machines provided by the cloud computing service provider need to be isolated by the firewall, in order to allow users to use the servers provided by the virtual machine normally, the cloud computing service provider provides users with a public network IP and configures the internal network IP in the virtual machine. The mapping relationship between the public IP and the internal IP is configured through the firewall. When the user accesses the virtual machine through the public IP, the firewall forwards the connection to the corresponding virtual machine and allocates resources for the virtual machine, so that the user can Use of resources. In related technologies, there are two billing methods in the cloud platform, which are flow meter charging and bandwidth charging.
  • the platform can schedule resources for the user according to the user's selection, provide services for the user based on the resources scheduled for the user, and charge the resources used by the user using the user-selected charging method.
  • the related technology has at least the following problems: the user's business is not transparent in the use of resources during the process of using resources to work, and the user cannot know the flow and bandwidth of the resource in a timely manner,
  • the staff of the cloud platform needs to assist users in determining which billing method to use and how to schedule resources, which increases consulting costs, wastes a lot of manpower, and is less intelligent.
  • this application provides a resource scheduling method, device, computer equipment, and computer-readable storage medium.
  • the main purpose is to solve the need for current cloud platform staff to assist users in determining which billing method to use and how to perform resource scheduling. , Raise the cost of consulting, waste a lot of manpower, the problem of poor intelligence.
  • a resource scheduling method includes: performing statistics on services to be detected in a virtual machine to obtain statistical data, where the statistical data includes at least service types and services of the services to be detected.
  • Historical operation data which includes at least historical traffic data and historical bandwidth data; obtaining a modeling warehouse, and determining a target sample model in the modeling warehouse based on the statistical data, the modeling warehouse including at least A sample model; determining a target resource scheduling template corresponding to the target sample model, and performing resource scheduling for the service to be detected based on the target resource scheduling template.
  • a resource scheduling device configured to collect statistics on services to be detected in a virtual machine, and obtain statistical data, where the statistics include at least the services to be detected.
  • Business type and business operation historical data the business operation historical data includes at least historical traffic data and historical bandwidth data;
  • a first determining module configured to obtain a modeling warehouse, and based on the statistical data, in the modeling warehouse Determine a target sample model, the modeling warehouse includes at least one sample model;
  • a scheduling module configured to determine a target resource scheduling template corresponding to the target sample model, and perform resources for the service to be detected based on the target resource scheduling template Scheduling.
  • a computer device including a memory and a processor, where the memory stores computer-readable instructions, and the processor implements any of the first aspects when the computer-readable instructions are executed. Steps of the method described in item.
  • a computer non-volatile readable storage medium on which computer readable instructions are stored, and when the computer readable instructions are executed by a processor, any one of the foregoing first aspects is implemented. The steps of the method.
  • a resource scheduling method, device, computer equipment and computer-readable storage medium provided by the present application, and the staff of the current cloud platform assist users to determine which charging method to use and how to perform resource scheduling.
  • this application determines the target sample model based on the statistical data of the business to be tested, and schedules the resources to be tested based on the target resource scheduling template corresponding to the target sample model, without the need for staff to evaluate based on the business to be tested, reducing consulting costs. It saves a lot of manpower and has better intelligence.
  • FIG. 1A shows a schematic flowchart of a resource scheduling method provided by an embodiment of the present application
  • FIG. 1B shows a schematic structure diagram of a resource scheduling system provided by an embodiment of the present application
  • FIG. 1C shows a A schematic flowchart of a resource scheduling method
  • FIG. 1D illustrates a schematic flowchart of a resource scheduling method provided by an embodiment of the present application
  • FIG. 2A illustrates a schematic structural diagram of a resource scheduling device provided by an embodiment of the present application
  • FIG. 2B illustrates A schematic structural diagram of a resource scheduling apparatus provided by an embodiment of the present application
  • FIG. 2C illustrates a schematic structural diagram of a resource scheduling apparatus provided by an embodiment of the present application
  • FIG. 1A shows a schematic flowchart of a resource scheduling method provided by an embodiment of the present application
  • FIG. 1B shows a schematic structure diagram of a resource scheduling system provided by an embodiment of the present application
  • FIG. 1C shows a A schematic flowchart of a resource scheduling method
  • FIG. 2D illustrates a structure of a resource scheduling apparatus provided by an embodiment of the present application
  • 2E shows a schematic structural diagram of a resource scheduling apparatus provided by an embodiment of the present application
  • FIG. 2F shows a schematic structural diagram of a resource scheduling apparatus provided by an embodiment of the present application
  • FIG. 2G shows an example provided by the embodiment of the present application
  • FIG. 2H is a schematic structural diagram of a resource scheduling apparatus according to an embodiment of the present application.
  • An embodiment of the present application provides a resource scheduling method. As shown in FIG. 1A, the method includes: 101. Statistic services to be detected in a virtual machine are obtained to obtain statistical data, and the statistical data includes at least the service type and Business operation historical data. The business operation historical data includes at least historical traffic data and historical bandwidth data.
  • step 105 For the specific process, see step 105 in the following embodiment. 102. Obtain a modeling warehouse. Based on statistical data, determine the target sample model in the modeling warehouse. The modeling warehouse includes at least one sample model. For the specific process, refer to step 106 in the following embodiment. 103. Determine a target resource scheduling template corresponding to the target sample model, and perform resource scheduling for the service to be detected based on the target resource scheduling template. For details, refer to step 107 in the following embodiment.
  • the resource scheduling system includes an intelligent system, a management system, and an underlying system.
  • the intelligent system is used to generate at least one sample model, create a modeling warehouse based on the at least one sample model, and instruct the management system to adjust the resources scheduled to the service to be tested according to the change of the statistical data of the service to be tested, and according to Generate a resource bill for the operation of the service to be tested;
  • the management system user performs resource scheduling for the service to be tested according to the determined target resource scheduling template, monitors the remaining resources of the service to be tested, and statistics the resource consumption of the service to be tested;
  • bottom layer The system is used to provide bandwidth resources and public network address resources for the resource scheduling system, thereby ensuring the normal operation of services in the resource scheduling system.
  • the embodiment of the present application provides a resource scheduling method, which can determine a target sample model according to the statistical data of the service to be detected, and perform resource scheduling for the service to be detected according to the target resource scheduling template corresponding to the target sample model, without the need for staff to
  • the evaluation of the business reduces the cost of consulting, saves a lot of manpower, and is more intelligent.
  • the method includes:
  • the applicant recognizes that during the operation of the same type of service, the historical data generated is substantially the same.
  • multiple sample services of different business types can be determined, and then sample models of different business types can be generated, and the optimal resource scheduling template can be set for different sample models, so that the target can be determined later based on user statistics
  • Sample model and resource scheduling according to the resource scheduling template corresponding to the target sample model simplifying the resource scheduling process, without the need for the cloud platform staff to evaluate the user's users, saving a lot of manpower and resources.
  • the sample models generated will also be relatively large.
  • Step 1 Detect the operation of multiple sample services and obtain multiple sample historical data.
  • the sample historical data includes at least sample flow data and sample bandwidth data.
  • the sample model is used to reflect the operation data of a certain type of sample service, after determining multiple sample services, the operations of the multiple sample services are detected to obtain multiple samples.
  • Historical data specifically, the sample historical data may be sample flow data and sample bandwidth data.
  • level 1 level 2 and level 3 can be set, in which the sample historical data of the sample flow data between 0 and 100 is divided into the level 1; the sample historical data of the sample flow data is between 100 and 1000 The sample historical data between them is classified as level 2; the sample data of the sample flow data between 1000 and 5000 in the sample historical data is classified as level 3.
  • Step 2 Determine the sample service types of the multiple sample services, and classify the multiple sample services according to the sample service types.
  • the application industry can be finance, medical, government, games, Internet, industry, and accordingly, the sample business types can be Web (network), APP (Application, Application), API (Application Programming Interface), application program call interface ) And live.
  • finance can include banks, securities, peer-to-peer (peer-to-peer) networks, and stocks. Both are used as the sample business of the financial type, and the business operation historical data of the above specific industries are used as the sample historical data of the financial type.
