WO2022048560A1 - 提供柔性实例的云数据中心以及柔性实例的调度方法 - Google Patents

提供柔性实例的云数据中心以及柔性实例的调度方法 Download PDF

Info

Publication number
WO2022048560A1
WO2022048560A1 PCT/CN2021/115896 CN2021115896W WO2022048560A1 WO 2022048560 A1 WO2022048560 A1 WO 2022048560A1 CN 2021115896 W CN2021115896 W CN 2021115896W WO 2022048560 A1 WO2022048560 A1 WO 2022048560A1
Authority
WO
WIPO (PCT)
Prior art keywords
flexible
instance
size
management system
resource management
Prior art date
Application number
PCT/CN2021/115896
Other languages
English (en)
French (fr)
Inventor
顾炯炯
闵小勇
黄朝意
蔡智源
Original Assignee
华为云计算技术有限公司
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 华为云计算技术有限公司 filed Critical 华为云计算技术有限公司
Priority to EP21863624.9A priority Critical patent/EP4195751A4/en
Publication of WO2022048560A1 publication Critical patent/WO2022048560A1/zh
Priority to US18/175,937 priority patent/US20230205582A1/en

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
    • G06F9/46—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
    • H—ELECTRICITY
    • H04—ELECTRIC COMMUNICATION TECHNIQUE
    • H04L—TRANSMISSION OF DIGITAL INFORMATION, e.g. TELEGRAPHIC COMMUNICATION
    • H04L41/00—Arrangements for maintenance, administration or management of data switching networks, e.g. of packet switching networks
    • H04L41/50—Network service management, e.g. ensuring proper service fulfilment according to agreements
    • H04L41/5003—Managing SLA; Interaction between SLA and QoS
    • H04L41/5019—Ensuring fulfilment of SLA
    • H04L41/5025—Ensuring fulfilment of SLA by proactively reacting to service quality change, e.g. by reconfiguration after service quality degradation or upgrade
    • 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
    • G06F11/00—Error detection; Error correction; Monitoring
    • G06F11/30—Monitoring
    • G06F11/3051—Monitoring arrangements for monitoring the configuration of the computing system or of the computing system component, e.g. monitoring the presence of processing resources, peripherals, I/O links, software programs
    • 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
    • G06F9/50—Allocation of resources, e.g. of the central processing unit [CPU]
    • G06F9/5005—Allocation of resources, e.g. of the central processing unit [CPU] to service a request
    • G06F9/5011—Allocation of resources, e.g. of the central processing unit [CPU] to service a request the resources being hardware resources other than CPUs, Servers and Terminals
    • G06F9/5016—Allocation of resources, e.g. of the central processing unit [CPU] to service a request the resources being hardware resources other than CPUs, Servers and Terminals the resource being the memory
    • 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
    • G06F9/50—Allocation of resources, e.g. of the central processing unit [CPU]
    • G06F9/5005—Allocation of resources, e.g. of the central processing unit [CPU] to service a request
    • G06F9/5011—Allocation of resources, e.g. of the central processing unit [CPU] to service a request the resources being hardware resources other than CPUs, Servers and Terminals
    • G06F9/5022—Mechanisms to release resources
    • 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
    • G06F9/50—Allocation of resources, e.g. of the central processing unit [CPU]
    • G06F9/5005—Allocation of resources, e.g. of the central processing unit [CPU] to service a request
    • G06F9/5027—Allocation of resources, e.g. of the central processing unit [CPU] to service a request the resource being a machine, e.g. CPUs, Servers, Terminals
    • 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
    • G06F9/50—Allocation of resources, e.g. of the central processing unit [CPU]
    • G06F9/5061—Partitioning or combining of resources
    • G06F9/5072—Grid computing
    • 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
    • G06F9/50—Allocation of resources, e.g. of the central processing unit [CPU]
    • G06F9/5061—Partitioning or combining of resources
    • G06F9/5077—Logical partitioning of resources; Management or configuration of virtualized resources
    • 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
    • G06F9/50—Allocation of resources, e.g. of the central processing unit [CPU]
    • G06F9/5083—Techniques for rebalancing the load in a distributed system
    • G06F9/5088—Techniques for rebalancing the load in a distributed system involving task migration
    • 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
    • G06F9/54—Interprogram communication
    • H—ELECTRICITY
    • H04—ELECTRIC COMMUNICATION TECHNIQUE
    • H04L—TRANSMISSION OF DIGITAL INFORMATION, e.g. TELEGRAPHIC COMMUNICATION
    • H04L41/00—Arrangements for maintenance, administration or management of data switching networks, e.g. of packet switching networks
    • H04L41/08—Configuration management of networks or network elements
    • H04L41/0896—Bandwidth or capacity management, i.e. automatically increasing or decreasing capacities
    • H04L41/0897—Bandwidth or capacity management, i.e. automatically increasing or decreasing capacities by horizontal or vertical scaling of resources, or by migrating entities, e.g. virtual resources or entities
    • 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
    • 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
    • G06F2009/45583—Memory management, e.g. access or allocation
    • 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
    • G06F2009/45595—Network integration; Enabling network access in virtual machine instances
    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06F—ELECTRIC DIGITAL DATA PROCESSING
    • G06F2209/00—Indexing scheme relating to G06F9/00
    • G06F2209/50—Indexing scheme relating to G06F9/50
    • G06F2209/5011—Pool
    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06F—ELECTRIC DIGITAL DATA PROCESSING
    • G06F2209/00—Indexing scheme relating to G06F9/00
    • G06F2209/50—Indexing scheme relating to G06F9/50
    • G06F2209/5022—Workload threshold
    • H—ELECTRICITY
    • H04—ELECTRIC COMMUNICATION TECHNIQUE
    • H04L—TRANSMISSION OF DIGITAL INFORMATION, e.g. TELEGRAPHIC COMMUNICATION
    • H04L41/00—Arrangements for maintenance, administration or management of data switching networks, e.g. of packet switching networks
    • H04L41/50—Network service management, e.g. ensuring proper service fulfilment according to agreements
    • H04L41/5029—Service quality level-based billing, e.g. dependent on measured service level customer is charged more or less
    • H—ELECTRICITY
    • H04—ELECTRIC COMMUNICATION TECHNIQUE
    • H04L—TRANSMISSION OF DIGITAL INFORMATION, e.g. TELEGRAPHIC COMMUNICATION
    • H04L41/00—Arrangements for maintenance, administration or management of data switching networks, e.g. of packet switching networks
    • H04L41/50—Network service management, e.g. ensuring proper service fulfilment according to agreements
    • H04L41/508—Network service management, e.g. ensuring proper service fulfilment according to agreements based on type of value added network service under agreement
    • H04L41/5096—Network service management, e.g. ensuring proper service fulfilment according to agreements based on type of value added network service under agreement wherein the managed service relates to distributed or central networked applications
    • H—ELECTRICITY
    • H04—ELECTRIC COMMUNICATION TECHNIQUE
    • H04L—TRANSMISSION OF DIGITAL INFORMATION, e.g. TELEGRAPHIC COMMUNICATION
    • H04L43/00—Arrangements for monitoring or testing data switching networks
    • H04L43/08—Monitoring or testing based on specific metrics, e.g. QoS, energy consumption or environmental parameters
    • H04L43/0805—Monitoring or testing based on specific metrics, e.g. QoS, energy consumption or environmental parameters by checking availability
    • H04L43/0817—Monitoring or testing based on specific metrics, e.g. QoS, energy consumption or environmental parameters by checking availability by checking functioning
    • 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
    • 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
    • Y02D30/00—Reducing energy consumption in communication networks
    • Y02D30/70—Reducing energy consumption in communication networks in wireless communication networks

