WO2016206441A1 - 一种虚拟资源分配方法及装置、计算机存储介质 - Google Patents

一种虚拟资源分配方法及装置、计算机存储介质 Download PDF

Info

Publication number
WO2016206441A1
WO2016206441A1 PCT/CN2016/078610 CN2016078610W WO2016206441A1 WO 2016206441 A1 WO2016206441 A1 WO 2016206441A1 CN 2016078610 W CN2016078610 W CN 2016078610W WO 2016206441 A1 WO2016206441 A1 WO 2016206441A1
Authority
WO
WIPO (PCT)
Prior art keywords
virtual
vnf
virtual resource
resource request
currently
Prior art date
Application number
PCT/CN2016/078610
Other languages
English (en)
French (fr)
Inventor
程骐
Original Assignee
中兴通讯股份有限公司
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 中兴通讯股份有限公司 filed Critical 中兴通讯股份有限公司
Publication of WO2016206441A1 publication Critical patent/WO2016206441A1/zh

Links

Images

Classifications

    • GPHYSICS
    • G06COMPUTING; CALCULATING OR COUNTING
    • G06FELECTRIC DIGITAL DATA PROCESSING
    • G06F9/00Arrangements for program control, e.g. control units
    • G06F9/06Arrangements for program control, e.g. control units using stored programs, i.e. using an internal store of processing equipment to receive or retain programs
    • G06F9/46Multiprogramming arrangements
    • G06F9/50Allocation of resources, e.g. of the central processing unit [CPU]