  • the daily activity level is divided for each sample historical data in step 1, after classifying multiple sample services, you can filter the sample services in each business type based on the daily activity level. , Remove the sample business that has a large difference between the sample historical data and the sample historical data of other sample businesses in the business type, so as to ensure that the subsequent sample models generated for each business type are more accurate.
  • a quantity threshold can be set to count the daily activity level of the sample historical data of the sample service included in the service type. Number of samples and the amount of sample historical data included in each daily activity level. If the number of sample historical data included in a daily activity level is greater than the data threshold, the daily activity in the business type that is different from the daily activity level can be The volume corresponding to the sample service is removed from the service type. For example, if the quantity threshold is 1000, if the sample historical data of daily activity level 1 included in service type A is 2000, the sample historical data of daily activity level 2 is 4, and the sample historical data of daily activity level 3 is 2.
  • Step 3 For any sample service type, calculate a first average of at least one sample traffic data of at least one sample service in the sample service type, calculate a second average of at least one sample bandwidth data of at least one sample service, and divide the first A mean and a second mean are used as model parameters to generate a sample model of a sample business type.
  • a sample model can be established for the sample services in each sample service type.
  • a first average of at least one sample traffic data of at least one sample service in the sample service type may be calculated, and a second average of at least one sample bandwidth data of the at least one sample service may be calculated.
  • the average number using the first average number and the second average number as model parameters, and then generating a sample model for each sample business type. For example, suppose sample service type 1 includes sample service A, sample service B, and sample service C, where the sample traffic data in the sample history data of sample service A is 1000 and the sample bandwidth data is 200; the sample history of sample service B The sample flow data in the data is 1500 and the sample bandwidth data is 200; the sample traffic data in the sample history data of sample service C is 800 and the sample bandwidth data is 500.
  • Step 4 Perform statistics on sample models of multiple sample business types to generate a modeling warehouse.
  • a modeling warehouse can be generated so that when the statistical data of the user's business to be detected is obtained later, the target sample can be determined in the modeling warehouse. model.
  • any sample model among multiple sample models determine the bandwidth usage and traffic usage of the sample model. Based on the bandwidth usage and traffic usage, create a resource scheduling template for the sample model, and schedule the sample model and resources. Template corresponding storage.
  • the history can be In the case of resource scheduling of each sample service, a resource scheduling template is set for a sample model generated based on the sample service, so that the resource scheduling template of the target sample model can be directly applied to perform resource scheduling for the service to be detected.
  • the public network IP Internet Protocol Address
  • public network IP group the total bandwidth used by each sample model
  • the data forms a resource scheduling template for reference.
  • the resource scheduling template may include the bandwidth usage corresponding to the public network IP or public network IP group that needs to be allocated under the current daily workload.
  • each sample model has a corresponding resource scheduling template, and the number of sample models is huge, the number of resource scheduling templates is also huge.
  • the resource scheduling template can be stored in the modeling warehouse along with the sample model.
  • a scheduling template warehouse can be established for the resource scheduling template, and each resource in the scheduling template warehouse can be scheduled.
  • the templates correspond to each sample model in the modeling warehouse.
  • the 105 Perform statistics on the services to be detected in the virtual machine to obtain statistical data.
  • the statistical data includes at least the service type of the service to be detected and historical service operation data.
  • the historical service operation data includes at least historical traffic data and historical bandwidth data.
  • the virtual machine may provide a data statistics entry.
  • the service to be detected is determined according to the service number provided by the user, and statistical data of the service to be detected corresponding to the service number is obtained.
  • a unit time can be set, and the size of the service usage traffic in the unit time is used as the historical traffic data of each unit time; the bandwidth occupation rate in the unit time is calculated, and the bandwidth occupation rate is calculated.
  • the public network IP uses 600MBit in 1 minute.
  • the public network IP uses 600MBit in 1 minute.
  • the historical traffic data of the public network IP within 1 minute is 600MBit / 1min.
  • the corresponding sample model is used as the target sample model.
  • At least one model parameter of at least one sample model can be obtained in the modeling warehouse, and at least one similarity between the at least one model parameter and the statistical data is calculated. Furthermore, it is determined which sample model to use as the target sample model according to at least one similarity. For any sample model in at least one sample model, when calculating the similarity between the sample model and the statistical data, the number of parameters of the parameters that are consistent with the sample parameters of the sample data can be determined, and the number of parameters is calculated in the total of the parameters. The proportion of the number, which is used as the similarity between the sample model and the statistical data.
  • the method can be used to determine the similarity between the sample model and the statistical data, so that at least one similarity of at least one sample model can be obtained.
  • the at least one obtained similarity can be sorted from large to small, and the similarity ranked first, that is, the largest similarity corresponds
  • the sample model is used as the target sample model, and then the resources are scheduled for the service to be detected according to the resource scheduling template corresponding to the target sample model.
  • at least one similarity may be sorted from small to large. Accordingly, the sample model corresponding to the similarity ranked last is used as the target sample model.
  • the manner is not specifically limited.
  • 107 Determine a target resource scheduling template corresponding to the target sample model, obtain a target bandwidth resource amount of the target resource scheduling template, deploy a public network address on a public network device, allocate a bandwidth resource indicated by the target bandwidth resource amount to the public network address, The network address is assigned to the service to be detected.
  • a target resource scheduling template corresponding to the target sample model may be further determined, and then resource scheduling is performed for the service to be detected according to the target resource scheduling template.
  • scheduling can be performed based on public network devices, and the public network devices may specifically include routers, firewalls, load balancing, and gateways.
  • the public network devices may specifically include routers, firewalls, load balancing, and gateways.
  • the public network address can be entered into the public network device.
  • the preset bandwidth resource indicated in the target resource scheduling template is assigned to the public network address. And limit the maximum bandwidth that the public network address can use.
  • the public network bandwidth resource is also the public network BGP (Border Gateway Protocol) bandwidth resource reserve. It is an indispensable resource reserve for a cloud computing service provider. It is carried out through the network with domestic operators and other sub-operators. BGP interconnects to obtain public network bandwidth resources. After setting the public network address, assign the public network address to the service to be tested for use. For example, a public network address is set to 101.1.1.1, a 100 MBps bandwidth resource is allocated to the public network address, and the public network address is allocated to a service to be detected. It should be noted that, since the statistical data of the service to be detected is constantly changing with time, in this way, the optimal resource amount for resource scheduling for the service to be detected is also constantly changing.
  • BGP Border Gateway Protocol
  • an adjustment period can be set, and every adjustment period , Then execute the processes in steps 102 to 104 above to re-determine a new target sample model for the business to be detected, and perform resource scheduling for the resources to be detected according to the new target resource scheduling template corresponding to the new target sample model.
  • the determination of the new target The process of the sample model and the new target resource scheduling template will not be repeated here.
  • an interception period can be set, and the trend sample data is intercepted from the statistical data of the business to be detected based on the interception period, and the prediction result is generated by the big data operation.
  • the trend sample data includes at least trend flow data and trend bandwidth data; the prediction result may be a function that matches the change of the trend sample data, for example, an exponential function, a logarithmic function, a power function, a linear function, a quadratic function, and the like. It should be noted that when generating the prediction results, a rectangular coordinate system can be established for the trend traffic data and the trend bandwidth data, and the trend sample data is reflected in the rectangular coordinate system.
  • the changes over time The trend initially determines the formula of the function that matches the change, and brings the trend sample data into the formula to determine the specific function.
  • y Trend flow data
  • x time
  • bring the data of trend flow data with time into the function formula y kx + b
  • determine the value of k and b then you can determine the specific function, and use this specific function as the prediction result .
  • the resources of the service to be detected can be adjusted according to the prediction result.
  • the bandwidth size and traffic size that the service to be detected needs to be set at each time are predicted, and automatic bandwidth adjustment is implemented.
  • the resource scheduling system may generate a resource bill according to the consumption of the resource by the service to be detected during operation, and return the generated resource bill to user.