Definitions

  • the present application relates to the field of computers, and in particular, to a cloud data center providing flexible instances, a scheduling method for flexible instances used in the cloud data center, a resource management system running in the cloud data center, and corresponding storage media and computer program products.
  • the number of servers in a data center managed by a cloud operator can be as high as 100,000 or even 1 million.
  • the utilization rate of physical resources in the server is generally limited to the interval of 10% to 20%.
  • the present application provides a cloud data center providing flexible instances, the flexible instances are flexibly scheduled, and the resource utilization rate of the cloud data center is improved.
  • a first aspect provides a cloud data center
  • the cloud data center includes a resource management system and a computing resource pool
  • the computing resource pool includes at least one host.
  • the resource management system is used to monitor the running parameters of the flexible instances running in the computing resource pool, and instruct the computing resource pool to adjust the size of the flexible instances according to adjustment requirements, and the adjustment requirements include QoS guarantee requirements and The operating parameters of the flexible instance.
  • the computing resource pool is used to run the flexible instance, collect the running parameters of the flexible instance, and adjust the size of the flexible instance according to the instruction, wherein the process of adjusting the size of the flexible instance conforms to the QoS guarantee requirements.
  • the size of the flexible instance can be flexibly scheduled, so that the physical resources of the host can be better allocated, which improves the resource utilization of the host, saves the power consumption of the cloud data center, and produces environmental benefits.
  • the scheduling of flexible instances complies with QoS guarantee requirements, which ensures the stability of tenants' services running on flexible instances.
  • the adjustment requirement further includes a specification configuration
  • the specification configuration includes a size range of the flexible instance, and the adjusted size of the flexible instance is within the size range.
  • the specification configuration may include a maximum size and a minimum size of the flexible instance, or a maximum size, or a standard size and a floating interval.
  • the specification configuration limits the variation range of the size of the flexible instance, and improves the stability of the flexible instance, so that the size of the flexible instance will not be adjusted to the limit and cause economic or performance risks.
  • the specification configuration further includes an instance type, and the process of adjusting the size of the flexible instance does not adjust the instance type of the flexible instance.
  • the resource management system is configured to instruct the computing resource pool to reduce the flexible instance when the operating parameter of the flexible instance indicates that the resource utilization rate of the flexible instance is lower than a threshold. size of.
  • the resource management system may instruct the deployment host of the flexible instance in the computing resource pool to reduce the size of the flexible instance, or instruct the flexible instance to migrate to a new deployment host, and the new deployment host reduces the flexibility The size of the instance.
  • the resource management system is configured to instruct the computing resource pool to increase the size of the computing resource pool when the operating parameter of the flexible instance indicates that the QoS parameter of the flexible instance does not meet the QoS guarantee requirement. size of the flexible instance described above.
  • the QoS guarantee requirement may specify an allowable degradation ratio, where the allowable degradation ratio is the degradation ratio of the lowest QoS parameter of the flexible instance acceptable to the tenant relative to the QoS parameter of the flexible instance at the maximum size.
  • the resource management system is configured to instruct the computing resource pool to increase the size of the flexible instance when the running parameter of the flexible instance indicates that the QoS parameter of the flexible instance does not meet the allowable degradation ratio.
  • the resource management system may instruct the deployment host of the flexible instance to increase the size of the flexible instance. If there are not enough free resources on the deployment host for the flexible instance to expand, the resource management system can instruct the flexible instance to be migrated to a new deployment host and increase the size of the flexible instance by the new deployment host, or the original deployment host can be Migrate the existing instance on the flexible instance to another host to free up spare resources for the flexible instance to expand.
  • the QoS guarantee requirement includes a duration requirement, wherein the time used in the process of adjusting the size of the flexible instance is shorter than or equal to the duration requirement.
  • the duration requirement limits the time from when the QoS parameters of the flexible instance do not meet the QoS guarantee requirements to the completion of the size expansion of the flexible instance, which ensures the stability of the tenant's business running on the flexible instance and avoids the scheduling process of the flexible instance. The performance of the tenant's business is degraded.
  • the resource management system is further configured to calculate the benchmark fee of the flexible instance according to the QoS guarantee requirement, wherein the higher the QoS guarantee requirement, the higher the benchmark fee.
  • the resource management system is configured to calculate the base fee of the flexible instance according to the duration requirement, wherein the smaller the value of the duration requirement is, the higher the base fee is.
  • the resource management system is configured to calculate the actual cost of the flexible instance according to the reference cost and the size of the flexible instance.
  • the size of the flexible instance changes during the life cycle of the flexible instance, so the final actual cost of the flexible instance needs to be calculated based on the base cost and the size of the flexible instance.
  • the resource management system is configured to select a deployment host of the flexible instance from the at least one host according to the specification configuration and the QoS guarantee requirement.
  • the resource management system When the resource management system creates the flexible instance, it can comprehensively consider the specification configuration and QoS guarantee requirements of the flexible instance to select a deployment host for the newly-created flexible instance, and ensure that the newly-created flexible instance and other instances on the deployment host are as good as possible. running quality.
  • a flexible instance scheduling method is provided, where the scheduling method is executed in a cloud data center, where the cloud data center includes a resource management system and a computing resource pool.
  • the scheduling method includes: the resource management system monitors the running parameters of the flexible instances running in the computing resource pool; the resource management system instructs the computing resource pool to adjust the size of the flexible instances according to adjustment requirements, the The adjustment requirements include QoS guarantee requirements and operating parameters of the flexible instance, wherein the process of adjusting the size of the flexible instance conforms to the QoS guarantee requirements.
  • the adjustment requirement further includes a specification configuration
  • the specification configuration includes a size range of the flexible instance
  • the adjusted size of the flexible instance is within the size range.
  • the resource management system instructs the computing resource pool to adjust the size of the flexible instance according to the adjustment requirement, including: when the running parameter of the flexible instance indicates that the resource utilization rate of the flexible instance is low. At a threshold, the resource management system instructs the computing resource pool to reduce the size of the flexible instance.
  • the resource management system instructs the computing resource pool to adjust the size of the flexible instance according to the adjustment requirement, including: when the running parameter of the flexible instance indicates that the QoS parameter of the flexible instance does not meet the When the QoS guarantee is required, the resource management system instructs the computing resource pool to increase the size of the flexible instance.
  • the QoS guarantee requirement includes an allowable degradation ratio.
  • the resource management system instructs the computing resource pool to adjust the size of the flexible instance according to the adjustment requirement, including: when the operation parameter of the flexible instance indicates that the QoS parameter of the flexible instance does not meet the allowable degradation ratio, the The resource management system instructs the computing resource pool to increase the size of the flexible instance.
  • the QoS guarantee requirement includes a duration requirement.
  • the time used by the computing resource pool to resize the flexible instance is shorter than or equal to the duration requirement.
  • the scheduling method further includes: the resource management system calculates the benchmark fee of the flexible instance according to the QoS guarantee requirement, wherein the higher the QoS guarantee requirement, the higher the benchmark fee. .
  • the scheduling method further includes: the resource management system calculates, by the resource management system, a benchmark fee for the flexible instance according to the duration requirement, wherein the smaller the duration requirement value is, the higher the benchmark fee is. .
  • the scheduling method further includes: the resource management system calculates the actual cost of the flexible instance according to the reference cost and the size of the flexible instance.
  • the scheduling method further includes: the resource management system selects the deployment of the flexible instance from at least one host included in the computing resource pool according to the specification configuration and the QoS guarantee requirement. host.
  • a resource management system in a third aspect, includes a monitoring module and a scheduling module: the monitoring module is used to monitor the running parameters of the flexible instances running in the computing resource pool; the scheduling module is used to instruct the computing resource pool to adjust the The size of the flexible instance, the adjustment requirement includes a QoS guarantee requirement and an operation parameter of the flexible instance, wherein the process of adjusting the size of the flexible instance by the computing resource pool according to the instruction complies with the QoS guarantee requirement.
  • the adjustment requirement further includes a specification configuration
  • the specification configuration includes a size range of the flexible instance
  • the adjusted size of the flexible instance is within the size range.
  • the scheduling module is configured to instruct the computing resource pool to reduce the resource utilization of the flexible instance when the operating parameter of the flexible instance indicates that the resource utilization of the flexible instance is lower than a threshold. size.
  • the scheduling module is configured to instruct the computing resource pool to increase the The size of the flexible instance.
  • the QoS guarantee requirement includes an allowable degradation ratio.
  • the scheduling module is configured to instruct the computing resource pool to increase the size of the flexible instance when the running parameter of the flexible instance indicates that the QoS parameter of the flexible instance does not meet the allowable degradation ratio.
  • the QoS guarantee requirement includes a duration requirement.
  • the time used by the computing resource pool to resize the flexible instance is shorter than or equal to the duration requirement.
  • the resource management system further includes a charging module.
  • the billing module is configured to calculate the benchmark fee of the flexible instance according to the QoS guarantee requirement, wherein the higher the QoS guarantee requirement, the higher the benchmark fee.
  • the billing module is configured to calculate the base fee of the flexible instance according to the duration requirement, wherein the smaller the value of the duration requirement is, the higher the base fee is.
  • the billing module is configured to calculate the actual fee of the flexible instance according to the reference fee and the size of the flexible instance.
  • the scheduling module is configured to select a deployment host of the flexible instance from at least one host included in the computing resource pool according to the specification configuration and the QoS guarantee requirement.
  • the resource management system further includes a configuration module.
  • the configuration module is configured to provide a flexible instance configuration interface, and receive the configuration parameters of the flexible instance through the flexible instance configuration interface, the flexible instance configuration interface includes a QoS guarantee requirement configuration area, and the QoS guarantee requirement configuration area is used for receiving the QoS guarantee requirement input by the tenant of the flexible instance.
  • the use of the flexible instance configuration interface improves the tenant's configuration experience for flexible instances.
  • the flexible instance configuration interface further includes a specification configuration area.
  • the specification configuration area is used for receiving specification configuration input by the tenant of the flexible instance.
  • the specification configuration can be selected according to the needs of tenants, which improves the flexibility of flexible instance configuration.
  • a fourth aspect provides a computer, including a memory and a processor, where the memory stores program instructions, and the processor executes the program instructions to execute the method provided by the second aspect and possible implementations thereof. Specifically, the processor executes the program instructions to execute the resource management system provided by the second aspect and possible implementations thereof.
  • a fifth aspect provides a readable storage medium, which may be the non-transitory.
  • the cloud data center causes the cloud data center to execute the method provided by the second aspect and possible implementations thereof.
  • Program instructions are stored in the readable storage medium.
  • the readable storage medium includes, but is not limited to, volatile storage devices, such as random access storage devices, and non-volatile storage devices, such as flash storage devices, hard disk drives (HDD), solid state drives (solid state drives) , SSD).
  • a sixth aspect provides a computer program product, when the instructions contained in the computer program product are executed by the resource management system of the cloud data center and the host where the computing resource pool is located, causing the cloud data center to execute the aforementioned second aspect and a possible implementation thereof method provided.
  • the computer program product may be a software installation package.
  • the computer program product may be downloaded and executed on the host where the resource management system and the computing resource pool are located. the computer program product.
  • 1 is a schematic diagram of the organizational structure of a cloud data center provided by the application
  • FIG. 2 is a schematic diagram of a flexible instance configuration interface provided by this application.
  • FIG. 3 is a schematic diagram of the organizational structure of the resource management system provided by the application.
  • Fig. 4 is the charging schematic diagram of the charging module provided by this application.
  • FIG. 5 is a schematic diagram of a scheduling flow of a flexible instance provided by the present application.
  • FIG. 6 is a schematic diagram of the organizational structure of another cloud data center provided by the present application.
  • FIG. 7 is a schematic diagram of the organizational structure of another cloud data center provided by the present application.
  • Host A physical server deployed in a cloud data center.
  • the physical resources of the host include physical central processing units (CPUs) and memory devices.
  • Virtualization software runs on each host, and the virtualization software virtualizes some physical resources into virtual resources for the instance to use.
  • virtualization software virtualizes CPUs into virtual CPUs (vCPUs).
  • vCPUs virtual CPUs
  • the CPU is abbreviated as U
  • vU virtual CPUs
  • Instance A computing node running on a host. Common examples include virtual machines (VMs) or containers. Each instance occupies some or all of the virtual resources of the host.
  • An instance's specification configuration includes the instance type (also known as flavor) and instance size.
  • the instance type indicates the resource characteristics of the instance. For example, economical instances use relatively cheap CPUs, occupying less computing resources and lower costs; computing-enhanced instances use high-performance CPUs, which occupy sufficient computing resources and higher costs; network-enhanced instances occupy sufficient network resources For example, if high IO bandwidth is configured, the cost is higher than that of economical instances.
  • Different instance types represent the resource requirements of tenants for their business running on flexible instances.
  • the instance size indicates the amount of resources occupied by the instance.
  • the instance size generally includes the number of vCPUs and memory size (in gigabytes (GB)), and can also include memory bandwidth, network bandwidth, graphics processing unit, The number of GPUs), the size of non-volatile storage devices (generally high-speed storage media, such as solid state drives (SSD), NVMe SSDs), etc.
  • SSD solid state drives
  • the size of a flexible instance is the resource actually occupied by the flexible instance at a certain moment or time period. Since the size of the flexible instance is floating in the life cycle of the flexible instance, the size of the flexible instance mentioned in this application includes the size of the flexible instance in different time periods.
  • Instances are classified into rigid instances and flexible instances according to whether the virtual resources occupied by the instance in its life cycle change or not.
  • the specification of a rigid instance remains unchanged during its lifetime. For example, when a rigid instance is created, the specified size is 16vU, 32GB, and the size of the rigid instance remains 16vU, 32GB during its life cycle, until the end of its life cycle, the resources occupied by the rigid instance are released.
  • the size of a flexible instance is variable during its lifetime.
  • the specified size is 16vU, 32GB (maximum size)
  • the size of the flexible instance can be floated during its life cycle
  • the maximum size within the floating interval is 16vU, 32GB
  • the minimum size can be defined according to preset rules , for example 1vU, 2GB.
  • Resource utilization the utilization of the host/instance of its physical/virtual resources at a certain moment or within a certain period of time.
  • the resource utilization of the host/instance can be calculated from parameters such as CPU utilization and memory utilization.
  • QoS parameters represent the running status of the instance. Common QoS parameters include any one or a combination of the following: packets per second (PPS), request response success rate, input output per second (IOPS), various types of time Delay (for example, network delay, service response delay), network bandwidth, storage bandwidth.
  • PPS packets per second
  • IOPS input output per second
  • time Delay for example, network delay, service response delay
  • network bandwidth storage bandwidth.
  • FIG. 1 provides a cloud data center, which includes a resource management system and a computing resource pool, and the computing resource pool includes multiple hosts.
  • the computing resource pool provides resources to the resource management system for running instances.
  • the resource management system establishes a communication connection with each host, and monitors the running parameters of each host and the running parameters of the existing instances running on each host.
  • the resource management system is also responsible for instance scheduling, including the issuance of new instances and the migration of existing instances.
  • the resource management system mainly considers the resource occupancy of the host. If the unoccupied virtual resources of a host are smaller than the size of the rigid instance, the resource management system will not be able to issue/migrate the rigid instance to the host. Since the size of the flexible instance is floating, the scheduling of the flexible instance is more flexible, which helps to improve the resource utilization of the host.
  • Rigid instance 1 (16vU, 32GB) and rigid instance 2 (16vU, 32GB) are running on host 1 (32U, 64GB).
  • the resource utilization rate of rigid instance 1 (16vU, 32GB) is 11%
  • the resource utilization rate of rigid instance 2 (16vU, 32GB) is 19%, so the resource utilization rate of host 1 is about 15%.
  • virtual resources assigned to the rigid instance cannot be assigned to other instances or used to create new instances even if they are not fully utilized by the rigid instance. Even if about 85% of the virtual resources on host 1 are not actually used (low resource utilization), because all virtual resources on the host are occupied (high resource occupancy), the resource management system will not be used on host 1.
  • the QoS parameters of the instance may remain stable.
  • the QoS parameters of an instance do not meet expectations, the size of the instance can be increased.
  • the configuration parameters of flexible instances include specification configuration and QoS guarantee requirements. Specification configuration and QoS assurance requirements each include one or more configuration items.
  • the configuration parameters of the flexible instance may be input to the resource management system through a configuration interface of the flexible instance, or may be input to the resource management system through a configuration instruction (for example, an application programming interface (API)).
  • a configuration instruction for example, an application programming interface (API)
  • the cloud data center can also preset the configuration parameters of the flexible instance, and the tenant does not need to input the configuration parameters.
  • the types of flexible instances can include general-purpose, compute-intensive, memory-intensive, network-intensive, machine-learning compute-intensive, storage-intensive, etc.
  • the instance size of a flexible instance indicates the size range of the flexible instance.
  • the size range of a flexible instance may include two configuration items in total, a minimum size and a maximum size, or only the configuration item of the maximum size of the flexible instance, or include the standard size of the flexible instance and set a floating interval (floating up/floating down). ratio) for a total of two configuration items.
  • the instance size of a flexible instance in the flexible instance configuration interface shown in FIG. 2 includes a minimum size and a maximum size. Tenants can enter the minimum number of vCPUs and the maximum number of vCPUs in the vCPU fill box, and can also enter the minimum memory size and maximum memory size in the memory fill box. Generally, the minimum number of vCPUs cannot be less than 1, the maximum vCPU cannot be less than the minimum number of vCPUs, the minimum memory size cannot be less than 1GB, and the maximum memory size cannot be less than the minimum memory size.
  • the size range of a flexible instance includes only the maximum size or the case where the standard size and the floating range of the standard size are included, and so on.
  • the QoS guarantee requirements Since the size of the flexible instance is floating, the tenant needs to input the QoS guarantee requirements of the flexible instance.
  • the QoS guarantee requirements indicate the tenant's requirements for the QoS parameters of the flexible instance.
  • the resource management system will schedule flexible instances according to QoS guarantee requirements.
  • the QoS guarantee requirement is the service level commitment of the cloud data center to the tenant, which reflects the quality guarantee of the QoS parameters of the cloud data center to the tenant's flexible instances during its life cycle, such as degradation guarantee (permissible degradation ratio), recovery guarantee (duration requirement) ).
  • degradation guarantee permissible degradation ratio
  • recovery guarantee recovery guarantee
  • the process and result of adjusting the size of the flexible instance should meet the QoS guarantee requirements.
  • the cloud data center can preset QoS guarantee requirements, and tenants can also configure QoS guarantee requirements according to their own requirements.
  • QoS assurance requirements include any one or more of configuration items such as QoS assurance priority, allowable degradation ratio, duration requirements, high availability (HA) HA assurance capability, and additional attention to QoS parameters: general, QoS assurance requirements At least the duration requirement is included.
  • QoS guarantee priority can be divided into multiple levels, high/medium/low, or high/low, etc.
  • the resource management system has different scheduling methods for flexible instances with different QoS guarantee priorities.
  • the resource management system will give priority to ensuring that flexible instances with higher priority have a lower QoS parameter degradation ratio and a shorter QoS parameter degradation duration.
  • the resource management system will also give priority to Ensure that the degradation ratio of QoS parameters of flexible instances with higher priorities can be quickly recovered. For example, flexible instances with higher priorities are migrated to idle hosts, or flexible instances with higher priorities are located on hosts with lower priorities. Instances are migrated to other hosts.
  • the resource management system does not strictly manage the QoS parameter degradation ratio and QoS parameter degradation duration of the flexible instance, and only guarantees the QoS parameter degradation of the flexible instance with higher priority. Under the premise that the situation (the degradation ratio and/or the degradation duration) does not exceed the allowable degradation parameter, scheduling is performed on the low-priority flexible instance.
  • the tenant can fill in the requirements for the allowable degradation ratio and duration.
  • the tenant only needs to fill in the allowable degradation ratio and duration requirements when selecting some of the QoS guarantee priorities.
  • the tenant does not need to fill in the allowable degradation ratio and duration requirements. For example, when the QoS assurance priority is high/medium, you need to fill in the allowable degradation ratio and duration requirements.
  • the QoS assurance priority is low, you do not need to fill in the allowable degradation ratio and duration requirements.
  • the QoS assurance requirement does not include the QoS assurance priority, or if the tenant does not select any QoS assurance priority, after the tenant selects the QoS assurance priority, the tenant needs to fill in the requirements for the allowable degradation ratio and duration.
  • Allowable degradation ratio The degradation ratio of the lowest QoS parameter of the flexible instance relative to the QoS parameter of the flexible instance at the maximum size. For example, if the maximum size of a flexible instance is 16vU, 32GB and the allowable degradation rate is 20%, then the actual running QoS parameters of the flexible instance during the running process are at least 80% of the QoS parameters of the flexible instance under the maximum size.
  • the calculation of the actual running QoS parameters of the flexible instance during the running process can adopt the average value or the instantaneous value.
  • the average running QoS parameters of the flexible instance in the period are lower than the allowable degradation ratio and the duration should not exceed the duration requirement.
  • the flexible instance scheduling will be more flexible, and the resource utilization of the cloud data center will be reduced. higher.
  • the actual running QoS parameter adopts the instantaneous value it is required that the duration that the instantaneous running QoS parameter is lower than the allowable degradation ratio shall not exceed the QoS parameter under the duration requirement * maximum size, and the performance of the flexible instance will be more guaranteed.
  • the ratio of the actual running QoS parameters of the flexible instance and the QoS parameters under the maximum size of the flexible instance is lower than the allowable degradation ratio (if the allowable degradation ratio is not specified or the QoS guarantee requirement does not include the allowable degradation ratio, the value here is 100%).
  • the time used in the process of adjusting the size of the flexible instance should be shorter than or equal to the set duration need.
  • the actual operating QoS parameters here can be average or instantaneous values.
  • the flexible instance may be scheduled to Other hosts, or other instances running on the host where the flexible instance resides, will be scheduled to other hosts.
  • the allowable degradation ratio can also be replaced by the minimum degradation size, that is, the minimum degradation level of the size of the flexible instance acceptable to the tenant.
  • Minimum degraded size includes vCPU and memory.
  • the minimum degradation size of the flexible instance can be directly defined as 12.8vU, 25.6GB.
  • the flexible instance configuration interface will provide one or more of a recommended default value of the allowable degradation ratio, a recommended value, and a value range.
  • a recommended default value of the allowable degradation ratio For example, high priority: default 10%, recommended 5%-15%, range 3%-30%; medium priority: default 15%, recommended 10%-50%, range 5% -50%.
  • the flexible instance configuration interface will provide one or more of the default value, recommended value and value range of the recommended duration requirement.
  • high priority the default value is 30 seconds, the recommended value is 10 seconds-60 seconds, and the value range is 5 seconds-100 seconds
  • medium priority the default value is 60 seconds, the recommended value is 30 seconds-180 seconds, and the value range is 15 seconds - 300 seconds.
  • Figure 2 shows the default allowable degradation ratio, the default duration requirement, and the recommended value when the QoS guarantee priority is high.
  • HA guarantee capability A means of guaranteeing the high availability requirements of flexible instances.
  • the flexible instance needs to be migrated when the original deployment host of the flexible instance fails or has insufficient performance.
  • HA assurance capabilities can include hot migration and cold migration.
  • the hot migration instruction allows the resource management system to hot migrate the flexible instance
  • the cold migration instruction allows the resource management system to cold migrate the flexible instance.
  • the HA guarantee capability is hot migration, the cost of flexible instance migration is higher, and the suspension time of the flexible instance caused by the flexible instance migration process is shorter.
  • the HA guarantee capability is cold migration, the cost of flexible instance migration is lower, and the suspension time of the flexible instance caused by the flexible instance migration process is longer.
  • Additional attention to QoS parameters The quality parameters of this flexible instance that require additional attention. Tenants can select one or more QoS parameters to fill in as additional QoS parameters. In order to further ensure the performance of the flexible instance during the running process and improve the tenant experience, the tenant can select the QoS parameters that the flexible instance needs to pay extra attention to according to the business characteristics of the flexible instance, so that the resource management system can schedule the flexible instance in the process of scheduling the flexible instance. As much as possible, ensure the stability of this additional concern QoS parameter.
  • the resource management system includes a configuration module, a billing module, a scheduling module, a monitoring module, and optionally, a prediction auxiliary module.
  • the configuration module is used to receive configuration parameters of the flexible instance, for example, to provide the flexible instance configuration interface shown in FIG. 2 , or to receive configuration instructions.
  • the billing module is used to calculate the base cost according to the configuration parameters of the flexible instance, and calculate the actual cost according to the running parameters of the flexible instance.
  • the scheduling module is used to schedule flexible instances based on various types of information.
  • the monitoring module is used to monitor the running parameters of the host and existing flexible instances, and provide various running parameters to the scheduling module and the prediction auxiliary module.
  • the prediction auxiliary module is used to predict the future resource utilization rate of each host, and predict the future QoS parameter degradation of the flexible instance.
  • the billing module can generate a benchmark fee for the flexible instance, so that the tenant can estimate the usage cost of the flexible instance.
  • the base fee indicates the estimated base usage fee per unit time of the flexible instance.
  • the unit time can be seconds/minutes/hours/days/weeks/months/years, etc.
  • the billing module calculates the benchmark fee with reference to scale configuration and QoS guarantee requirements. Specifically, the billing module may use one or more configuration items in the specification configuration and any one or more configuration items in the QoS guarantee requirement to calculate the base charge. Under a certain scale configuration, the higher the QoS guarantee requirement, the higher the benchmark fee of the flexible instance, and the lower the QoS guarantee requirement, the lower the benchmark fee of the flexible instance.
  • Instance type The base fee for different instance types is generally different
  • Minimum size Generally, the larger the value of this configuration item, the higher the QoS guarantee requirement, and the higher the benchmark cost;
  • QoS guarantee priority Generally, the higher the configuration item, the higher the QoS guarantee requirement and the higher the benchmark cost;
  • Allowable degradation ratio Generally, the lower the value of this configuration item, the higher the QoS guarantee requirement and the higher the benchmark cost;
  • HA Guarantee Capability Generally, the QoS guarantee requirement of a flexible instance whose configuration item is hot migration is higher than the QoS guarantee requirement of a flexible instance whose configuration item is cold migration, and the base cost of a flexible instance whose configuration item is hot migration is also higher. Baseline fee for flexible instances that are cold migrated for this configuration item;
  • the QoS guarantee requirement of a flexible instance with this configuration item is filled in is higher than that of a flexible instance without this configuration item, and the base fee of a flexible instance with this configuration item filled in is also lower than that without this configuration item.
  • the base fee for a flexible instance is also lower than that without this configuration item.
  • the base fee is equal to the product of the fee calculated from the specification configuration and the discount ratio calculated from the QoS guarantee requirement.