  • the running time of the service to be detected may be determined first, and the bandwidth resource scheduled as the service to be detected may be determined. Then, a product of the running time and the bandwidth resource is calculated, and a resource bill is generated based on the product. For example, if the resource cost is a yuan Mbps / hour, the running time is T1, and the bandwidth resource is B1, then the generated resource bill is a * T1 * B1.
  • each interval time period is T1, T2, T3, T4, T5
  • the bandwidth value will be adjusted to B1 each time.
  • B2, B3, B4, then the charging value during this time is a * (B1 * T1 + B2 * T2 + B3 * T3 + B4 * T4 + B5 * T5).
  • step 110 Monitor the remaining resource amount of the service to be detected. If it is detected that the remaining resource amount reaches the resource amount threshold, perform the following step 111; if it is detected that the remaining resource amount does not reach the resource amount threshold, perform the following step 112.
  • monitoring when monitoring the remaining resource amount of the service to be detected, monitoring can be performed from two aspects.
  • the consumption of bandwidth resources by the service to be detected is monitored to determine the remaining resources of the service to be detected;
  • the consumption of the public network address of the service to be detected is monitored to determine the remaining resources of the service to be detected the amount.
  • a first threshold may be set as the resource amount threshold. If it is detected that the bandwidth resources are about to be depleted or the amount of remaining resources has reached the resource amount threshold, a warning is issued.
  • step 111 when monitoring the consumption of resources of the public network address of the service to be detected, a second threshold may be set as the resource amount threshold. If the resources of the public network address are detected to be exhausted, or the remaining resources are used up If the amount has reached the resource amount threshold, a warning is issued, that is, step 111 described below is performed. In addition, if it is detected that the remaining resource amount of the service to be detected does not reach the resource amount threshold, it means that there is currently no need to replenish the resources scheduled for the service to be detected, and the service to be detected can work normally, that is, step 112 below is performed.
  • the remaining resource amount reaches the resource amount threshold, it means that the resources currently scheduled for the service to be detected are about to be exhausted. Therefore, it is necessary to generate a warning prompt based on the remaining resource amount and display the generated warning prompt So that the staff can replenish the resources dispatched to the business to be detected after obtaining the warning prompt.
  • the resource scheduling method can determine a target sample model according to the statistical data of the service to be detected, and perform resource scheduling for the service to be detected according to the target resource scheduling template corresponding to the target sample model. Evaluation reduces consulting costs, saves a lot of manpower, and is more intelligent.
  • an embodiment of the present application provides a resource scheduling apparatus.
  • the apparatus includes a statistics module 201, a first determination module 202, and a scheduling module 203.
  • the statistics module 201 is configured to collect statistics on the services to be detected in the virtual machine, and obtain statistical data.
  • the statistical data includes at least the service type and historical service operation data of the service to be detected.
  • the historical service operation data includes at least historical traffic data and historical bandwidth.
  • the first determining module 202 is configured to obtain a modeling warehouse, and determine a target sample model in the modeling warehouse based on statistical data; the modeling warehouse includes at least one sample model; the scheduling module 203 is used to determine the target sample model The corresponding target resource scheduling template is based on the target resource scheduling template to perform resource scheduling for the service to be detected.
  • the device further includes a detection module 204, a classification module 205, a calculation module 206, a model generation module 207, and a establishment module 208.
  • the detection module 204 is configured to detect the operation of multiple sample services and obtain multiple sample historical data.
  • the sample historical data includes at least sample flow data and sample bandwidth data.
  • the classification module 205 is used to determine A sample service type, which classifies multiple sample services according to the sample service type; the calculation module 206 is configured to calculate, for any sample service type, a first average number of sample traffic data of all sample services in the sample service type, Calculate the second average of the sample bandwidth data of all sample services; the model generation module 207 is used to use the first average and the second average as model parameters to generate a sample model of the sample service type; the establishment module 208 uses It is used for statistics on sample models of multiple sample business types to generate a modeling warehouse.
  • the device further includes a second determination module 209, a creation module 210, and a storage module 211.
  • the second determining module 209 is configured to determine the bandwidth usage and the traffic usage of the sample model for any one of the multiple sample models.
  • the creation module 210 is configured to create a bandwidth based on the bandwidth usage and the traffic usage.
  • the resource scheduling template of the sample model; the storage module 211 is configured to store the sample model and the resource scheduling template correspondingly.
  • the first determination module 202 includes an acquisition submodule 2021, a calculation submodule 2022, and a sorting submodule 2023.
  • the obtaining sub-module 2021 is used to obtain the model parameters of at least one sample model in the modeling warehouse; the calculation sub-module 2022 is used to calculate the similarity between the model parameters and statistical data of each sample model; the ranking sub-module 2023, Used to sort the similarity corresponding to at least one sample model from large to small, and use the sample model corresponding to the similarity ranked first as the target sample model.
  • the scheduling module 203 includes an acquisition submodule 2031, a deployment submodule 2032, and an allocation submodule 2033.
  • the acquisition sub-module 2031 is used to determine a target resource scheduling template corresponding to the target sample model, and to obtain the target bandwidth resource amount of the target resource scheduling template.
  • the deployment sub-module 2032 is used to deploy a public network address on a public network device.
  • the device includes at least a router, a firewall, a load balancer, and a gateway; the allocation submodule 2033 is configured to allocate a bandwidth resource indicated by a target bandwidth resource amount to a public network address, and allocate a public network address to a service to be detected.
  • the first determining module 202 is further configured to obtain an adjustment period, and re-execute the above-mentioned process of determining a target sample model every adjustment period to determine a new target sample model of a service to be detected; the scheduling module 203 Is also used to determine a new target resource scheduling template corresponding to the new target sample model, and based on the new target resource scheduling template, perform resource scheduling for the service to be detected.
  • the device further includes an interception module 212, a result generation module 213, and an adjustment module 214.
  • the interception module 212 is used to determine the interception period, and based on the interception period, the trend sample data is intercepted from the statistical data of the service to be detected; the result generation module 213 is configured to generate the prediction result of the service to be detected based on the trend sample data; An adjustment module 214 is configured to adjust resources of a service to be detected according to a prediction result.
  • the device further includes a monitoring module 215, a warning module 216, and an operation module 217.
  • the monitoring module is used to monitor the remaining resource amount of the service to be detected; the warning module is used to generate a warning prompt based on the remaining resource amount if it is detected that the remaining resource amount reaches the resource threshold value; the operation module is used to If it is detected that the remaining resource amount does not reach the resource amount threshold, the running state of the service to be detected is maintained.
  • the device further includes an acquisition module 218 and a bill generation module 219.
  • the obtaining module 218 is configured to obtain the running time, bandwidth resources, and resource costs of the services to be detected after detecting the end of the services to be detected, and the bill generation module 219 is configured to be based on the operating hours, bandwidth resources, and resource costs.
  • an embodiment of the present application further provides a computer device including a memory and a processor.
  • the memory stores computer-readable instructions
  • the processor executes the processor.
  • the computer-readable instructions implement the steps of the method described in FIGS. 1C and 1D.
  • an embodiment of the present application further provides a computer non-volatile readable storage medium, on which Computer-readable instructions are stored, and when executed by a processor, the steps of the method shown in FIG. 1C and FIG. 1D are implemented.
  • a target sample model can be determined according to the statistical data of the service to be detected, and the resource to be tested is scheduled according to the target resource scheduling template corresponding to the target sample model, without the need for staff to evaluate the service to be detected.
  • the consulting cost is reduced, a lot of manpower is saved, and the intelligence is better.
  • this application can be implemented by hardware, or by software plus necessary general hardware platform.
  • the technical solution of this application can be embodied in the form of a software product, which can be stored in a non-volatile storage medium (which can be a CD-ROM, a U disk, a mobile hard disk, etc.), including several The instructions are used to cause a computer device (which may be a personal computer, a server, or a network device, etc.) to execute the methods described in each implementation scenario of this application.
  • a computer device which may be a personal computer, a server, or a network device, etc.
  • modules in the device in the implementation scenario may be distributed among the devices in the implementation scenario according to the description of the implementation scenario, or may be correspondingly located in one or more devices different from the implementation scenario.
  • the modules of the above implementation scenario can be combined into one module, or further divided into multiple sub-modules.