  • the larger the minimum size and maximum size the higher the calculated cost of the specification configuration.
  • Each configuration item in the QoS guarantee requirement will affect the specific discount rate.
  • the billing module can generate actual charges according to the base charge and the actual running size of the flexible instance.
  • the actual cost can include the base cost and at least a period of floating cost, each of which includes the size of the flexible instance and the corresponding cost within a period of time. For example, a flexible instance is created at 6:00 and resized at 9:00, then the actual cost includes:
  • Figure 5 shows the scheduling process of flexible instances, including:
  • the monitoring module continuously monitors the running parameters of each host and each existing flexible instance, sends the running parameters of the host and the existing flexible instances to the scheduling module and the prediction auxiliary module (optional), and sends the running parameters of each existing flexible instance to billing module.
  • the running parameters of the host include resource utilization of the host, and the running parameters of the flexible instance include one or more of resource utilization, current size, QoS parameter degradation ratio, and QoS parameter degradation duration of the flexible instance.
  • Each host is installed with an agent of the monitoring module, and the agent periodically collects the running parameters of the host where it is located and each existing flexible instance running on the host where it is located, and reports it to the monitoring module.
  • S200 may be performed in parallel with other steps in the scheduling process.
  • the client of the tenant sends the configuration parameters of the newly created flexible instance to the configuration module.
  • the tenant's client can run on the tenant's local server or on a host in the cloud data center.
  • the tenant's client sends the configuration parameters of the newly created flexible instance to the configuration module through the configuration instruction or the flexible instance configuration interface.
  • the configuration module sends the configuration parameters of the newly created flexible instance to the billing module.
  • the execution of S201 and S202 is optional, and the cloud data center may preset the configuration parameters of the flexible instance without the tenant's input.
  • the billing module calculates the base fee for the newly created flexible instance according to the configuration parameters of the newly created flexible instance.
  • the billing module provides the base cost of the newly created flexible instance to the client.
  • the billing module can send the benchmark fee for the newly created flexible instance to the configuration module, and the configuration module displays the benchmark fee for the newly created flexible instance on the flexible instance configuration interface provided by the configuration module.
  • S202 to S204 are optional steps.
  • the configuration module needs to confirm the confirmation instruction of the tenant after obtaining the benchmark fee for the newly-created flexible instance, and the confirmation instruction is used to confirm the deployment of the newly-created flexible instance according to the configuration parameter instance.
  • the scheduling process (S205 to S207) of the resource management system without including the prediction auxiliary module and the scheduling process (S208 to S212) when the prediction auxiliary module is included are respectively introduced.
  • the configuration module sends the configuration parameters of the newly created flexible instance to the scheduling module.
  • the scheduling module selects the deployment host of the newly-created flexible instance according to the configuration parameters of the newly-created flexible instance, each host and the running parameters of each existing flexible instance.
  • the scheduling module instructs the selected deployment host to deploy the newly created flexible instance according to the configuration parameters of the newly created flexible instance.
  • the following describes in detail how the scheduling module selects a host for deploying the newly created flexible instance in S206.
  • the scheduling module calculates the resource utilization rate of each host after the newly created flexible instance is deployed to each host according to the resource utilization rate of the host and the configuration parameters of the newly created flexible instance. Hosts that exceed a resource utilization threshold (eg, 50%) are removed from the candidate host list. If no host can enter the candidate host list, then further determine which hosts can stay on the candidate host if the existing flexible instance whose QoS guarantee priority is lower than the QoS guarantee priority of the newly created flexible instance is migrated to other hosts or terminated. list. If it still cannot find any candidate host, the scheduling module can notify the client that the deployment of the new flexible instance fails.
  • a resource utilization threshold eg, 50%
  • the scheduling module calculates the existing flexible instances running on each host after the newly created flexible instances are deployed to each host according to the resource utilization of the hosts, the configuration parameters of the newly created flexible instances, and the degradation ratio of the QoS parameters of the existing flexible instances. Degradation of QoS parameters of flexible instances. Hosts running existing flexible instances that may exceed the allowable degradation rate are further removed from the candidate host list. Then, select the host with the lowest resource utilization in the candidate host list as the deployment host for the newly created flexible instance. If each host in the candidate host list deploys a new flexible instance, the QoS parameter degradation ratio of at least one existing flexible instance exceeds its allowable degradation ratio, then further judge if the QoS guarantee priority is lower than that of the newly created flexible instance. Existing flexible instances of QoS guaranteed priorities are migrated to other hosts or terminated, which hosts can remain in the candidate host list. Then, select the host with the lowest resource utilization in the candidate host list as the deployment host for the newly created flexible instance.
  • the process of selecting a deployment host for the migrated existing flexible instance refers to the process of selecting a deployment host for a newly created flexible instance (S206).
  • the scenario where a flexible instance needs to be migrated or terminated can occur when a new flexible instance is created, or when the QoS parameter degradation ratio of an existing flexible instance exceeds the allowable degradation ratio, and the host has no remaining resources.
  • the QoS guarantee level of the existing flexible instance on the host and the type of the existing flexible instance on the host may also be comprehensively considered to select the flexible instance to be migrated or terminated.
  • the configuration module sends the configuration parameters of the newly created flexible instance to the scheduling module and the prediction auxiliary module.
  • the prediction auxiliary module predicts the future operation parameters of each host, each existing flexible instance, and the newly created flexible instance.
  • the prediction auxiliary module is internally provided with an intelligent module generated based on the learning history data. Predict the future running parameters of the newly created flexible instance.
  • the prediction time can be n unit times in the future, where n is an integer greater than 0.
  • the prediction auxiliary module sends the predicted operation parameters of each host, the predicted operation parameters of the existing flexible instances, and the predicted operation parameters of the newly created flexible instances to the scheduling module.
  • the prediction auxiliary module may be performed in parallel with other steps to predict the future operating parameters of each host and each existing flexible instance.
  • the scheduling module selects the deployment host of the newly created flexible instance according to the configuration parameters of the newly created flexible instance, the predicted running parameters of each host, the predicted running parameters of the existing flexible instance, and the predicted running parameters of the newly created flexible instance.
  • the scheduling module instructs the selected deployment host to deploy the newly created flexible instance according to the configuration parameters of the newly created flexible instance.
  • how the scheduling module selects the host for deploying the newly created flexible instance is different from that in S206 in that the scheduling module can further use various predicted operation parameters to select the deployment host for the newly created flexible instance.
  • the predicted resource utilization rate of each host in the future period after the newly created flexible instance is deployed to each host can be calculated according to the predicted resource utilization rate of the host and the configuration parameters of the newly created flexible instance, and the candidate host can be determined accordingly. list.
  • the scheduling module can calculate, according to the predicted resource utilization rate of the host, the configuration parameters of the newly created flexible instance, and the predicted running parameters of the existing flexible instance, the number of new flexible instances running on each host after the newly created flexible instance is deployed to each host. Based on the deterioration of QoS parameters of existing flexible instances in the future, the deployment host of the new flexible instance is selected accordingly.
  • the scheduling module determines the flexible instance that needs to be adjusted according to the adjustment requirement.
  • the adjustment requirements include QoS guarantee requirements and operating parameters of the flexible instance.
  • the scheduling module instructs the host of the computing resource pool to adjust the size of the flexible instance to be adjusted, and the process of adjusting the size of the flexible instance meets the QoS guarantee requirement. Specifically, the time used in the process of adjusting the size of the flexible instance is shorter than or equal to the duration requirement.
  • the scheduling module determines the flexible instances to be adjusted according to the running parameters and QoS guarantee requirements of each flexible instance running in the cloud data center provided by the monitoring module. Flexible instances that need to be adjusted are divided into situations:
  • the scheduling module instructs the computing resource pool to increase the size of the flexible instance when it determines that the running parameter of the flexible instance indicates that the QoS parameter of the flexible instance does not meet the QoS guarantee requirement. Specifically, the scheduling module determines that the QoS parameter degradation ratio of the flexible instance has exceeded or is about to exceed the allowable degradation ratio. In this case, the scheduling module instructs the deployment host of the flexible instance to be adjusted to increase the size of the flexible instance to be adjusted, that is, the scheduling module instructs the deployment host to allocate more resources to the flexible instance.
  • the indication may include an adjustment parameter, and the adjustment parameter may be determined by the size range of the flexible instance, the QoS parameter degradation ratio, and the QoS parameter degradation duration that the scheduling module needs to adjust.
  • the deployment host increases the size of the flexible instance that needs to be adjusted according to the adjustment parameters. If there are no remaining resources available for allocation on the deployment host, the scheduling module instructs to abort or migrate the existing flexible instance on the deployment host, and then increase the size of the flexible instance, or migrate the flexible instance to another host with free resources , and then increase the size of the flexible instance by the host to which it is migrated.
  • the scheduling module adjusts the size of the flexible instance to be adjusted within the size range of the flexible instance to be adjusted. Generally, the larger the QoS parameter degradation ratio is, the larger the size of the flexible instance with the longer QoS parameter degradation duration is increased.
  • the process of resizing the flexible instance takes less than or equal to the duration requirement.
  • the time used by the process of resizing the flexible instance includes the time required to migrate an existing flexible instance (possible), the time required to migrate the flexible instance (possible), and the time to increase the size of the flexible instance.
  • the scheduling module instructs the computing resource pool to reduce the size of the flexible instance when the operation parameter of the flexible instance indicates that the resource utilization of the flexible instance is lower than a threshold.
  • the scheduling module instructs the deployment host of the flexible instance to be adjusted to reduce the size of the flexible instance to be adjusted, that is, the scheduling module instructs the deployment host to reduce the resources allocated to the flexible instance.
  • the indication may include an adjustment parameter, and the adjustment parameter may be determined by the size range and resource utilization of the flexible instance to be adjusted by the scheduling module as required.
  • the deployment host reduces the size of the flexible instance that needs to be adjusted according to the adjustment parameters.
  • the scheduling module adjusts the size of the flexible instance to be adjusted within the size range of the flexible instance to be adjusted. Generally, the smaller the resource utilization rate is, the greater the reduction in the size of the flexible instance.
  • S213 and S214 may be performed in parallel with other steps in the scheduling process.
  • the billing module generates an actual fee according to the base fee of the flexible instance and the running parameters of the flexible instance.
  • the billing module provides the actual cost of the flexible instance to the client.
  • the billing module will calculate the running parameters of the flexible instance provided by the monitoring module. Generate the actual cost, and provide the actual cost to the client.
  • the maximum size of the flexible instance 1, the flexible instance 2, the flexible instance 3, and the flexible instance 4 is the same as that of the rigid instance in FIG. 1 .
  • the size allocated by the resource management system to the four flexible instances at the current moment is smaller than the set maximum size, which greatly improves the resource utilization of host 1.
  • the resource management system can allocate more resources from host 1 to each flexible instance, or migrate some flexible instances to other hosts to ensure the flexible instances running on host 1.
  • the cloud data center provided by the present application has higher utilization rate of host resources, reduces energy consumption in cloud data, and generates environmental benefits.
  • FIG. 7 provides a cloud data center including a computer 400 and at least one computer 600 .
  • the computer 400 and the computer 600 are connected through a communication network.
  • the computer 400 includes a processor 401 , a network device 402 , a bus 403 , and a storage device 404 .
  • the processor 401 , the network device 402 , and the storage device 404 communicate through the bus 403 .
  • the processor 401 may be a central processing unit (central processing unit, CPU).
  • the storage device 403 may include a volatile storage device (volatile memory), such as a random access storage device (random access memory, RAM), or a non-volatile storage device (non-volatile memory), such as a read-only storage device (read -only memory, ROM), flash storage device, HDD or SSD, etc.
  • Network device 402 is a network interface card.
  • the storage device 404 stores executable instructions, and the processor 401 executes the executable instructions to execute various modules of the resource management system to execute the method shown in FIG. 5 .
  • Storage device 404 may also include executable instructions required to run an operating system (OS).
  • the OS can be LINUX TM , UNIX TM , WINDOWS TM and so on.
  • the computer 600 that is, the host, includes a processor 601 , a network device 602 , a bus 603 , and a storage device 604 .
  • the organizational structure of the computer 600 is the same as that of the computer 400 .
  • the storage device 604 stores executable instructions, and the processor 601 executes the executable instructions to at least one flexible instance and an agent of the monitoring module.
  • the above-mentioned embodiments it may be implemented in whole or in part by software, hardware, firmware or any combination thereof.
  • software or firmware it may be implemented in whole or in part in the form of a computer program product.
  • the computer program product includes one or more computer instructions. When the computer program instructions are loaded and executed on a computer, the procedures or functions according to the embodiments of the present invention result in whole or in part.
  • the computer may be a general purpose computer, special purpose computer, computer network, or other programmable device.
  • the computer instructions may be stored on or transmitted from one computer readable storage medium to another computer readable storage medium, for example, the computer instructions may be transmitted over a wire from a website site, computer, server or data center (eg coaxial cable, fiber optic, twisted pair) or wireless (eg infrared, wireless, microwave, etc.) means to another website site, computer, server or data center.
  • the computer-readable storage medium can be any medium that can be accessed by a computer, or a data storage device such as a server, data center, etc. that includes one or more mediums integrated.
  • the media may be magnetic media (eg, floppy disks, hard disks, magnetic tapes), optical media (eg, optical disks), or semiconductor media (eg, SSDs), and the like.

Landscapes

  • Engineering & Computer Science (AREA)
  • Theoretical Computer Science (AREA)
  • Software Systems (AREA)
  • Physics & Mathematics (AREA)
  • General Engineering & Computer Science (AREA)
  • General Physics & Mathematics (AREA)
  • Quality & Reliability (AREA)
  • Computer Networks & Wireless Communication (AREA)
  • Signal Processing (AREA)
  • Computing Systems (AREA)
  • Mathematical Physics (AREA)
  • Debugging And Monitoring (AREA)
  • Information Retrieval, Db Structures And Fs Structures Therefor (AREA)
  • Computer And Data Communications (AREA)

Abstract

本申请提供一种云数据中心，包括资源管理系统和计算资源池。其中，资源管理系统监控所述计算资源池中运行的柔性实例的运行参数，并根据调整需求指示所述计算资源池调整所述柔性实例的尺寸，所述调整需求包括QoS保障需求和所述柔性实例的运行参数。所述计算资源池运行所述柔性实例，采集所述柔性实例的运行参数，并根据所述指示调整所述柔性实例的尺寸，调整的过程符合所述QoS保障需求。该云数据中心提供了柔性实例的运行，提高了云数据中心的资源利用率，降低了能源消耗，产生了环保的效益。

Description

提供柔性实例的云数据中心以及柔性实例的调度方法 技术领域
本申请涉及计算机领域，尤其涉及一种提供柔性实例的云数据中心、该云数据中心中使用的柔性实例的调度方法、该云数据中心运行的资源管理系统以及相应的存储介质和计算机程序产品。
背景技术
随着公有云技术的兴起，云运营商管理的数据中心的规模也逐渐增大，一家云运营商管理的一个数据中内的服务器数量可以高达10万台甚至100万台。然而，为了保证租户运行在服务器上的实例的运行质量，目前服务器内物理资源的利用率普遍限制于10％至20％这个区间。
那么，如何提升这些服务器的利用率成为了一个重要的问题。提升这些服务器的资源利用率不仅可以降低云运营商的设备成本，同时也可以降低租户的使用成本，减少数据中心的电能消耗，产生经济和环保等多方面的效益。
发明内容
本申请提供了一种提供柔性实例的云数据中心，该柔性实例的调度灵活，提升了云数据中心的资源利用率。
第一方面提供了一种云数据中心，所述云数据中心包括资源管理系统和计算资源池，所述计算资源池包括至少一个主机。所述资源管理系统，用于监控所述计算资源池中运行的柔性实例的运行参数，并根据调整需求指示所述计算资源池调整所述柔性实例的尺寸，所述调整需求包括QoS保障需求和所述柔性实例的运行参数。所述计算资源池，用于运行所述柔性实例，采集所述柔性实例的运行参数，并根据所述指示调整所述柔性实例的尺寸，其中，调整所述柔性实例的尺寸的过程符合所述QoS保障需求。
柔性实例的尺寸能够灵活调度，使得主机的物理资源能够得到更好的分配，提升了主机的资源利用率，节省了云数据中心对电能的消耗，产生了环保的效益。同时，柔性实例的调度符合QoS保障需求，保证了租户运行在柔性实例上的业务的稳定性。
一种可能的实现方式中，所述调整需求还包括规格配置，所述规格配置包括所述柔性实例的尺寸范围，调整后的所述柔性实例的尺寸在所述尺寸范围内。所述规格配置可以包括所述柔性实例的最大尺寸和最小尺寸，或包括最大尺寸，或包括标准尺寸以及浮动区间。
规格配置限制了柔性实例的尺寸的变化区间，提升了柔性实例的稳定性，使得柔性实例的尺寸不至于被调整到过于极限造成经济或性能方面的风险。
一种可能的实现方式中，规格配置还包括实例类型，调整所述柔性实例的尺寸的过程不调整所述柔性实例的实例类型。
一种可能的实现方式中，所述资源管理系统，用于当所述柔性实例的运行参数指 示所述柔性实例的资源利用率低于阈值时，指示所述计算资源池减小所述柔性实例的尺寸。
所述资源管理系统可以指示所述计算资源池中所述柔性实例的部署主机减小所述柔性实例的尺寸，或者指示柔性实例迁移到新的部署主机上，并由新的部署主机减小柔性实例的尺寸。
一种可能的实现方式中，所述资源管理系统，用于当所述柔性实例的运行参数指示所述柔性实例的QoS参数不符合所述QoS保障需求时，指示所述计算资源池增大所述柔性实例的尺寸。
QoS保障需求可以指定允许劣化比例，该允许劣化比例为租户能够接受的柔性实例的最低QoS参数相对于柔性实例在最大尺寸下的QoS参数的劣化比例。那么，所述资源管理系统，用于当所述柔性实例的运行参数指示所述柔性实例的QoS参数不符合所述允许劣化比例时，指示所述计算资源池增大所述柔性实例的尺寸。
所述资源管理系统可以指示所述柔性实例的部署主机增大所述柔性实例的尺寸。如果部署主机上没有足够的空余资源供柔性实例扩容，那么资源管理系统可以指示将该柔性实例迁移到新的部署主机上并由新的部署主机增大柔性实例的尺寸，或者可以将原部署主机上的现存实例迁移到其他主机，以释放空余资源供该柔性实例扩容。
一种可能的实现方式中，所述QoS保障需求包括时长需求，其中，调整所述柔性实例的尺寸的过程所使用的时间短于或等于所述时长需求。
时长需求限制了从柔性实例的QoS参数不符合所述QoS保障需求至对柔性实例的尺寸扩容完毕所需的时间，保证了租户运行在柔性实例上的业务的稳定性，避免造成柔性实例调度过程中租户的业务的性能下降。
一种可能的实现方式中，所述资源管理系统，还用于根据所述QoS保障需求计算所述柔性实例的基准费用，其中，所述QoS保障需求越高所述基准费用越高。
一种可能的实现方式中，所述资源管理系统，用于根据所述时长需求计算所述柔性实例的基准费用，其中，所述时长需求的值越小所述基准费用越高。
时长需求越低，代表要求柔性实例的调度速度更快，劣化的QoS参数恢复速度更快，基准费用也会提升。
一种可能的实现方式中，所述资源管理系统，用于根据所述基准费用和所述柔性实例的尺寸计算所述柔性实例的实际费用。
柔性实例的尺寸在柔性实例的生命周期内变化，因此，柔性实例最终的实际费用需要根据基准费用和柔性实例的尺寸计算。
一种可能的实现方式中，所述资源管理系统，用于根据所述规格配置和所述QoS保障需求从所述至少一个主机中选择所述柔性实例的部署主机。
资源管理系统在创建该柔性实例的时候，可以综合考虑柔性实例的规格配置和QoS保障需求为该新建的柔性实例选择部署主机，尽可能的保证该新建的柔性实例以及部署主机上的其余实例的运行质量。
第二方面，提供了一种柔性实例的调度方法，所述调度方法执行于云数据中心，所述云数据中心包括资源管理系统和计算资源池。所述调度方法包括：所述资源管理系统监控所述计算资源池中运行的柔性实例的运行参数；所述资源管理系统根据调整 需求指示所述计算资源池调整所述柔性实例的尺寸，所述调整需求包括QoS保障需求和所述柔性实例的运行参数，其中，调整所述柔性实例的尺寸的过程符合所述QoS保障需求。
一种可能的实现方式中，所述调整需求还包括规格配置，所述规格配置包括所述柔性实例的尺寸范围，调整后的所述柔性实例的尺寸在所述尺寸范围内。
一种可能的实现方式中，所述资源管理系统根据调整需求指示所述计算资源池调整所述柔性实例的尺寸，包括:当所述柔性实例的运行参数指示所述柔性实例的资源利用率低于阈值时，所述资源管理系统指示所述计算资源池减小所述柔性实例的尺寸。
一种可能的实现方式中，所述资源管理系统根据调整需求指示所述计算资源池调整所述柔性实例的尺寸，包括:当所述柔性实例的运行参数指示所述柔性实例的QoS参数不符合所述QoS保障需求时，所述资源管理系统指示所述计算资源池增大所述柔性实例的尺寸。
一种可能的实现方式中，所述QoS保障需求包括允许劣化比例。所述资源管理系统根据调整需求指示所述计算资源池调整所述柔性实例的尺寸，包括:当所述柔性实例的运行参数指示所述柔性实例的QoS参数不符合所述允许劣化比例时，所述资源管理系统指示所述计算资源池增大所述柔性实例的尺寸。
一种可能的实现方式中，所述QoS保障需求包括时长需求。所述计算资源池调整所述柔性实例的尺寸的过程所使用的时间短于或等于所述时长需求。
一种可能的实现方式中，所述调度方法还包括：所述资源管理系统根据所述QoS保障需求计算所述柔性实例的基准费用，其中，所述QoS保障需求越高所述基准费用越高。
一种可能的实现方式中，所述调度方法还包括：所述资源管理系统根据所述时长需求计算所述柔性实例的基准费用，其中，所述时长需求的值越小所述基准费用越高。
一种可能的实现方式中，所述调度方法还包括：所述资源管理系统根据所述基准费用和所述柔性实例的尺寸计算所述柔性实例的实际费用。
一种可能的实现方式中，所述调度方法还包括：所述资源管理系统根据所述规格配置和所述QoS保障需求从所述计算资源池包括的至少一个主机中选择所述柔性实例的部署主机。
第三方面，提供了一种资源管理系统。该资源管理系统包括监控模块和调度模块：所述监控模块，用于监控所述计算资源池中运行的柔性实例的运行参数；所述调度模块，用于根据调整需求指示计算资源池调整所述柔性实例的尺寸，所述调整需求包括QoS保障需求和所述柔性实例的运行参数，其中，所述计算资源池根据所述指示调整所述柔性实例的尺寸的过程符合所述QoS保障需求。
一种可能的实现方式中，所述调整需求还包括规格配置，所述规格配置包括所述柔性实例的尺寸范围，调整后的所述柔性实例的尺寸在所述尺寸范围内。
一种可能的实现方式中，所述调度模块，用于当所述柔性实例的运行参数指示所述柔性实例的资源利用率低于阈值时，指示所述计算资源池减小所述柔性实例的尺寸。
一种可能的实现方式中，所述调度模块，用于当所述柔性实例的运行参数指示所述柔性实例的QoS参数不符合所述QoS保障需求时，指示所述计算资源池增大所述柔 性实例的尺寸。
一种可能的实现方式中，所述QoS保障需求包括允许劣化比例。所述调度模块，用于当所述柔性实例的运行参数指示所述柔性实例的QoS参数不符合所述允许劣化比例时，指示所述计算资源池增大所述柔性实例的尺寸。
一种可能的实现方式中，所述QoS保障需求包括时长需求。所述计算资源池调整所述柔性实例的尺寸的过程所使用的时间短于或等于所述时长需求。
一种可能的实现方式中，所述资源管理系统还包括计费模块。所述计费模块，用于根据所述QoS保障需求计算所述柔性实例的基准费用，其中，所述QoS保障需求越高所述基准费用越高。
一种可能的实现方式中，所述计费模块，用于根据所述时长需求计算所述柔性实例的基准费用，其中，所述时长需求的值越小所述基准费用越高。
一种可能的实现方式中，所述计费模块，用于根据所述基准费用和所述柔性实例的尺寸计算所述柔性实例的实际费用。
一种可能的实现方式中，所述调度模块，用于根据所述规格配置和所述QoS保障需求从所述计算资源池包括的至少一个主机中选择所述柔性实例的部署主机。
一种可能的实现方式中，所述资源管理系统还包括配置模块。所述配置模块，用于提供柔性实例配置界面，通过所述柔性实例配置界面接收所述柔性实例的配置参数，所述柔性实例配置界面包括QoS保障需求配置区域，所述QoS保障需求配置区域用于接收所述柔性实例的租户输入的所述QoS保障需求。柔性实例配置界面的使用提升了租户对柔性实例的配置体验。
一种可能的实现方式中，所述柔性实例配置界面还包括规格配置区域。所述规格配置区域用于接收所述柔性实例的租户输入的规格配置。规格配置可以按照租户的需求来选择，提升了柔性实例的配置的灵活性。
第四方面提供了一种计算机，包括存储器和处理器，该存储器存储有程序指令，该处理器运行该程序指令以执行第二方面及其可能的实现方式提供的方法。具体的，该处理器运行该程序指令以运行第二方面及其可能的实现方式提供的资源管理系统。
第五方面提供了一种可读存储介质，该可读存储介质可以是该非瞬态的。该可读存储介质中存储的指令被云数据中心的资源管理系统和计算资源池所在的主机执行时，导致云数据中心执行前述第二方面及其可能的实现方式提供的方法。该可读存储介质中存储了程序指令。该可读存储介质包括但不限于易失性存储设备，例如随机访问存储设备，和非易失性存储设备，例如快闪存储设备、硬盘(hard disk drive，HDD)、固态硬盘(solid state drive，SSD)。
第六方面提供了一种计算机程序产品，该计算机程序产品包含的指令被云数据中心的资源管理系统和计算资源池所在的主机执行时，导致云数据中心执行前述第二方面及其可能的实现方式提供的方法。该计算机程序产品可以为一个软件安装包，在需要使用前述第二方面及其可能的实现方式提供的方法的情况下，可以下载该计算机程序产品并在资源管理系统和计算资源池所在的主机执行该计算机程序产品。
附图说明
图1为本申请提供的云数据中心的组织结构示意图；
图2为本申请提供的柔性实例配置界面示意图；
图3为本申请提供的资源管理系统的组织结构示意图；
图4为本申请提供的计费模块的计费示意图；
图5为本申请提供的柔性实例的调度流程示意图；
图6为本申请提供的另一云数据中心的组织结构示意图；
图7为本申请提供的另一云数据中心的组织结构示意图。
具体实施方式
首先，介绍本申请中涉及的一些术语。