Landscapes

  • Engineering & Computer Science (AREA)
  • Software Systems (AREA)
  • Theoretical Computer Science (AREA)
  • Physics & Mathematics (AREA)
  • General Engineering & Computer Science (AREA)
  • General Physics & Mathematics (AREA)
  • Data Exchanges In Wide-Area Networks (AREA)
  • Management, Administration, Business Operations System, And Electronic Commerce (AREA)

Abstract

一种资源调度方法、装置、计算机设备及计算机可读存储介质，涉及互联网技术领域，可以根据待检测业务的统计数据确定目标样本模型，并根据目标样本模型对应的目标资源调度模板对待检测业务进行资源的调度，无需工作人员根据待检测业务进行评估，降低了咨询成本，节省了大量人力，智能性较好。所述方法包括：对虚拟机中的待检测业务进行统计，得到统计数据；获取建模仓库，基于统计数据，在建模仓库中确定目标样本模型，建模仓库包括至少一个样本模型；确定目标样本模型对应的目标资源调度模板，基于目标资源调度模板，为待检测业务进行资源调度。

Description

资源调度方法、装置、计算机设备及计算机可读存储介质
本申请要求与2018年8月1日提交中国专利局、申请号为2018108664314、申请名称为“资源调度方法、装置、计算机设备及计算机可读存储介质”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
本申请涉及互联网技术领域，特别是涉及一种资源调度方法、装置、计算机设备及计算机可读存储介质。
背景技术
在云计算时代，虚拟机通过服务的形式向用户提供各种业务，公网的用户需要通过公网IP(Internet Protocol，网络之间互连的协议)才能对虚拟机进行访问。由于云计算服务商提供的虚拟机都需要经过防火墙的隔离，因此，为了使用户可以正常使用虚拟机提供的服务器，云计算服务商为用户提供公网IP，并在虚拟机中配置内网IP，通过防火墙配置公网IP与内网IP的映射关系，当用户通过公网IP访问虚拟机时，防火墙把连接转发到对应的虚拟机上，并为该虚拟机分配资源，从而为使用户可以对资源进行利用。相关技术中，由于云平台中有两种计费方式，分别为流量计费和带宽计费，采用哪种计费方式进行计费需要用户根据资源的调度情况及使用情况进行选择，从而使云平台可以根据用户的选择，为用户调度其所需资源量的资源，基于为用户调度的资源为用户提供服务，并采用用户选择的计费方式对用户使用的资源进行计费。
在实现本申请的过程中，申请人发现相关技术至少存在以下问题：用户的业务在使用资源进行工作的过程中其资源的使用情况并不是透明的，用户无法及时获知资源的流量及带宽情况，云平台的工作人员需要协助用户确定采用哪种计费方式以及如何进行资源调度，提高了咨询成本，浪费大量人力，智能性较差。
发明内容
有鉴于此，本申请提供了一种资源调度方法、装置、计算机设备及计算机可读存储介质，主要目的在于解决目前云平台的工作人员需要协助用户确定采用哪种计费方式以及如何进行资源调度，提高了咨询成本，浪费大量人力，智能性较差的问题。
依据本申请第一方面，提供了一种资源调度方法，该方法包括：对虚拟机中的待检测业务进行统计，得到统计数据，所述统计数据至少包括所述待检测业务的业务类型和业务运营历史数据，所述业务运营历史数据至少包括历史流量数据和历史带宽数据；获取建模仓库，基于所述统计数据，在所述建模仓库中确定目标样本模型，所述建模仓库包括至少一个样本模型；确定所述目标样本模型对应的目标资源调度模板，基于所述目标资源调度模板，为所述待检测业务进行资源调度。
依据本申请第二方面，提供了一种资源调度装置，该装置包括：统计模块，用于对虚拟机中的待检测业务进行统计，得到统计数据，所述统计数据至少包括所述待检测业务的业务类型和业务运营历史数据，所述业务运营历史数据至少包括历史流量数据和历史带宽数据；第一确定模块，用于获取建模仓库，基于所述统计数据，在所述建模仓库中确定目标样本模型，所述建模仓库包括至少一个样本模型；调度模块，用于确定所述目标样本模型对应的目标资源调度模板，基于所述目标资源调度模板，为所述待检测业务进行资源调度。
依据本申请第三方面，提供了一种计算机设备，包括存储器和处理器，所述存储器存储有计算机可读指令，所述处理器执行所述计算机可读指令时实现上述第一方面中任一项所述方法的步骤。
依据本申请第四方面，提供了一种计算机非易失性可读存储介质，其上存储有计算机可读指令，所述计算机可读指令被处理器执行时实现上述第一方面中任一项所述的方法的步骤。
借由上述技术方案，本申请提供的一种资源调度方法、装置、计算机设备及计算机可读存储介质，与目前云平台的工作人员协助用户确定采用哪种计费方式以及如何进 行资源调度的方式相比，本申请根据待检测业务的统计数据确定目标样本模型，并根据目标样本模型对应的目标资源调度模板对待检测业务进行资源的调度，无需工作人员根据待检测业务进行评估，降低了咨询成本，节省了大量人力，智能性较好。上述说明仅是本申请技术方案的概述，为了能够更清楚了解本申请的技术手段，而可依照说明书的内容予以实施，并且为了让本申请的上述和其它目的、特征和优点能够更明显易懂，以下特举本申请的具体实施方式。
附图说明
通过阅读下文优选实施方式的详细描述，各种其他的优点和益处对于本领域普通技术人员将变得清楚明了。附图仅用于示出优选实施方式的目的，而并不认为是对本申请的限制。而且在整个附图中，用相同的参考符号表示相同的部件。在附图中：
图1A示出了本申请实施例提供的一种资源调度方法流程示意图；图1B示出了本申请实施例提供的一种资源调度系统结构示意图；图1C示出了本申请实施例提供的一种资源调度方法流程示意图；图1D示出了本申请实施例提供的一种资源调度方法流程示意图；图2A示出了本申请实施例提供的一种资源调度装置结构示意图；图2B示出了本申请实施例提供的一种资源调度装置结构示意图；图2C示出了本申请实施例提供的一种资源调度装置结构示意图；图2D示出了本申请实施例提供的一种资源调度装置结构示意图；图2E示出了本申请实施例提供的一种资源调度装置结构示意图；图2F示出了本申请实施例提供的一种资源调度装置结构示意图；图2G示出了本申请实施例提供的一种资源调度装置结构示意图；图2H示出了本申请实施例提供的一种资源调度装置结构示意图。
具体实施方式
下面将参照附图更详细地描述本申请的示例性实施例。虽然附图中显示了本申请的示例性实施例，然而应当理解，可以以各种形式实现本申请而不应被这里阐述的实施例 所限制。相反，提供这些实施例是为了能够更透彻地理解本申请，并且能够将本申请的范围完整的传达给本领域的技术人员。本申请实施例提供了一种资源调度方法，如图1A所示，该方法包括：101、对虚拟机中的待检测业务进行统计，得到统计数据，统计数据至少包括待检测业务的业务类型和业务运营历史数据，业务运营历史数据至少包括历史流量数据和历史带宽数据。具体过程参见下述实施例的步骤105。102、获取建模仓库，基于统计数据，在建模仓库中确定目标样本模型，建模仓库包括至少一个样本模型。具体过程参见下述实施例的步骤106。103、确定目标样本模型对应的目标资源调度模板，基于目标资源调度模板，为待检测业务进行资源调度。具体过程参见下述实施例的步骤107。
在对本申请实施例进行详细的解释说明之前，先对本申请实施例涉及的资源调度系统的结构进行简单介绍。参加图1B，资源调度系统中包括智能系统、管理系统和底层系统。其中，智能系统用于生成至少一个样本模型，基于至少一个样本模型创建建模仓库，并根据待检测业务的统计数据的变化情况，指示管理系统对调度给待检测业务的资源进行调整，以及根据待检测业务的运行情况生成资源账单；管理系统用户根据确定的目标资源调度模板为待检测业务进行资源调度，对待检测业务的剩余资源量进行监控，以及对待检测业务耗费的资源量进行统计；底层系统用于为资源调度系统提供带宽资源以及公网地址资源，进而保证资源调度系统中业务的正常运行。
本申请实施例提供了一种资源调度方法，可以根据待检测业务的统计数据确定目标样本模型，并根据目标样本模型对应的目标资源调度模板对待检测业务进行资源的调度，无需工作人员根据待检测业务进行评估，降低了咨询成本，节省了大量人力，智能性较好的目的，如图1C所示，该方法包括：
104、基于多个样本业务，创建多个样本模型，生成建模仓库。
在本申请实施例中，申请人认识到，同一类型的业务，在运行的过程中，产生的历史数据大体上是相同的，为了根据用户的业务类型和业务运营历史数据来进行资源的 调度，从而避免资源的浪费，可以确定多个不同业务类型的样本业务，进而生成不同业务类型的样本模型，为不同的样本模型设置最佳的资源调度模板，使得后续可以根据用户的统计数据来确定目标样本模型，并根据目标样本模型对应的资源调度模板进行资源的调度，简化资源调度的过程，无需云平台的工作人员为用户的用户进行评估，节省大量的人力物力。其中，由于涉及到的业务类型较多，生成的样本模型也会相对较多，因此，可以建立包括至少一个业务模型的建模仓库，从而后续在获取到用户的统计数据时，在建模仓库中确定与用户的待检测业务匹配的目标样本模型。在生成建模仓库时，可以通过执行下述步骤一至步骤五实现。步骤一、对多个样本业务的运行进行检测，获取多个样本历史数据，样本历史数据至少包括样本流量数据和样本带宽数据。在本申请实施例中，由于样本模型是用于体现某一类型的样本业务的运营数据的，因此，在确定多个样本业务后，对多个样本业务的运行进行检测，从而获取多个样本历史数据，具体地，样本历史数据可为样本流量数据和样本带宽数据。在实际应用的过程中，为了对样本业务的运行进行评估，进而更加合理的为生成的各个样本模型设置资源调度模板，在获取到至少一个样本历史数据后，可以为每个样本历史数据划分日活量级别，并在后续对多个样本业务类型进行分类时，综合考虑样本业务的样本历史数据的日活量级别。例如，可设置级别1，级别2和级别3，其中，将样本历史数据中样本流量数据在0至100之间的样本历史数据划分为级别1；将样本历史数据中样本流量数据在100至1000之间的样本历史数据划分为级别2；将样本历史数据中样本流量数据在1000至5000之间的样本数据划分为级别3。