主机：部署在云数据中心中的物理服务器。主机的物理资源包括物理中央处理器(CPU)和内存设备。每个主机上运行有虚拟化软件，虚拟化软件将部分物理资源虚拟化为虚拟资源供实例使用。例如，虚拟化软件将CPU虚拟化为虚拟CPU(vCPU)。在后文及附图中，将CPU简写为U，vCPU简写为vU。主机上还有内存通道、缓存通道、缓存、网络输入输出(input output，IO)带宽，存储IO带宽等资源供主机上运行的实例共享。
实例：运行在主机上的计算节点，常见的实例包括虚拟机(virtual machine，VM)或容器(container)。每个实例占用了主机的部分或全部的虚拟资源。实例的规格(specification)配置包括实例类型(也称为风味(flavor))和实例尺寸(size)。
实例类型指示了实例的资源特点。例如经济型实例采用较为便宜的CPU，占有的计算资源较低，费用较低；计算增强型实例采用高性能的CPU，占有的计算资源充足，费用较高；网络增强型实例占用的网络资源充足，例如配置了高IO带宽，费用相比经济型实例也更高。不同的实例类型代表了租户对于其运行在柔性实例上业务的资源需求。
实例尺寸指示实例占用的资源的量，实例尺寸一般包括了vCPU数量和内存尺寸(单位为吉字节(gigabyte，GB))，还可以包括内存带宽、网络带宽、图像处理器(graphics processing unit，GPU)个数，非易失性存储设备大小(一般为高速存储介质，例如固态驱动器(solid state drive，SSD)、NVMe SSD)等。下文中，示例性的仅展示了实例尺寸包括了vCPU数量和内存尺寸(gigabyte，GB)的场景。柔性实例的尺寸为柔性实例在某一时刻或时间段内实际占用的资源。由于柔性实例的尺寸在柔性实例的生命周期内是浮动的，本申请中提及的柔性实例的尺寸包括了柔性实例在不同时间段内的尺寸。
依据实例在其生命周期内占用的虚拟资源的变化与否，实例分为刚性实例和柔性实例。其中，刚性实例的规格在其生命周期内保持不变。例如，刚性实例在创建时指定的尺寸为16vU，32GB，该刚性实例在其生命周期内尺寸保持为16vU，32GB，直至其生命周期结束，该刚性实例占用的资源被释放。柔性实例的尺寸在其生命周期内可变。例如，柔性实例在创建时指定的尺寸为16vU，32GB(最大尺寸)，该柔性实例在其生命周期内尺寸可以浮动，浮动区间内最大尺寸为16vU，32GB，最小尺寸可以依据预设的规则定义，例如为1vU，2GB。
资源利用率：主机/实例在某一时刻或者某一时间段内对其物理资源/虚拟资源的利用率。主机/实例的资源利用率可以通过CPU利用率、内存利用率等参数计算得出。
服务质量(quality of service，QoS)参数：QoS参数表现了实例的运行状况。常见的服务质量参数包括以下任意一种或多种的组合：每秒包数(packet per second，PPS)，请求响应成功率，每秒输入输出数(input output per second，IOPS)，各类时延(例如，网络时延、业务响应时延)，网络带宽，存储带宽。
图1提供了一种云数据中心，其中包括了资源管理系统和计算资源池，计算资源池包括多个主机。计算资源池向资源管理系统提供资源以供实例的运行。资源管理系统与每个主机建立通信连接，对每个主机的运行参数和各主机上运行的现存实例的运行参数进行监控。资源管理系统还负责实例的调度，包括新实例的发放和现存实例的迁移。调度刚性实例时，资源管理系统主要考虑主机的资源占用情况，如果一主机的未被占用的虚拟资源小于刚性实例的尺寸，那么资源管理系统将无法向该主机上发放/迁移该刚性实例。由于柔性实例的尺寸是浮动的，柔性实例的调度更加灵活，有助于主机的资源利用率的提升。
主机1(32U，64GB)上运行了刚性实例1(16vU，32GB)和刚性实例2(16vU，32GB)。刚性实例1(16vU，32GB)的资源利用率为11％，刚性实例2(16vU，32GB)的资源利用率为19％，那么主机1的资源利用率为大概为15％。对于刚性实例，由于其尺寸固定，分配给该刚性实例的虚拟资源即使未被该刚性实例充分使用，也不可以被分配给其他实例或者用于创建新的实例。主机1上即使约有85％的虚拟资源实际上没有被使用(资源利用率低)，但由于主机上的虚拟资源已经全部被占用(资源占用率高)，导致资源管理系统不会在主机1上分配新的实例或者往主机1上迁移实例。然而，在资源利用率较低的情况下，即使缩减实例的尺寸，实例的QoS参数也可能保持稳定，相对的，如果一个实例的QoS参数达不到预期，可以增大实例的尺寸。
基于此，以下介绍柔性实例的配置参数，柔性实例的配置参数包括规格配置和QoS保障需求。规格配置和QoS保障需求各包括一个或多个配置项。
柔性实例的配置参数可以通过柔性实例的配置界面输入至资源管理系统，也可以通过配置指令(例如，应用程序接口(application programming interface，API))输入至资源管理系统。
云数据中心也可以预设有柔性实例的配置参数，无须租户输入配置参数。
实例类型：柔性实例的类型可以包括通用类型，计算密集类型，内存密集类型，网络密集类型，机器学习计算密集类型，存储密集类型等。
实例尺寸：柔性实例的实例尺寸指示了柔性实例的尺寸区间。示例性的，柔性实例的尺寸区间可以包括最小尺寸和最大尺寸共计两个配置项，或仅包括柔性实例的最大尺寸这一配置项，或包括柔性实例的标准尺寸并设置浮动区间(上浮/下浮比例)共计两个配置项。
图2中展示的柔性实例配置界面中柔性实例的实例尺寸包括最小尺寸和最大尺寸。租户可以在vCPU填框内输入最小的vCPU个数和最大的vCPU个数，还可以在内存填框内输入最小的内存尺寸和最大的内存尺寸。一般的，最小vCPU个数不能小于1，最大vCPU不能小于最小vCPU数，最小内存尺寸不能小于1GB，最大内存尺寸不能小于最小 内存尺寸。柔性实例的尺寸区间仅包括最大尺寸或包括标准尺寸和标准尺寸的浮动区间的情况依次类推。
QoS保障需求：由于柔性实例的尺寸是浮动的，因此，租户需要输入柔性实例的QoS保障需求，QoS保障需求指示了租户对该柔性实例的QoS参数的要求。资源管理系统会根据QoS保障需求进行柔性实例的调度。QoS保障需求作为云数据中心对租户的服务等级承诺，体现了云数据中心对租户的柔性实例在其生命周期内QoS参数的质量保障，例如，劣化保障(允许劣化比例)，恢复保障(时长需求)。对柔性实例的尺寸的调整的过程和结果应该符合QoS保障需求。云数据中心可以预设QoS保障需求，租户也可以按照自身的要求配置QoS保障需求。
QoS保障需求包括QoS保障优先级、允许劣化比例、时长需求、高可用性(high availability，HA)HA保障能力、额外关注QoS参数等配置项中的任意一种或多种：一般的，QoS保障需求至少包括时长需求。
QoS保障优先级：可以分为多个级别，高/中/低，或高/低等。资源管理系统对拥有不同的QoS保障优先级的柔性实例的调度方式有所区别。资源管理系统会优先保证优先级更高的柔性实例的QoS参数劣化比例更低、QoS参数劣化时长更短，在发生超出QoS参数劣化比例和QoS参数劣化时长的情况下，资源管理系统也会优先保证优先级更高的柔性实例的QoS参数劣化比例快速恢复，例如把优先级更高的柔性实例迁移到空闲的主机上，或者把优先级更高的柔性实例所在的主机上优先级较低的实例迁移到其他主机上。
可选的，QoS保障优先级为低的情况下，资源管理系统不对该柔性实例的QoS参数劣化比例以及QoS参数劣化时长进行严格的管理，仅在保障更高优先级的柔性实例的QoS参数劣化情况(劣化比例和/或劣化时长)不超出允许劣化参数的前提下，对低优先级的柔性实例实行调度。
QoS保障需求包括QoS保障优先级的情况下，租户在选择QoS保障优先级后，可以填写允许劣化比例、时长需求。或者，租户在选择部分QoS保障优先级的情况下，才需要填写允许劣化比例、时长需求，在选择其余QoS保障优先级的情况下，无须填写允许劣化比例、时长需求。例如，QoS保障优先级为高/中的情况下，需要填写允许劣化比例、时长需求，QoS保障优先级为低的情况下，无须填写允许劣化比例、时长需求。
QoS保障需求不包括QoS保障优先级，或者租户未选择任何QoS保障优先级的情况下，租户在选择QoS保障优先级后，需要填写允许劣化比例、时长需求。
允许劣化比例：柔性实例的最低QoS参数相对于柔性实例在最大尺寸下的QoS参数的劣化比例。例如，一个柔性实例的最大尺寸为16vU，32GB且允许劣化比例为20％，那么该柔性实例在运行过程中实际运行QoS参数最低为最大尺寸下该柔性实例的QoS参数的80％。
柔性实例在运行过程中的实际运行QoS参数的计算，可以采用平均值或瞬时值。实际运行QoS参数采用平均值的情况下，要求周期内该柔性实例的平均运行QoS参数低于允许劣化比例的时长不得超过时长需求，柔性实例的调度将更加灵活，云数据中心的资源利用率将更高。实际运行QoS参数采用瞬时值的情况下，要求瞬时运行QoS 参数低于允许劣化比例的时长不得超过时长需求*最大尺寸下的QoS参数，柔性实例的性能将更有保障。
时长需求：允许该柔性实例的实际运行QoS参数和该柔性实例的最大尺寸下QoS参数的比例低于允许劣化比例(如果未指定允许劣化比例或QoS保障需求不包括允许劣化比例，这里的取值为100％)的时长。为了保障允许该柔性实例的实际运行QoS参数和该柔性实例的最大尺寸下QoS参数的比例低于允许劣化比例的时长，调整柔性实例的尺寸的过程所使用的时间应该短于或等于设置的时长需求。如前文所述，这里的实际运行QoS参数可以是平均值或瞬时值。如果一个柔性实例的实际运行QoS参数低于该柔性实例的最大尺寸的劣化比例的时长已经超过或即将超过或被预测出可能超过该柔性实例的时长需求，那么该柔性实例可能将会被调度至其他主机，或者该柔性实例所在的主机上运行的其他实例将会被调度至其他主机。
时长需求越低，柔性实例的实际运行QoS参数的稳定性越高，性能更有保障。与之相对的，时长需求越高，租户对柔性实例的实际运行QoS参数下降的容忍度越高，柔性实例的调度将更加灵活。
可选的，允许劣化比例也可以用最低劣化尺寸替代，最低劣化尺寸也即租户能够接受的柔性实例的尺寸最低劣化水平。最低劣化尺寸包括vCPU和内存。如前例，采用最低劣化尺寸的情况下，可以直接定义柔性实例的最低劣化尺寸为12.8vU，25.6GB。
可选的，针对租户选择的不同的QoS保障优先级，柔性实例配置界面会提供推荐的允许劣化比例的缺省值、推荐值和取值区间中的一个或多个。例如，高优先级：缺省值10％，推荐值5％-15％，取值区间3％-30％；中优先级：缺省15％，推荐10％-50％，取值区间5％-50％。
可选的，针对租户选择的不同的QoS保障优先级，柔性实例配置界面会提供推荐的时长需求的缺省值、推荐值和取值区间中的一个或多个。例如，高优先级：缺省值30秒，推荐值10秒-60秒，取值区间5秒-100秒；中优先级：缺省值60秒，推荐值30秒-180秒，取值区间15秒-300秒。图2中展示了QoS保障优先级为高的情况下的缺省允许劣化比例，缺省时长需求，以及推荐值。
HA保障能力：柔性实例的高可用性要求的保障手段。一般而言，为了保证柔性实例的高可用性，在柔性实例的原部署主机故障或性能不足的情况下，需要迁移柔性实例。HA保障能力可以包括热迁移和冷迁移。其中，热迁移指示允许资源管理系统热迁移柔性实例，冷迁移指示允许资源管理系统冷迁移柔性实例。HA保障能力为热迁移的情况下，柔性实例迁移的成本更高，柔性实例迁移过程的造成的柔性实例中止的时间更短。HA保障能力为冷迁移的情况下，柔性实例迁移的成本更低，柔性实例迁移过程的造成的柔性实例中止的时间更长。
额外关注QoS参数:需要额外关注的该柔性实例的质量参数。租户可以选择一项或多项QoS参数作为额外关注QoS参数填入。为了进一步保障柔性实例运行过程中的性能，提升租户体验，租户可以根据自己运行在柔性实例上的业务特点，选择该柔性实例需要额外关注的QoS参数，以便资源管理系统在调度柔性实例的过程中尽可能保证该额外关注QoS参数的稳定性。
图3介绍了资源管理系统包括的各个模块。资源管理系统包括配置模块、计费模 块、调度模块、监控模块，可选的，还包括预测辅助模块。其中，配置模块用于接收柔性实例的配置参数，例如提供图2中的柔性实例配置界面，或者接收配置指令。计费模块用于根据柔性实例的配置参数计算基准费用，以及根据柔性实例的运行参数计算实际费用。调度模块用于基于各类信息对柔性实例进行调度。监控模块用于对主机和现存柔性实例的运行参数进行监控，将各类运行参数提供至调度模块和预测辅助模块。预测辅助模块用于对各主机的未来资源利用率进行预测，以及对柔性实例的未来QoS参数劣化情况进行预测。
柔性实例配置完毕后，计费模块可以为该柔性实例生成基准费用，以便租户预估该柔性实例的使用成本。基准费用指示了预估的该柔性实例的单位时间的基础使用费用。单位时间可以是秒/分/小时/天/周/月/年等。
如图4所示，计费模块参考规模配置和QoS保障需求来计算基准费用。具体的，计费模块可以采用规格配置中的一个或多个配置项，和QoS保障需求中的任意一个或多个配置项来计算基准费用。规模配置一定的情况下，QoS保障需求越高，柔性实例的基准费用越高，而QoS保障需求越低，柔性实例的基准费用越低。
以下介绍各规格配置项和QoS保障需求项对于基准费用的影响，其中，列举到某一配置项时是基于该配置项包括于柔性实例配置参数的假设。
实例类型：不同实例类型的基准费用一般不同；
最小尺寸：一般的，该配置项的值越大，QoS保障需求越高，从而基准费用也越高；
最大尺寸：一般的，该配置项的值越大，QoS保障需求越高，从而基准费用也越高；
QoS保障优先级：一般的，该配置项越高，QoS保障需求越高，从而基准费用也越高；
允许劣化比例：一般的，该配置项的值越低，QoS保障需求越高，从而基准费用越高；