步骤二、确定多个样本业务的样本业务类型，根据样本业务类型，对多个样本业务进行分类。在本申请实施例中，由于不同的样本业务是不同应用行业中的业务，不同应用行业中业务的业务类型是不同的，但是同一应用行业中业务的历史数据是相似的，因此，可以针对不同的应用行业建立不同的样本模型，这样，便需要确定多个样本业务的样本业务类型，并根据样本业务类型，对多个样本业务进行分类。其中，应用行业可为金融、医疗、政府、游戏、互联网、工业，相应地，样本业务类型可为 Web(网络)、APP(Application，应用程序)、API(Application Programming Interface，应用程序的调用接口)和直播。另外，由于不同的应用行业还可以细分为多个具体的行业，例如，金融具体可以包括银行、证券、P2P(Peer to peer，对等网络)和股票，因此，可以将上述具体行业的业务均作为金融类型的样本业务，并将上述具体行业的业务运营历史数据作为金融类型的样本历史数据。在实际应用的过程中，如果在步骤一中对各个样本历史数据划分了日活量级别，则在对多个样本业务进行分类后，可以基于日活量级别，对各个业务类型中的样本业务进行筛选，将样本历史数据与该业务类型中其他样本业务的样本历史数据相差较多的样本业务剔除，从而保证后续为每个业务类型生成的样本模型更加精确。其中，对于某一个业务类型来说，在基于日活量级别对该业务类型中的样本业务进行筛选时，可以设置数量阈值，统计该业务类型中包括的样本业务的样本历史数据的日活量级别的个数以及每个日活量级别中包括的样本历史数据的数量，如果某一日活量级别中包括的样本历史数据的数量大于数据阈值，则可将该业务类型中与该日活量级别不同的日活量级别对应样本业务从该业务类型中剔除。例如，设数量阈值为1000，如果业务类型A中包括的日活量级别1的样本历史数据为2000个，日活量级别2的样本历史数据为4个，日活量级别3的样本历史数据为2个，则可将日活量级别2和日活量级别3对应的样本业务从业务类型A中剔除。步骤三、对于任一样本业务类型，计算样本业务类型中至少一个样本业务的至少一个样本流量数据的第一平均数，计算至少一个样本业务的至少一个样本带宽数据的第二平均数，将第一平均数和第二平均数作为模型参数，生成样本业务类型的样本模型。在本申请实施例中，当完成对多个样本业务的分类后，便可以为每个样本业务类型中的样本业务建立样本模型。由于一个样本业务类型中包括多个样本业务，因此，可以计算样本业务类型中至少一个样本业务的至少一个样本流量数据的第一平均数，计算至少一个样本业务的至少一个样本带宽数据的第二平均数，将第一平均数和第二平均数作为模型参数，进而生成每个样本业务类型的样本模型。例如，设样本业务类型1中包括样本业务A、样本业务B和样本业务C， 其中，样本业务A的样本历史数据中的样本流量数据为1000，样本带宽数据为200；样本业务B的样本历史数据中的样本流量数据为1500，样本带宽数据为200；样本业务C的样本历史数据中的样本流量数据为800，样本带宽数据为500，则可以计算得到样本流量数据的第一平均数为1100，样本带宽数据的第二平均数为300，这样，便可以将1100和300作为模型参数，生成样本业务类型1的样本模型。步骤四、对多个样本业务类型的样本模型进行统计，生成建模仓库。在本申请实施例中，由于样本业务类型为多个，不同样本业务类型中样本业务的样本历史数据是不同的，因此，可以生成多个样本业务类型的多个样本模型。为了便于对样本模型进行整合管理，因此，可以对多个样本业务模型进行统计，进而生成建模仓库，以便后续在获取到用户待检测业务的统计数据时，可以在建模仓库中确定目标样本模型。步骤五、对于多个样本模型中的任一样本模型，确定样本模型的带宽使用量和流量使用量，基于带宽使用量和流量使用量，创建样本模型的资源调度模板，将样本模型与资源调度模板对应存储。在本申请实施例中，由于后续需要根据确定的与待检测业务的统计数据匹配的目标样本模型，来为用户的待检测业务进行资源调度，因此，在生成建模仓库后，可以根据历史为各个样本业务进行资源调度的情况，来为根据样本业务生成的样本模型设置资源调度模板，从而在后续可以直接应用目标样本模型的资源调度模板为待检测业务进行资源调度。其中，在生成每个样本模型的资源调度模板时，可以根据每个样本模型实际使用的公网IP(Internet Protocol Address，互联网协议地址)(或公网IP群)以及各个样本模型总共使用的带宽数据形成可供参考的资源调度模板。资源调度模板中可以包含当前日活量下需要分配的公网IP或公网IP群对应的带宽使用量。另外，由于每一个样本模型均存在与其对应的资源调度模板，而样本模型的数量是庞大的，因此，资源调度模板的数量也是庞大的，这样，在进行资源调度模板的存储时，一方面，可以将资源调度模板与样本模型一起对应存储至建模仓库中；另一方面，为了减轻建模仓库的存储负担，可以为资源调度模板建立调度模板仓库，并将调度模板仓库中的各个资源调度模板与建模仓库中的各个样本模型一 一对应。
105、对虚拟机中的待检测业务进行统计，得到统计数据，统计数据至少包括待检测业务的业务类型和业务运营历史数据，业务运营历史数据至少包括历史流量数据和历史带宽数据。
在本申请实施例中，为了确定与待检测业务匹配的目标样本模型，进而根据目标样本模型确定待检测业务的资源调度模板，需要对待检测业务进行统计，得到待检测业务的统计数据。其中，虚拟机中可以提供数据统计入口，当检测到用户触发该数据统计入口时，根据用户提供的业务编号，确定待检测业务，并获取该业务编号对应的待检测业务的统计数据。其中，在获取统计数据中的业务运营历史数据时，可以设置单位时间，统计单位时间内业务使用流量的大小作为各个单位时间的历史流量数据；统计单位时间内带宽占用率，将该带宽占用率作为历史带宽数据。例如，设公网IP为101.1.1.1，单位时间为1分钟，该公网IP在1分钟内使用了600MBit，这样，便可以确定该公网IP在1分钟内的历史流量数据为600MBit/1min，也即10MBit/s＝10Mbps；继续以上述数据为例，如果该公网IP101.1.1.1被分配了100MBps的带宽资源，则得到的历史带宽数据为10/100＝10％
106、获取建模仓库中至少一个样本模型的至少一个模型参数，计算至少一个模型参数与统计数据的至少一个相似度，将至少一个相似度从大到小进行排序，将排在首位的相似度对应的样本模型作为目标样本模型。
在本申请实施例中，当获取到待检测业务的统计数据后，便可以在建模仓库中获取至少一个样本模型的至少一个模型参数，计算至少一个模型参数与统计数据的至少一个相似度，进而根据至少一个相似度确定将哪一个样本模型作为目标样本模型。对于至少一个样本模型中的任一样本模型，在计算该样本模型与统计数据的相似度时，可以确定统计数据与样本模型的样本参数中一致的参数的参数数量，计算参数数量在参数的总数量中所占的比例，将该比例作为该样本模型与统计数据的相似度。例如，设样本模型 A与统计数据中一致的参数的参数数量为3个，而参数的总数量为5，则计算得到的参数数量在参数的总数量中所占的比例为3/5等于60％，这样，便可以确定样本模型A与统计数据之间的相似度为60％。对于至少一个样本模型中的每一个样本模型，均可以采用该方法确定样本模型与统计数据之间的相似度，这样，便可以获取至少一个样本模型的至少一个相似度。由于相似度最大的样本模型是与待检测业务最相似的样本模型，因此，可将得到的至少一个相似度从大到小进行排序，将排在首位的相似度，也即最大的相似度对应的样本模型作为目标样本模型，进而在后续根据目标样本模型对应的资源调度模板为待检测业务进行资源的调度。在实际应用的过程中，还可以将至少一个相似度从小到大进行排序，相应地，将排在末位的相似度对应的样本模型作为目标样本模型，本申请实施例对确定目标样本模型的方式不进行具体限定。
107、确定目标样本模型对应的目标资源调度模板，获取目标资源调度模板的目标带宽资源量，在公网设备上部署公网地址，为公网地址分配目标带宽资源量指示的带宽资源，将公网地址分配给待检测业务。
在本申请实施例中，当确定目标样本模型后，进一步可以确定目标样本模型对应的目标资源调度模板，进而根据目标资源调度模板为待检测业务进行资源的调度。其中，在为待检测业务进行资源调度时，可以基于公网设备进行调度，公网设备具体可以包括路由器、防火墙、负载均衡以及网关。在公网设备上部署公网地址时，可将公网地址输入至公网设备中，当公网地址部署完毕后，将目标资源调度模板中指示的预设带宽资源分配给该公网地址，并对该公网地址能够使用的最大带宽进行限制。其中，公网带宽资源也即公网BGP(Border Gateway Protocol，边界网关协议)带宽资源储备，是一家云计算服务商必备的资源储备，通过与国内运营商以及其他次级运营商的网络进行BGP协议的互联，获取公网带宽资源。当完成公网地址的设置后，将公网地址分配给该待检测业务进行使用。例如，设公网地址为101.1.1.1，为该公网地址分配100MBps的带宽资源，并将该公网地址分配给待检测业务使用。需要说明的是，由于待检测业务的统计 数据是随着时间不断变化的，这样，为待检测业务进行资源调度的最佳资源量也是不断变化的，因此，可以设置调整周期，每隔调整周期，便执行上述步骤102至104中的过程，重新为待检测业务确定新目标样本模型，并根据新目标样本模型对应的新目标资源调度模板为待检测资源进行资源调度，此处对确定新目标样本模型以及新目标资源调度模板的过程不再进行赘述。
108、确定截取周期，基于截取周期，在待检测业务的统计数据中截取趋势样本数据，基于趋势样本数据，生成待检测业务的预测结果，根据预测结果，对待检测业务的资源进行调整。
申请人认识到，待检测业务的统计数据是随着时间变化的，而不同时间的统计数据的最佳资源调度方式均是不同的，为了满足待检测业务的基本需求，避免对待检测业务的运行造成影响，因此，可以根据统计数据随着时间变化的变化趋势，对待检测业务在后续的时间中统计数据的变化进行预设，生成预测结果，进而根据预测结果，对待检测业务的资源进行调整，从而保证待检测业务的正常且高效的运行。在对待检测业务的统计数据进行预测时，首先，可以设置截取周期，基于截取周期在待检测业务的统计数据中截取趋势样本数据，并通过大数据运算生成预测结果。其中，趋势样本数据至少包括趋势流量数据和趋势带宽数据；预测结果可为与趋势样本数据的变化匹配的函数，例如，指数函数、对数函数、幂函数、一次函数、二次函数等。需要说明的是，在生成预测结果时，可以分别为趋势流量数据和趋势带宽数据建立直角坐标系，将趋势样本数据在直角坐标系中体现，根据趋势流量数据和趋势带宽数据随着时间变化的趋势初步确定变化相符的函数公式，并将趋势样本数据带入公式中，进而确定具体的函数。例如，设置截取周期可为30天，对30天之内的统计数据的趋势进行匹配确定与一次函数y＝kx+b的变化趋势一致，对于趋势样本数据中的趋势流量数据来说，y为趋势流量数据，x为时间，将趋势流量数据随时间变化的数据带入函数公式y＝kx+b，确定k和b的取值，即可确定具体的函数，将该具体的函数作为预测结果。随后，在确定预测结果后，便可 以根据预测结果，对待检测业务的资源进行调整。具体地，根据预测结果，预测待检测业务在各个时间需要设置的带宽大小和流量大小，并实现实施自动进行带宽调整。其中，由于预测结果为函数形式的，则可以将时间带入函数中，进而实现对带宽及流量的预测。例如，继续以上述生成的一天中关于带宽大小的函数为y＝3x+1，那么，如果时间x为3点，则可以根据函数确定在3点时，需要的带宽大小为10MBps。
109、当检测到待检测业务运行结束后，获取待检测业务运行的运行时间、带宽资源以及资源费用，基于运行时间、带宽资源和资源费用，生成待检测业务的资源账单，将资源账单返回给用户。
在本申请实施例中，为了使用户可以确定待检测业务所消耗的费用，资源调度系统可以根据待检测业务在运行过程中对资源的消耗情况，生成资源账单，并将生成的资源账单返回给用户。其中，在生成资源账单时，可以先确定待检测业务的运行时间，并确定为待检测业务调度的带宽资源，随后，计算运行时间与带宽资源的乘积，基于该乘积，生成资源账单。例如，设资源费用为a元Mbps/小时，运行时间为T1，带宽资源为B1，则生成的资源账单即为a*T1*B1。在实际应用的过程中，由于待检测资源的统计数据是不断变化的，使得为待检测业务调度的资源也是不断变化的，因此，需要分别统计不同时间段内的费用，进而生成资源账单。例如，设a元Mbps/小时，假设1周内，对带宽进行了5次变化，每次的间隔时间段分别为T1、T2、T3、T4、T5，每次的带宽值会调整到B1、B2、B3、B4，那么这段时间的计费值为，a*(B1*T1+B2*T2+B3*T3+B4*T4+B5*T5)。需要说明的是，由于为待检测业务调度的资源是有限的，如果为待检测业务调度的资源耗光，则待检测业务便无法正常运行，这样，可能会对待检测业务造成影响，因此，参见下述图1D中所示的内容，可以对调度给待检测业务的资源的剩余资源量进行监控，并当监控到调度给待检测业务的资源快要耗尽的时候，及时发出警告，使得工作人员可以及时对资源进行补给，避免对待检测业务造成影响。
110、对待检测业务的剩余资源量进行监控，如果检测到剩余资源量达到资源量阈值，则执行下述步骤111；如果检测到剩余资源量未达到资源量阈值，则执行下述步骤112。
在本申请实施例中，在对待检测业务的剩余资源量进行监控时，可以从两方面进行监控。一方面，对待检测业务对带宽资源的消耗情况进行监控，从而确定待检测业务的剩余资源量；另一方面，对待检测业务的公网地址的消耗情况进行监控，从而确定待检测业务的剩余资源量。其中，在对待检测业务对带宽资源的消耗情况进行监控时，可以设置第一阈值作为资源量阈值，如果检测到带宽资源快要消耗尽，或者剩余资源量已经到达了资源量阈值，则发出警告，也即执行下述步骤111；在对待检测业务的公网地址的资源的消耗情况进行监控时，可以设置第二阈值作为资源量阈值，如果检测到公网地址的资源快要消耗尽，或者剩余资源量已经达到了资源量阈值，则发出警告，也即执行下述步骤111。另外，如果检测到待检测业务的剩余资源量未达到资源量阈值，则表示当前无需对待检测业务调度的资源进行补给，待检测业务可以正常进行工作，也即执行下述步骤112。
111、如果检测到剩余资源量达到资源量阈值，则基于剩余资源量，生成警告提示，展示警告提示。
在本申请实施例中，如果检测到剩余资源量达到资源量阈值，则表示当前调度给待检测业务的资源即将耗尽，因此，需要基于剩余资源量，生成警告提示，并展示生成的警告提示，以便工作人员在获取该警告提示后，可以对调度给待检测业务的资源进行补给。
112、如果检测到剩余资源量未达到资源量阈值，则保持待检测业务的运行状态。
在本申请实施例中，如果检测到剩余资源量未达到资源量阈值，则表示当前调度给待检测业务的资源是充足的，因此，保持待检测业务的正常运行状态即可。本申请实施例提供的资源调度方法，可以根据待检测业务的统计数据确定目标样本模型，并根据 目标样本模型对应的目标资源调度模板对待检测业务进行资源的调度，无需工作人员根据待检测业务进行评估，降低了咨询成本，节省了大量人力，智能性较好。
进一步地，作为图1B和图1C方法的具体实现，本申请实施例提供了一种资源调度装置，如图2A所示，装置包括：统计模块201、第一确定模块202和调度模块203。该统计模块201，用于对虚拟机中的待检测业务进行统计，得到统计数据，统计数据至少包括待检测业务的业务类型和业务运营历史数据，业务运营历史数据至少包括历史流量数据和历史带宽数据；该第一确定模块202，用于获取建模仓库，基于统计数据，在建模仓库中确定目标样本模型，建模仓库包括至少一个样本模型；该调度模块203，用于确定目标样本模型对应的目标资源调度模板，基于目标资源调度模板，为待检测业务进行资源调度。在具体的应用场景中，如图2B所示，该装置还包括检测模块204，分类模块205，计算模块206，模型生成模块207和建立模块208。该检测模块204，用于对多个样本业务的运行进行检测，获取多个样本历史数据，样本历史数据至少包括样本流量数据和样本带宽数据；该分类模块205，用于确定多个样本业务的样本业务类型，根据样本业务类型，对多个样本业务进行分类；该计算模块206，用于对于任一样本业务类型，计算样本业务类型中的所有样本业务的样本流量数据的第一平均数，计算所有样本业务的样本带宽数据的第二平均数；该模型生成模块207，用于将第一平均数和第二平均数作为模型参数，生成样本业务类型的样本模型；该建立模块208，用于对多个样本业务类型的样本模型进行统计，生成建模仓库。在具体的应用场景中，如图2C所示，该装置还包括第二确定模块209，创建模块210和存储模块211。该第二确定模块209，用于对于多个样本模型中的任一样本模型，确定样本模型的带宽使用量和流量使用量；该创建模块210，用于基于带宽使用量和流量使用量，创建样本模型的资源调度模板；该存储模块211，用于将样本模型与资源调度模板对应存储。在具体的应用场景中，如图2D所示，该第一确定模块202，包括获取子模块2021，计算子模块2022和排序子模块2023。该获取子模块2021，用于获取建模仓库中至少一个样本模型的模型参 数；该计算子模块2022，用于计算每个样本模型的模型参数与统计数据的相似度；该排序子模块2023，用于将至少一个样本模型对应的相似度从大到小进行排序，将排在首位的相似度对应的样本模型作为目标样本模型。在具体的应用场景中，如图2E所示，该调度模块203，包括获取子模块2031，部署子模块2032和分配子模块2033。该获取子模块2031，用于确定目标样本模型对应的目标资源调度模板，获取目标资源调度模板的目标带宽资源量；该部署子模块2032，用于在公网设备上部署公网地址，公网设备至少包括路由器、防火墙、负载均衡以及网关；该分配子模块2033，用于为公网地址分配目标带宽资源量指示的带宽资源，将公网地址分配给待检测业务。在具体的应用场景中，该第一确定模块202，还用于获取调整周期，每隔调整周期，重新执行上述确定目标样本模型的过程，确定待检测业务的新目标样本模型；该调度模块203，还用于确定新目标样本模型对应的新目标资源调度模板，基于新目标资源调度模板，为待检测业务进行资源调度。在具体的应用场景中，如图2F所示，该装置还包括截取模块212，结果生成模块213和调整模块214。该截取模块212，用于确定截取周期，基于截取周期，在待检测业务的统计数据中截取趋势样本数据；该结果生成模块213，用于基于趋势样本数据，生成待检测业务的预测结果；该调整模块214，用于根据预测结果，对待检测业务的资源进行调整。在具体的应用场景中，如图2G所示，该装置还包括监控模块215，警告模块216和运行模块217。监控模块，用于对待检测业务的剩余资源量进行监控；警告模块，用于如果检测到剩余资源量达到资源量阈值，则基于剩余资源量，生成警告提示，展示警告提示；运行模块，用于如果检测到剩余资源量未达到资源量阈值，则保持待检测业务的运行状态。在具体的应用场景中，如图2H所示，该装置还包括获取模块218和账单生成模块219。该获取模块218，用于当检测到待检测业务运行结束后，获取待检测业务运行的运行时间、带宽资源以及资源费用；该账单生成模块219，用于基于运行时间、带宽资源和资源费用，生成待检测业务的资源账单，将资源账单返回给用户。需要说明的是，本申请实施例提供的一种资源调度装置所涉及各功能单元的其他 相应描述，可以参考图1C和图1D中的对应描述，在此不再赘述。基于上述如图1C和图1D所示方法，相应的，本申请实施例还提供了一种计算机设备，包括存储器和处理器，所述存储器存储有计算机可读指令，所述处理器执行所述计算机可读指令时实现图1C和图1D所述方法的步骤。基于上述如图1C和图1D所示方法和2A至图2H所示虚拟装置的实施例，为了实现上述目的，本申请实施例还提供了一种计算机非易失性可读存储介质，其上存储有计算机可读指令，所述计算机可读指令被处理器执行时实现如图1C和图1D所示方法的步骤。通过应用本申请的技术方案，可以根据待检测业务的统计数据确定目标样本模型，并根据目标样本模型对应的目标资源调度模板对待检测业务进行资源的调度，无需工作人员根据待检测业务进行评估，降低了咨询成本，节省了大量人力，智能性较好。通过以上的实施方式的描述，本领域的技术人员可以清楚地了解到本申请可以通过硬件实现，也可以借助软件加必要的通用硬件平台的方式来实现。基于这样的理解，本申请的技术方案可以以软件产品的形式体现出来，该软件产品可以存储在一个非易失性存储介质(可以是CD-ROM，U盘，移动硬盘等)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场景所述的方法。本领域技术人员可以理解附图只是一个优选实施场景的示意图，附图中的模块或流程并不一定是实施本申请所必须的。本领域技术人员可以理解实施场景中的装置中的模块可以按照实施场景描述进行分布于实施场景的装置中，也可以进行相应变化位于不同于本实施场景的一个或多个装置中。上述实施场景的模块可以合并为一个模块，也可以进一步拆分成多个子模块。