时长需求：一般的，该配置项的值越低，QoS保障需求越高，从而基准费用也越高；
HA保障能力：一般的，该配置项为热迁移的柔性实例的QoS保障需求高于该配置项为冷迁移的柔性实例的QoS保障需求，该配置项为热迁移的柔性实例的基准费用也高于该配置项为冷迁移的柔性实例的基准费用；
额外关注QoS参数：一般的，填写了该配置项的柔性实例的QoS保障需求高于未填写该配置项的柔性实例，填写了该配置项的柔性实例的基准费用也低于未填写该配置项的柔性实例的基准费用。
示例性的，基准费用等于规格配置算出来的费用与QoS保障需求算出来的折扣比例的乘积。对于相同类型的实例，最小尺寸和最大尺寸越大的情况下，规格配置算出来的费用越高。QoS保障需求中各个配置项会影响到具体的折扣比例。
如图4所示，随着柔性实例的实际使用，计费模块可以根据基准费用和柔性实例的实际运行尺寸，生成实际费用。
实际费用可以包括基准费用和至少一段浮动费用，每段浮动费用包括一段时间内 柔性实例的尺寸以及对应的费用。例如，柔性实例在6:00被创建，并在9:00被调整了规格，那么实际费用包括：
基准费用；
6:00-9:00，尺寸1，浮动费用1；
9:00-14:00，尺寸2，浮动费用2；
……。
图5展示了柔性实例的调度流程，包括：
S200，监控模块持续监控各个主机和各现存柔性实例的运行参数，将主机和现存柔性实例的运行参数发送至调度模块以及预测辅助模块(可选的)，将各现存柔性实例的运行参数发送至计费模块。
主机的运行参数包括主机的资源利用率，柔性实例的运行参数包括柔性实例的资源利用率、当前尺寸、QoS参数劣化比例、QoS参数劣化时长中的一个或多个。
每个主机安装有监控模块的代理，代理周期性的采集其所在的主机及其所在的主机上运行的各现存柔性实例的运行参数，并上报给监控模块。
S200可以与调度流程中的其他步骤并列执行。
S201，租户的客户端向配置模块发送新建柔性实例的配置参数。
租户的客户端可以运行在租户本地的服务器上，也可以运行在云数据中心的某一主机上。租户的客户端通过配置指令或柔性实例配置界面将新建柔性实例的配置参数发送至配置模块。
S202，配置模块将新建柔性实例的配置参数发送至计费模块。
S201和S202的执行为可选的，云数据中心可以预设有柔性实例的配置参数，无须租户输入。
S203，计费模块根据新建柔性实例的配置参数计算新建柔性实例的基准费用。
S204，计费模块将新建柔性实例的基准费用提供给客户端。
计费模块可以将新建柔性实例的基准费用发送给配置模块，配置模块在其提供的柔性实例配置界面上展示新建柔性实例的基准费用。
S202至S204为可选步骤。在执行了S202至S204的情况下，一般的，在执行S205前，配置模块需要确认租户在获得新建柔性实例的基准费用后的确认指令，该确认指令用于确认按照该配置参数部署该新建柔性实例。
以下，分别介绍资源管理系统不包括预测辅助模块的调度流程(S205至S207)和包括预测辅助模块情况下的调度流程(S208至S212)。
S205，配置模块将新建柔性实例的配置参数发送至调度模块。
S206，调度模块根据新建柔性实例的配置参数、各个主机和各现存柔性实例的运行参数，选择该新建柔性实例的部署主机。
S207，调度模块指示被选中的部署主机按照新建柔性实例的配置参数部署该新建柔性实例。
以下详细介绍S206中，调度模块如何选择一个部署新建柔性实例的主机。
调度模块选择新建柔性实例的部署主机可以综合考虑以下两个维度：
1.主机部署该新建柔性实例后的资源利用率是否超出资源利用率阈值。
2.主机部署该新建柔性实例后，主机上的现存柔性实例的QoS参数劣化比例是否超出各现存柔性实例的允许劣化比例。
调度模块根据主机的资源利用率和新建柔性实例的配置参数，计算该新建柔性实例部署到每个主机后每个主机的资源利用率。将超出资源利用率阈值(例如，50％)的主机从待选主机列表剔除。如果没有主机能够进入待选主机列表，那么进一步判断如果QoS保障优先级低于新建柔性实例的QoS保障优先级的现存柔性实例被迁移到其他主机或者被中止的话，哪些主机能够留在待选主机列表。如果仍旧无法找出任何待选主机，调度模块可以通知客户端新建柔性实例部署失败。
在待选主机列表中，调度模块根据主机的资源利用率、新建柔性实例的配置参数、现存柔性实例的QoS参数劣化比例，计算新建柔性实例部署到每个主机后，每个主机上运行的现存柔性实例的QoS参数劣化情况。将运行有可能超出允许劣化比例的现存柔性实例的主机进一步从待选主机列表中剔除。随后，在待选主机列表中选择资源利用率最低的主机作为新建柔性实例的部署主机。如果待选主机列表中每个主机在部署新建柔性实例后，都会出现至少一个现存的柔性实例的QoS参数劣化比例会超出其允许劣化比例，那么进一步判断如果QoS保障优先级低于新建柔性实例的QoS保障优先级的现存柔性实例被迁移到其他主机或者被中止的话，哪些主机能够留在待选主机列表。随后，在待选主机列表中选择资源利用率最低的主机作为新建柔性实例的部署主机。
如果任一主机上运行的现存柔性实例需要被迁移或者被中止，那么综合考虑以下维度来选择该主机上哪些柔性实例需要被迁移或者被中止：
1.现存柔性实例的QoS保障等级。QoS保障等级越低的现存柔性实例优先被迁移或者中止。
2.现存柔性实例的类型。为了尽量避免现存柔性实例和新建柔性实例部署在同一主机上互相影响QoS参数的情况，可以考虑优先将类型与新建柔性实例相同的现存柔性实例迁移或者中止。
根据设计，可以指定只有QoS保障等级为低的现存柔性实例会被中止。需要被迁移的柔性实例的迁移方式参考其HA保障能力。为被迁移的现存柔性实例挑选部署主机的过程参考为新建柔性实例挑选部署主机的过程(S206)。
柔性实例需要被迁移或者被中止的场景，可以发生在新建柔性实例的情况，也可以发生在出现某一现存柔性实例的QoS参数劣化比例超出允许劣化比例，且该主机没有剩余资源的情况。在后一情况下，也可以综合考虑主机上现存柔性实例的QoS保障等级，以及主机上现存柔性实例的类型来选择被迁移或者被中止的柔性实例。
S208，配置模块将新建柔性实例的配置参数发送至调度模块和预测辅助模块。
S209，预测辅助模块对各主机、各现存柔性实例、新建柔性实例未来的运行参数进行预测。
预测辅助模块内部设置有基于学习历史数据生成的智能模块，该智能模块根据新建柔性实例的配置参数、监控模块提供的各个主机和各现存柔性实例的运行参数，对各主机、各现存柔性实例、新建柔性实例的未来运行参数进行预测。预测时间可以为未来n个单位时间，n为大于0的整数。
S210，预测辅助模块将各主机的预测运行参数、现存柔性实例的预测运行参数、新建柔性实例的预测运行参数发送至调度模块。
S209和S210中，预测辅助模块对于各主机、各现存柔性实例的未来的运行参数进行预测可以与其他步骤并行执行。
S211，调度模块根据新建柔性实例的配置参数、各主机的预测运行参数、现存柔性实例的预测运行参数、新建柔性实例的预测运行参数，选择该新建柔性实例的部署主机。
S212，调度模块指示被选中的部署主机按照新建柔性实例的配置参数部署该新建柔性实例。
S211中，调度模块如何选择部署新建柔性实例的主机的方式与S206的不同之处在于调度模块可以进一步的利用各类预测运行参数来选择新建柔性实例的部署主机。例如，可以根据主机的预测资源利用率和新建柔性实例的配置参数来计算该新建柔性实例部署到每个主机后每个主机的在未来一段时间内的预测资源利用率，据此确定待选主机列表。在待选主机列表中，调度模块可以根据主机的预测资源利用率、新建柔性实例的配置参数、现存柔性实例的预测运行参数，计算新建柔性实例部署到每个主机后，每个主机上运行的现存柔性实例未来一段时间内的QoS参数劣化情况，据此选择新建柔性实例的部署主机。
S213，调度模块根据调整需求，确定需要调整的柔性实例。调整需求包括QoS保障需求和该柔性实例的运行参数。
S214，调度模块指示计算资源池的主机对需要调整的柔性实例的尺寸进行调整，调整所述柔性实例的尺寸的过程符合所述QoS保障需求。具体的，调整所述柔性实例的尺寸的过程所使用的时间短于或等于时长需求。
调度模块根据监控模块提供的云数据中心中运行的各柔性实例运行参数和QoS保障需求来确定需要调整的柔性实例。需要调整的柔性实例分为情况：
第一类，调度模块确定所述柔性实例的运行参数指示所述柔性实例的QoS参数不符合所述QoS保障需求时，指示所述计算资源池增大所述柔性实例的尺寸。具体的，调度模块判断柔性实例的QoS参数劣化比例已经或即将超出允许劣化比例。这种情况下，调度模块指示需要调整的柔性实例的部署主机增大需要调整的柔性实例的尺寸，也即调度模块指示该部署主机分配更多的资源给该柔性实例。
该指示中可以包括调整参数，该调整参数可以由调度模块根据需要调整的柔性实例的尺寸范围和QoS参数劣化比例以及QoS参数劣化时长确定。部署主机根据调整参数提升需要调整的柔性实例的尺寸。如果部署主机上没有剩余可供分配的资源，则调度模块指示中止或迁移部署主机上的现存柔性实例，再增大该柔性实例的尺寸，或者将该柔性实例迁移到其他有空闲资源的主机上，再由迁移到的主机增大该柔性实例的尺寸。调度模块在需要调整的柔性实例的尺寸范围内调整需要调整的柔性实例的尺寸。一般的，QoS参数劣化比例越大的，QoS参数劣化时长越大的柔性实例的尺寸的提升幅度越大。
调整该柔性实例的尺寸的过程所使用的时间短于或等于所述时长需求。调整该柔性实例的尺寸的过程所使用的时间包括迁移现存柔性实例所需的时间(可能的)、迁 移该柔性实例所需的时间(可能的)、增大该柔性实例的尺寸的时间。
第二类，调度模块确定所述柔性实例的运行参数指示该柔性实例的资源利用率低于阈值时，指示所述计算资源池减小所述柔性实例的尺寸。这种情况下，调度模块指示需要调整的柔性实例的部署主机减少需要调整的柔性实例的尺寸，也即调度模块指示该部署主机减少分配给该柔性实例的资源。该指示中可以包括调整参数，该调整参数可以由调度模块根据需要调整的柔性实例的尺寸范围和资源利用率确定。部署主机根据调整参数降低需要调整的柔性实例的尺寸。调度模块在需要调整的柔性实例的尺寸范围内调整需要调整的柔性实例的尺寸。一般的，资源利用率越低的柔性实例的尺寸的降低幅度越大。
S213和S214可以与调度流程中的其他步骤并列执行。
S215，计费模块根据柔性实例的基准费用和柔性实例的运行参数生成实际费用。
S216，计费模块将柔性实例的实际费用提供给客户端。
随着柔性实例的运行，如果有任一柔性实例被租户主动中止，或者为了保障QoS保障级别更高的柔性实例的QoS参数被调度模块中止，计费模块根据监控模块提供的柔性实例的运行参数生成实际费用，将实际费用提供给客户端。
如图6所示，采用本申请提供的柔性实例的调度方法下，柔性实例1、柔性实例2、柔性实例3、柔性实例4的最大尺寸均与图1中的刚性实例相同。然而由于这4个柔性实例的当前负载并不重，在当前时刻资源管理系统为这4个柔性实例的分配的尺寸都小于设置的最大尺寸，使得主机1的资源利用率大幅提高。随着这4个柔性实例的负载的提升，资源管理系统可以为每个柔性实例从主机1分配更多的资源，或者把部分柔性实例迁移到其他主机上以保障主机1上运行的柔性实例的QoS参数。本申请提供的云数据中心的主机资源利用率更高，降低了云数据中的能源消耗，产生了环保效益。
图7提供了一种云数据中心，包括计算机400和至少一台计算机600。计算机400和计算机600间通过通信网络连接。
计算机400包括处理器401、网络设备402、总线403、存储设备404。处理器401、网络设备402、存储设备404之间通过总线403通信。处理器401可以为中央处理器(central processing unit，CPU)。存储设备403可以包括易失性存储设备(volatile memory)，例如随机存取存储设备(random access memory，RAM)，或非易失性存储设备(non-volatile memory)，例如只读存储设备(read-only memory，ROM)，快闪存储设备，HDD或SSD等。网络设备402为网络接口卡。
存储设备404中存储有可执行指令，处理器401执行该可执行指令以执行资源管理系统的各个模块以运行图5所示的方法。存储设备404还可以包括运行操作系统(operation system，OS)所需的可执行指令。OS可以为LINUX TM,UNIX TM,WINDOWS TM等。
计算机600，也即主机，包括处理器601、网络设备602、总线603、存储设备604。计算机600的组织结构与计算机400相同。存储设备604中存储有可执行指令，处理器601执行该可执行指令以至少一个柔性实例和监控模块的代理。
在上述实施例中，可以全部或部分地通过软件、硬件、固件或者其任意组合来实 现。当使用软件或固件实现时，可以全部或部分地以计算机程序产品的形式实现。该计算机程序产品包括一个或多个计算机指令。在计算机上加载和执行该计算机程序指令时，全部或部分地产生按照本发明实施例的流程或功能。该计算机可以是通用计算机、专用计算机、计算机网络、或者其他可编程装置。该计算机指令可以存储在计算机可读存储介质中，或者从一个计算机可读存储介质向另一个计算机可读存储介质传输，例如，该计算机指令可以从一个网站站点、计算机、服务器或数据中心通过有线(例如同轴电缆、光纤、双绞线)或无线(例如红外、无线、微波等)方式向另一个网站站点、计算机、服务器或数据中心进行传输。该计算机可读存储介质可以是计算机能够存取的任何介质或者是包含一个或多个介质集成的服务器、数据中心等数据存储设备。该介质可以是磁性介质，(例如，软盘、硬盘、磁带)、光介质(例如，光盘)、或者半导体介质(例如SSD)等。