Claims (20)

  1. 一种资源调度方法，其特征在于，包括：
    对虚拟机中的待检测业务进行统计，得到统计数据，所述统计数据至少包括所述待检测业务的业务类型和业务运营历史数据，所述业务运营历史数据至少包括历史流量数据和历史带宽数据；
    获取建模仓库，基于所述统计数据，在所述建模仓库中确定目标样本模型，所述建模仓库包括至少一个样本模型；
    确定所述目标样本模型对应的目标资源调度模板，基于所述目标资源调度模板，为所述待检测业务进行资源调度。
  2. 根据权利要求1所述的方法，其特征在于，所述对虚拟机中的待检测业务进行统计，得到统计数据之前，包括：
    对多个样本业务的运行进行检测，获取多个样本历史数据，所述样本历史数据至少包括样本流量数据和样本带宽数据；
    确定所述多个样本业务的样本业务类型，根据所述样本业务类型，对所述多个样本业务进行分类；
    对于任一样本业务类型，计算所述样本业务类型中的所有样本业务的样本流量数据的第一平均数，计算所述所有样本业务的样本带宽数据的第二平均数；
    将所述第一平均数和所述第二平均数作为模型参数，生成所述样本业务类型的样本模型；
    对所述多个样本业务类型的样本模型进行统计，生成所述建模仓库。
  3. 根据权利要求2所述的方法，其特征在于，所述对所述多个样本业务类型的样本模型进行统计，生成所述建模仓库之后，还包括：
    对于所述多个样本模型中的任一样本模型，确定所述样本模型的带宽使用量和流量使用量；
    基于所述带宽使用量和所述流量使用量，创建所述样本模型的资源调度模板；
    将所述样本模型与所述资源调度模板对应存储。
  4. 根据权利要求1所述的方法，其特征在于，所述获取建模仓库，基于所述统计数据，在所述建模仓库中确定目标样本模型，包括：
    获取所述建模仓库中所述至少一个样本模型的模型参数；
    计算每个样本模型的模型参数与所述统计数据的相似度；
    将所述至少一个样本模型对应的相似度从大到小进行排序，将排在首位的相似度对应的样本模型作为所述目标样本模型。
  5. 根据权利要求1所述的方法，其特征在于，所述确定所述目标样本模型对应的目标资源调度模板，基于所述目标资源调度模板，为所述待检测业务进行资源调度，包括：
    确定所述目标样本模型对应的目标资源调度模板，获取所述目标资源调度模板的目标带宽资源量；
    在公网设备上部署公网地址，所述公网设备至少包括路由器、防火墙、负载均衡以及网关；
    为所述公网地址分配所述目标带宽资源量指示的带宽资源，将所述公网地址分配给所述待检测业务。
  6. 根据权利要求1所述的方法，其特征在于，所述方法还包括：
    确定截取周期，基于所述截取周期，在所述待检测业务的统计数据中截取趋势样本数据；
    基于所述趋势样本数据，生成所述待检测业务的预测结果；
    根据所述预测结果，对所述待检测业务的资源进行调整。
  7. 根据权利要求1所述的方法，其特征在于，所述方法还包括：
    当检测到所述待检测业务运行结束后，获取所述待检测业务运行的运行时间、带 宽资源以及资源费用；
    基于所述运行时间、所述带宽资源和所述资源费用，生成所述待检测业务的资源账单，将所述资源账单返回给用户。
  8. 一种资源调度装置，其特征在于，包括：
    统计模块，用于对虚拟机中的待检测业务进行统计，得到统计数据，所述统计数据至少包括所述待检测业务的业务类型和业务运营历史数据，所述业务运营历史数据至少包括历史流量数据和历史带宽数据；
    第一确定模块，用于获取建模仓库，基于所述统计数据，在所述建模仓库中确定目标样本模型，所述建模仓库包括至少一个样本模型；
    调度模块，用于确定所述目标样本模型对应的目标资源调度模板，基于所述目标资源调度模板，为所述待检测业务进行资源调度。
  9. 根据权利要求8所述的装置，其特征在于，所述装置还包括：
    检测模块，用于对多个样本业务的运行进行检测，获取多个样本历史数据，所述样本历史数据至少包括样本流量数据和样本带宽数据；
    分类模块，用于确定所述多个样本业务的样本业务类型，根据所述样本业务类型，对所述多个样本业务进行分类；
    计算模块，用于对于任一样本业务类型，计算所述样本业务类型中的所有样本业务的样本流量数据的第一平均数，计算所述所有样本业务的样本带宽数据的第二平均数；
    模型生成模块，用于将所述第一平均数和所述第二平均数作为模型参数，生成所述样本业务类型的样本模型；
    建立模块，用于对所述多个样本业务类型的样本模型进行统计，生成所述建模仓库。
  10. 根据权利要求9所述的装置，其特征在于，所述装置还包括：
    第二确定模块，用于对于所述多个样本模型中的任一样本模型，确定所述样本模型的带宽使用量和流量使用量；
    创建模块，用于基于所述带宽使用量和所述流量使用量，创建所述样本模型的资源调度模板；
    存储模块，用于将所述样本模型与所述资源调度模板对应存储。
  11. 根据权利要求8所述的装置，其特征在于，所述第一确定模块，包括：
    获取子模块，用于获取所述建模仓库中所述至少一个样本模型的模型参数；
    计算子模块，用于计算每个样本模型的模型参数与所述统计数据的相似度；
    排序子模块，用于将所述至少一个样本模型对应的相似度从大到小进行排序，将排在首位的相似度对应的样本模型作为所述目标样本模型。
  12. 根据权利要求8所述的装置，其特征在于，所述调度模块，包括：
    获取子模块，用于确定所述目标样本模型对应的目标资源调度模板，获取所述目标资源调度模板的目标带宽资源量；
    部署子模块，用于在公网设备上部署公网地址，所述公网设备至少包括路由器、防火墙、负载均衡以及网关；
    分配子模块，用于为所述公网地址分配所述目标带宽资源量指示的带宽资源，将所述公网地址分配给所述待检测业务。
  13. 根据权利要求8所述的装置，其特征在于，所述装置还包括：
    截取模块，用于确定截取周期，基于所述截取周期，在所述待检测业务的统计数据中截取趋势样本数据；
    结果生成模块，用于基于所述趋势样本数据，生成所述待检测业务的预测结果；
    调整模块，用于根据所述预测结果，对所述待检测业务的资源进行调整。
  14. 根据权利要求8所述的装置，其特征在于，所述装置还包括：
    获取模块，用于当检测到所述待检测业务运行结束后，获取所述待检测业务运行 的运行时间、带宽资源以及资源费用；
    账单生成模块，用于基于所述运行时间、所述带宽资源和所述资源费用，生成所述待检测业务的资源账单，将所述资源账单返回给用户。
  15. 一种计算机设备，包括存储器和处理器，所述存储器存储有计算机可读指令，其特征在于，所述处理器执行所述计算机可读指令时实现资源调度方法，包括：
    对虚拟机中的待检测业务进行统计，得到统计数据，所述统计数据至少包括所述待检测业务的业务类型和业务运营历史数据，所述业务运营历史数据至少包括历史流量数据和历史带宽数据；获取建模仓库，基于所述统计数据，在所述建模仓库中确定目标样本模型，所述建模仓库包括至少一个样本模型；确定所述目标样本模型对应的目标资源调度模板，基于所述目标资源调度模板，为所述待检测业务进行资源调度。
  16. 根据权利要求15所述的计算机设备，其特征在于，所述对虚拟机中的待检测业务进行统计，得到统计数据之前，包括：
    对多个样本业务的运行进行检测，获取多个样本历史数据，所述样本历史数据至少包括样本流量数据和样本带宽数据；确定所述多个样本业务的样本业务类型，根据所述样本业务类型，对所述多个样本业务进行分类；对于任一样本业务类型，计算所述样本业务类型中的所有样本业务的样本流量数据的第一平均数，计算所述所有样本业务的样本带宽数据的第二平均数；将所述第一平均数和所述第二平均数作为模型参数，生成所述样本业务类型的样本模型；对所述多个样本业务类型的样本模型进行统计，生成所述建模仓库。
  17. 根据权利要求16所述的计算机设备，其特征在于，所述对所述多个样本业务类型的样本模型进行统计，生成所述建模仓库之后，还包括：
    对于所述多个样本模型中的任一样本模型，确定所述样本模型的带宽使用量和流量使用量；基于所述带宽使用量和所述流量使用量，创建所述样本模型的资源调度模板；将所述样本模型与所述资源调度模板对应存储。
  18. 一种计算机非易失性可读存储介质，其上存储有计算机可读指令，其特征在于，所述计算机可读指令被处理器执行时实现资源调度方法，包括：
    对虚拟机中的待检测业务进行统计，得到统计数据，所述统计数据至少包括所述待检测业务的业务类型和业务运营历史数据，所述业务运营历史数据至少包括历史流量数据和历史带宽数据；获取建模仓库，基于所述统计数据，在所述建模仓库中确定目标样本模型，所述建模仓库包括至少一个样本模型；确定所述目标样本模型对应的目标资源调度模板，基于所述目标资源调度模板，为所述待检测业务进行资源调度。
  19. 根据权利要求18所述的计算机非易失性可读存储介质，其特征在于，所述对虚拟机中的待检测业务进行统计，得到统计数据之前，包括：
    对多个样本业务的运行进行检测，获取多个样本历史数据，所述样本历史数据至少包括样本流量数据和样本带宽数据；确定所述多个样本业务的样本业务类型，根据所述样本业务类型，对所述多个样本业务进行分类；对于任一样本业务类型，计算所述样本业务类型中的所有样本业务的样本流量数据的第一平均数，计算所述所有样本业务的样本带宽数据的第二平均数；将所述第一平均数和所述第二平均数作为模型参数，生成所述样本业务类型的样本模型；对所述多个样本业务类型的样本模型进行统计，生成所述建模仓库。
  20. 根据权利要求18所述的计算机非易失性可读存储介质，其特征在于，所述对所述多个样本业务类型的样本模型进行统计，生成所述建模仓库之后，还包括：
    对于所述多个样本模型中的任一样本模型，确定所述样本模型的带宽使用量和流量使用量；基于所述带宽使用量和所述流量使用量，创建所述样本模型的资源调度模板；将所述样本模型与所述资源调度模板对应存储。
PCT/CN2018/111117 2018-08-01 2018-10-21 资源调度方法、装置、计算机设备及计算机可读存储介质 WO2020024443A1 (zh)

Applications Claiming Priority (2)

Application Number Priority Date Filing Date Title
CN201810866431.4A CN109189548B (zh) 2018-08-01 2018-08-01 资源调度方法、装置、计算机设备及计算机可读存储介质
CN201810866431.4 2018-08-01

Publications (1)

Publication Number Publication Date
WO2020024443A1 true WO2020024443A1 (zh) 2020-02-06

Family

ID=64920398

Family Applications (1)

Application Number Title Priority Date Filing Date
PCT/CN2018/111117 WO2020024443A1 (zh) 2018-08-01 2018-10-21 资源调度方法、装置、计算机设备及计算机可读存储介质

Country Status (2)

Country Link
CN (1) CN109189548B (zh)
WO (1) WO2020024443A1 (zh)

Cited By (2)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN113391985A (zh) * 2021-06-09 2021-09-14 北京猿力未来科技有限公司 资源分配方法及装置
CN117910929A (zh) * 2024-03-14 2024-04-19 浙江菜鸟供应链管理有限公司 仓储系统全链路处理方法、以及仓储系统全链路仿真平台

Families Citing this family (1)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN110059746A (zh) * 2019-04-18 2019-07-26 达闼科技（北京）有限公司 一种创建目标检测模型的方法、电子设备及存储介质

Citations (3)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN102546379A (zh) * 2010-12-27 2012-07-04 中国移动通信集团公司 一种虚拟化资源调度的方法及虚拟化资源调度系统
CN104065745A (zh) * 2014-07-07 2014-09-24 电子科技大学 云计算动态资源调度系统和方法
CN108259376A (zh) * 2018-04-24 2018-07-06 北京奇艺世纪科技有限公司 服务器集群业务流量的控制方法及相关设备

Family Cites Families (3)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
US8379518B2 (en) * 2007-01-23 2013-02-19 Agere Systems Llc Multi-stage scheduler with processor resource and bandwidth resource allocation
CN102780759B (zh) * 2012-06-13 2016-05-18 合肥工业大学 基于调度目标空间的云计算资源调度方法
CN104881325B (zh) * 2015-05-05 2018-09-21 中国联合网络通信集团有限公司 一种资源调度方法和资源调度系统

Patent Citations (3)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN102546379A (zh) * 2010-12-27 2012-07-04 中国移动通信集团公司 一种虚拟化资源调度的方法及虚拟化资源调度系统
CN104065745A (zh) * 2014-07-07 2014-09-24 电子科技大学 云计算动态资源调度系统和方法
CN108259376A (zh) * 2018-04-24 2018-07-06 北京奇艺世纪科技有限公司 服务器集群业务流量的控制方法及相关设备

Cited By (2)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN113391985A (zh) * 2021-06-09 2021-09-14 北京猿力未来科技有限公司 资源分配方法及装置
CN117910929A (zh) * 2024-03-14 2024-04-19 浙江菜鸟供应链管理有限公司 仓储系统全链路处理方法、以及仓储系统全链路仿真平台

Also Published As

Publication number Publication date
CN109189548B (zh) 2023-05-26
CN109189548A (zh) 2019-01-11

Similar Documents

Publication Publication Date Title
Ghahramani et al. Toward cloud computing QoS architecture: Analysis of cloud systems and cloud services
Dhaya et al. Energy-efficient resource allocation and migration in private cloud data centre
CN103699445B (zh) 一种任务调度方法、装置及系统
US9552231B2 (en) Client classification-based dynamic allocation of computing infrastructure resources
US20180198855A1 (en) Method and apparatus for scheduling calculation tasks among clusters
WO2020024443A1 (zh) 资源调度方法、装置、计算机设备及计算机可读存储介质
US11836535B1 (en) System and method of providing cloud bursting capabilities in a compute environment
CN110333937A (zh) 任务分发方法、装置、计算机设备和存储介质
Herbst et al. Ready for rain? A view from SPEC research on the future of cloud metrics
Cerroni et al. Live migration of virtual network functions in cloud-based edge networks
US11652720B2 (en) Allocating cloud resources in accordance with predicted deployment growth
WO2020015578A1 (zh) 一种调度缓存节点的方法、装置、系统、介质及设备
US20170017918A1 (en) Method and system for enabling dynamic capacity planning
Vos et al. Architectural tactics to optimize software for energy efficiency in the public cloud
CN115917510A (zh) 精简云计算环境中的虚拟机部署
CN105022823B (zh) 一种基于数据挖掘的云服务性能预警事件生成方法
CN114116157A (zh) 一种边缘环境下多边缘集群云结构及负载均衡调度方法
Lucanin et al. A cloud controller for performance-based pricing
Farooq et al. Adaptive and resilient revenue maximizing dynamic resource allocation and pricing for cloud-enabled IoT systems
CN114844791B (zh) 基于大数据的云服务自动管理分配方法、系统及存储介质
Mera-Gómez et al. Elasticity debt: a debt-aware approach to reason about elasticity decisions in the cloud
Altomare et al. Energy-aware migration of virtual machines driven by predictive data mining models
Sujan et al. A batchmode dynamic scheduling scheme for cloud computing
Mokhtari et al. Multi-objective task scheduling using smart MPI-based cloud resources
Wanis et al. Substrate network house cleaning via live virtual network migration

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: 18928537

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: 18928537

Country of ref document: EP

Kind code of ref document: A1