Claims (28)

  1. 一种云数据中心，其特征在于，所述云数据中心包括资源管理系统和计算资源池，所述计算资源池包括至少一个主机；
    所述资源管理系统，用于监控所述计算资源池中运行的柔性实例的运行参数，并根据调整需求指示所述计算资源池调整所述柔性实例的尺寸，所述调整需求包括服务质量QoS保障需求和所述柔性实例的运行参数；
    所述计算资源池，用于运行所述柔性实例，采集所述柔性实例的运行参数，并根据所述指示调整所述柔性实例的尺寸，其中，调整所述柔性实例的尺寸的过程符合所述QoS保障需求。
  2. 如权利要求1所述的云数据中心，其特征在于，所述调整需求还包括规格配置，所述规格配置包括所述柔性实例的尺寸范围，调整后的所述柔性实例的尺寸在所述尺寸范围内。
  3. 如权利要求2所述的云数据中心，其特征在于，所述资源管理系统，用于根据所述规格配置和所述QoS保障需求从所述至少一个主机中选择所述柔性实例的部署主机。
  4. 如权利要求1至3任一所述的云数据中心，其特征在于，
    所述资源管理系统，用于当所述柔性实例的运行参数指示所述柔性实例的资源利用率低于阈值时，指示所述计算资源池减小所述柔性实例的尺寸。
  5. 如权利要求1至4任一所述的云数据中心，其特征在于，所述QoS保障需求包括允许劣化比例，所述资源管理系统，用于当所述柔性实例的运行参数指示所述柔性实例的QoS参数不符合所述允许劣化比例时，指示所述计算资源池增大所述柔性实例的尺寸。
  6. 如权利要求1至5任一所述的云数据中心，其特征在于，所述QoS保障需求包括时长需求，其中，调整所述柔性实例的尺寸的过程所使用的时间短于或等于所述时长需求。
  7. 如权利要求1至6任一所述的云数据中心，其特征在于，所述资源管理系统，还用于根据所述QoS保障需求计算所述柔性实例的基准费用，其中，所述QoS保障需求越高所述基准费用越高。
  8. 如权利要求7所述的云数据中心，其特征在于，所述资源管理系统，还用于根据所述基准费用和所述柔性实例的尺寸计算所述柔性实例的实际费用。
  9. 一种柔性实例的调度方法，其特征在于，所述调度方法执行于云数据中心，所述云数据中心包括资源管理系统和计算资源池，所述调度方法包括：
    所述资源管理系统监控所述计算资源池中运行的柔性实例的运行参数；
    所述资源管理系统根据调整需求指示所述计算资源池调整所述柔性实例的尺寸，所述调整需求包括QoS保障需求和所述柔性实例的运行参数，其中，调整所述柔性实例的尺寸的过程符合所述QoS保障需求。
  10. 如权利要求9所述的调度方法，其特征在于，所述调整需求还包括规格配置，所述规格配置包括所述柔性实例的尺寸范围，调整后的所述柔性实例的尺寸在所述尺寸范围内。
  11. 如权利要求10所述的调度方法，其特征在于，所述调度方法还包括：
    所述资源管理系统根据所述规格配置和所述QoS保障需求从所述计算资源池包括的至少一个主机中选择所述柔性实例的部署主机。
  12. 如权利要求9至11任一所述的调度方法，其特征在于，所述资源管理系统根据调整需求指示所述计算资源池调整所述柔性实例的尺寸，包括:
    当所述柔性实例的运行参数指示所述柔性实例的资源利用率低于阈值时，所述资源管理系统指示所述计算资源池减小所述柔性实例的尺寸。
  13. 如权利要求9至12任一所述的调度方法，其特征在于，所述QoS保障需求包括允许劣化比例，所述资源管理系统根据调整需求指示所述计算资源池调整所述柔性实例的尺寸，包括：
    当所述柔性实例的运行参数指示所述柔性实例的QoS参数不符合所述允许劣化比例时，所述资源管理系统指示所述计算资源池增大所述柔性实例的尺寸。
  14. 如权利要求9至13任一所述的调度方法，其特征在于，所述QoS保障需求包括时长需求；
    所述计算资源池调整所述柔性实例的尺寸的过程所使用的时间短于或等于所述时长需求。
  15. 如权利要求9至14任一所述的调度方法，其特征在于，所述调度方法还包括：
    所述资源管理系统根据所述QoS保障需求计算所述柔性实例的基准费用，其中，所述QoS保障需求越高所述基准费用越高。
  16. 如权利要求15所述的调度方法，其特征在于，所述调度方法还包括：
    所述资源管理系统根据所述基准费用和所述柔性实例的尺寸计算所述柔性实例的实际费用。
  17. 一种资源管理系统，其特征在于，所述资源管理系统包括监控模块和调度模块：
    所述监控模块，用于监控所述计算资源池中运行的柔性实例的运行参数；
    所述调度模块，用于根据调整需求指示计算资源池调整所述柔性实例的尺寸，所述调整需求包括QoS保障需求和所述柔性实例的运行参数，其中，所述计算资源池根据所述指示调整所述柔性实例的尺寸的过程符合所述QoS保障需求。
  18. 如权利要求17所述的资源管理系统，其特征在于，所述调整需求还包括规格配置，所述规格配置包括所述柔性实例的尺寸范围，调整后的所述柔性实例的尺寸在所述尺寸范围内。
  19. 如权利要求18所述的资源管理系统，其特征在于，
    所述调度模块，用于根据所述规格配置和所述QoS保障需求从所述计算资源池包括的至少一个主机中选择所述柔性实例的部署主机。
  20. 如权利要求17至19任一所述的资源管理系统，其特征在于，
    所述调度模块，于当所述柔性实例的运行参数指示所述柔性实例的资源利用率低于阈值时，指示所述计算资源池减小所述柔性实例的尺寸。
  21. 如权利要求17至20任一所述的资源管理系统，其特征在于，所述QoS保障需求包括允许劣化比例；
    所述调度模块，用于当所述柔性实例的运行参数指示所述柔性实例的QoS参数不符合所述允许劣化比例时，指示所述计算资源池增大所述柔性实例的尺寸。
  22. 如权利要求17至21任一所述的资源管理系统，其特征在于，所述QoS保障需求包括时长需求；
    所述计算资源池调整所述柔性实例的尺寸的过程所使用的时间短于或等于所述时长需求。
  23. 如权利要求17至22任一所述的资源管理系统，其特征在于，所述资源管理系统还包括计费模块；
    所述计费模块，用于根据所述QoS保障需求计算所述柔性实例的基准费用，其中，所述QoS保障需求越高所述基准费用越高。
  24. 如权利要求23所述的资源管理系统，其特征在于，
    所述计费模块，用于根据所述基准费用和所述柔性实例的尺寸计算所述柔性实例的实际费用。
  25. 如权利要求17至24任一所述的资源管理系统，其特征在于，所述资源管理系统还包括配置模块；
    所述配置模块，用于提供柔性实例配置界面，通过所述柔性实例配置界面接收所述柔性实例的配置参数，所述柔性实例配置界面包括QoS保障需求配置区域，所述QoS保障需求配置区域用于接收所述柔性实例的租户输入的所述QoS保障需求。
  26. 一种计算机，其特征在于，包括存储器和处理器，所述存储器存储有指令，所述处理器运行所述指令以执行权利要求9至16任一所述的方法。
  27. 一种可读存储介质，其特征在于，所述可读存储介质中存储的指令被计算机执行时，导致所述计算机执行权利要求9至16任一所述的方法。
  28. 一种计算机程序产品，其特征在于，所述计算机程序产品中的指令被计算机执行时，导致所述计算机执行权利要求9至16任一所述的方法。
PCT/CN2021/115896 2020-09-01 2021-09-01 提供柔性实例的云数据中心以及柔性实例的调度方法 WO2022048560A1 (zh)

Priority Applications (2)

Application Number Priority Date Filing Date Title
EP21863624.9A EP4195751A4 (en) 2020-09-01 2021-09-01 CLOUD DATA CENTER FOR PROVIDING A FLEXIBLE INSTANCE AND METHOD FOR PLANNING A FLEXIBLE INSTANCE
US18/175,937 US20230205582A1 (en) 2020-09-01 2023-02-28 Cloud Data Center for Providing Flexible Instance and Scheduling Method for Flexible Instance

Applications Claiming Priority (4)

Application Number Priority Date Filing Date Title
CN202010906615 2020-09-01
CN202010906615.6 2020-09-01
CN202011636326.5 2020-12-31
CN202011636326.5A CN114116114A (zh) 2020-09-01 2020-12-31 提供柔性实例的云数据中心以及柔性实例的调度方法

Related Child Applications (1)

Application Number Title Priority Date Filing Date
US18/175,937 Continuation US20230205582A1 (en) 2020-09-01 2023-02-28 Cloud Data Center for Providing Flexible Instance and Scheduling Method for Flexible Instance

Publications (1)

Publication Number Publication Date
WO2022048560A1 true WO2022048560A1 (zh) 2022-03-10

Family

ID=80360552

Family Applications (1)

Application Number Title Priority Date Filing Date
PCT/CN2021/115896 WO2022048560A1 (zh) 2020-09-01 2021-09-01 提供柔性实例的云数据中心以及柔性实例的调度方法

Country Status (4)

Country Link
US (1) US20230205582A1 (zh)
EP (1) EP4195751A4 (zh)
CN (1) CN114116114A (zh)
WO (1) WO2022048560A1 (zh)

Citations (5)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN103079280A (zh) * 2011-10-26 2013-05-01 中国电信股份有限公司 网络资源调度方法和装置
CN107040479A (zh) * 2016-02-04 2017-08-11 华为软件技术有限公司 一种云计算资源调节的方法和装置
WO2018127107A1 (en) * 2017-01-05 2018-07-12 Huawei Technologies Co., Ltd. System and method for access network configuration
CN110662231A (zh) * 2019-09-25 2020-01-07 南京优速网络科技有限公司 面向5g环境的网络切片资源调整方法及系统
CN111340230A (zh) * 2018-12-18 2020-06-26 北京小桔科技有限公司 一种服务提供方法、装置、服务器及计算机可读存储介质

Patent Citations (5)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN103079280A (zh) * 2011-10-26 2013-05-01 中国电信股份有限公司 网络资源调度方法和装置
CN107040479A (zh) * 2016-02-04 2017-08-11 华为软件技术有限公司 一种云计算资源调节的方法和装置
WO2018127107A1 (en) * 2017-01-05 2018-07-12 Huawei Technologies Co., Ltd. System and method for access network configuration
CN111340230A (zh) * 2018-12-18 2020-06-26 北京小桔科技有限公司 一种服务提供方法、装置、服务器及计算机可读存储介质
CN110662231A (zh) * 2019-09-25 2020-01-07 南京优速网络科技有限公司 面向5g环境的网络切片资源调整方法及系统

Non-Patent Citations (1)

* Cited by examiner, † Cited by third party
Title
See also references of EP4195751A4 *

Also Published As

Publication number Publication date
EP4195751A1 (en) 2023-06-14
EP4195751A4 (en) 2024-02-07
CN114116114A (zh) 2022-03-01
US20230205582A1 (en) 2023-06-29

Similar Documents

Publication Publication Date Title
US9367340B2 (en) Resource management method and apparatus for virtual machine system, and virtual machine system
US11102282B2 (en) Method for managing and allocating resources in a clustered computing environment
US10387202B2 (en) Quality of service implementation in a networked storage system with hierarchical schedulers
US9588789B2 (en) Management apparatus and workload distribution management method
US8161260B2 (en) Optimal memory allocation for guested virtual machine(s)
JP5939740B2 (ja) 動的にリソースを割り当てる方法、システム及びプログラム
US9183016B2 (en) Adaptive task scheduling of Hadoop in a virtualized environment
US9304803B2 (en) Cooperative application workload scheduling for a consolidated virtual environment
EP2863304B1 (en) Allocation method for central processing unit resources and computation node
CN111796908B (zh) 一种资源自动弹性伸缩的系统、方法及云平台
JP5417287B2 (ja) 計算機システム、及び、計算機システムの制御方法
US10333859B2 (en) Multi-tenant resource coordination method
CN106133693B (zh) 虚拟机的迁移方法、装置及设备
WO2015132753A1 (en) Method for memory management in virtual machines, and corresponding system and computer program product
CN110058966A (zh) 用于数据备份的方法、设备和计算机程序产品
US20160196157A1 (en) Information processing system, management device, and method of controlling information processing system
EP3314424A1 (en) Techniques for virtual machine migration
CN112805682A (zh) 具有成本效益的高可用性的多单租户服务
CN112052072B (zh) 一种虚拟机的调度策略及超融合系统
WO2022048560A1 (zh) 提供柔性实例的云数据中心以及柔性实例的调度方法
CN114489922A (zh) 虚拟实例设置方法及装置
WO2016075771A1 (ja) 計算機システムおよび計算機システムにおけるオートスケール方法
WO2013095083A1 (en) A method and system of extending computing grid resources
US20210382760A1 (en) Resource adjustment device, non-transitory computer-readable medium and resource adjustment method
US20240118989A1 (en) Proactively perform placement operations to provide resizing recommendations for worker nodes

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: 21863624

Country of ref document: EP

Kind code of ref document: A1

ENP Entry into the national phase

Ref document number: 2021863624

Country of ref document: EP

Effective date: 20230306

NENP Non-entry into the national phase

Ref country code: DE