Definitions

  • the present invention relates to a network function virtualization technology, and in particular, to a virtual resource allocation method and apparatus, and a computer storage medium.
  • NFV Network Function Virtualization
  • VNF virtual network function
  • NFVO Virtual Resource Management Infrastructure
  • This method matches the idle virtual resources of the current Virtual Resource Management Infrastructure (NFVI) one by one according to the number of virtual CPU cores, virtual memory, virtual hard disk, etc. required by the VNF virtual resource application, and the allocation is confirmed when all the elements are satisfied. If a resource is not met, it fails. Since NFVO is to perform virtual resource orchestration and allocation for multiple VNFs, this method cannot guarantee the fairness and overall efficiency of virtual resource usage. For example, when the two VNFs of A and B simultaneously initiate virtual resource requests, the traditional method will queue the requests of A and B in the order of arrival. If A arrives first, then the A request is satisfied first and then the B request is satisfied. If the virtual resource after the request of A is satisfied at this time, the resource application of B fails. For example, some applications do not return to the system in time after using virtual resources, resulting in fewer and fewer idle virtual resources available to the system.
  • NFVI Virtual Resource Management Infrastructure
  • the embodiment of the present invention provides a virtual resource allocation method and device, and a computer storage medium, which avoids artificial allocation of virtual resources through application priority for virtual resource allocation in a multi-VNF competitive environment.
  • the blindness enhances the rationality and efficiency of the system's global virtual resource allocation.
  • the embodiment of the present invention adopts the following technical solutions:
  • a virtual resource allocation method including:
  • the VNF that currently has the virtual resource request is given a weight value
  • the idle virtual resource is allocated to the VNF having the virtual resource request according to the weight value assigned by the VNF.
  • the step of acquiring the usage efficiency value of the VNF currently having the virtual resource request includes:
  • the usage efficiency value of the virtual CPU of each VNF currently having the virtual resource request, the usage efficiency value of the virtual memory, and the usage efficiency value of the virtual hard disk are respectively obtained.
  • the formula is utilized: Calculate the usage efficiency value of the virtual CPU, where the VCPU is the virtual CPU, the sum of the allocated VCPUs is the sum of the VCPUs allocated by the virtual resource system, and the VCPU occupancy rate is the usage rate of the VCPU allocated to each VNF pair, and the resource return time is one time.
  • the VNF is assigned the total time that the VCPU starts to return the VCPU.
  • the formula is utilized: Calculate the usage efficiency value of the virtual memory, where VMEM is virtual memory, the sum of the allocated VMEMs is the sum of the VMEMs allocated by the virtual resource system, the VMEM occupancy rate is the usage rate of the VMEM allocated to each VNF pair, and the resource return time is one time.
  • the VNF is assigned the total time from the start of VMEM to the return of VMEM.
  • the formula is utilized: Calculate the usage efficiency of the virtual hard disk.
  • the VHD is the virtual hard disk.
  • the sum of the allocated VHDs is the sum of the VHDs allocated by the virtual resource system.
  • the VHD occupancy rate is the usage rate of the VHD allocated to each VNF pair.
  • the resource return time is once. The total time from when the VNF is assigned a VHD to when the VHD is returned.
  • the step of acquiring the performance score of the virtual network function VNF currently having the virtual resource request according to the usage efficiency value includes:
  • the step of determining the performance scores of the virtual CPU, the virtual memory, and the virtual hard disk according to the usage efficiency values of the virtual CPU, the virtual memory, and the virtual hard disk includes:
  • Determining a trend of the analysis sequence wherein the trend of the analysis sequence includes: a gradual trend, an up and down trend, and a gradual trend;
  • the trend of the analysis sequence is a gradual trend, the number of the predetermined use efficiency values after the time is selected, and the value with the highest efficiency value is used as the performance score;
  • the trend of the analysis sequence is a gradual trend, the number of the predetermined use efficiency values after the time is selected, and the value with the smallest efficiency value is used as the performance score;
  • the median of the efficiency values used in the analysis sequence is selected as the performance score.
  • the step of assigning a weight value to the VNF currently having the virtual resource request according to the performance score of the VNF includes:
  • the performance scores of the virtual CPU, the virtual memory, and the virtual hard disk of the VNF having the virtual resource request are respectively summed, and the virtual CPU total performance score, the virtual memory total performance score, and the virtual hard disk total performance score are obtained;
  • the ratio is sequentially used as the weight value of the virtual CPU given to each VNF, the weight value of the virtual memory, and the weight value of the virtual hard disk.
  • the method before the step of acquiring the usage efficiency value of the VNF currently having the virtual resource request, the method further includes:
  • the current idle virtual resource is allocated to the VNF having the virtual resource request according to the relationship between the current idle virtual resource and the virtual resource request requirement of the VNF having the virtual resource request.
  • the step of allocating the current idle virtual resource to the VNF having the virtual resource request according to the relationship between the current idle virtual resource and the virtual resource request requirement of the VNF having the virtual resource request includes:
  • the current idle virtual resource is allocated an idle virtual resource according to the virtual resource application requirement of the VNF;
  • the method further includes:
  • the current idle virtual resource is allocated the idle virtual resource according to the virtual resource application requirement of the VNF;
  • the usage efficiency value of the VNF currently having the virtual resource request is obtained.
  • the step of releasing the virtual resource occupied by the idle VNF includes:
  • a virtual resource of a VNF that is less active than a predetermined value is released.
  • the step of acquiring the activity level of each VNF of the currently allocated virtual resource includes:
  • DA represents the average activity of the second predetermined time period of the VNF every day
  • WA represents the average activity of the second preset time period of the VNF every day
  • MA represents the second of the VNF monthly.
  • D is the number of statistical days
  • W is the number of statistical weeks
  • M is the number of statistical months.
  • the average activity of the second preset time period of the VNF is calculated by using the formula: ( ⁇ DAT)/(T ⁇ D), wherein the DAT is VNF every day in the second preset time period.
  • the resource occupation time, ⁇ DAT is the sum of the resource occupation time of the VNF in the second preset time period within the statistical days, and T is the length of the second preset time period.
  • the average active degree of the second preset time period of the VNF on the day of the week is calculated, wherein the WAT is the VNF every day in the day.
  • the average active degree of the second preset time period of the VNF on the day of the month is calculated, wherein MAT is VNF and the day is in the day.
  • the duration of resource occupation in the second preset time period, ⁇ MAT sums the resource occupation time of the VNF in the second preset time period of the month in the statistical month, and T is the second The length of the preset time period.
  • a virtual resource allocation apparatus includes:
  • the first obtaining module is configured to obtain a usage efficiency value of the virtual network function VNF currently having the virtual resource request;
  • a second obtaining module configured to obtain, according to the usage efficiency value, a performance score of a virtual network function VNF currently having a virtual resource request;
  • the assigning module is configured to assign a weight value to the VNF currently having the virtual resource request according to the performance score of the VNF;
  • the allocation module is configured to allocate the idle virtual resource to the VNF having the virtual resource request according to the weight value assigned by the VNF.
  • the first obtaining module is further configured to obtain a usage efficiency value of the virtual CPU, a usage efficiency value of the virtual memory, and a use efficiency value of the virtual hard disk, respectively, of each VNF currently having the virtual resource request.
  • a computer storage medium is provided.
  • the computer storage medium provided by the embodiment of the present invention stores a computer program, where the computer program is used to execute the virtual resource allocation method.
  • the beneficial effect of the present invention is that the virtual resource allocation method provided by the embodiment of the present invention obtains the efficiency score of the VNF by analyzing the previous use efficiency of the VNF, and assigns weights to the virtual resource request of the VNF according to the performance score of the VNF.
  • the virtual resource allocation problem in the environment of VNF competition avoids the artificial allocation of virtual resources through application priority. Blindness improves the rationality and efficiency of the system's global virtual resource allocation.
  • the system can automatically adjust the allocation scheme by this method, which greatly reduces the virtual resource allocation management cost.
  • FIG. 1 is a flowchart of a virtual resource allocation method according to an embodiment of the present invention
  • FIG. 2 is a flowchart of a method for obtaining a performance score according to an embodiment of the present invention
  • FIG. 3 is a schematic diagram showing a trend of an analysis sequence according to an embodiment of the present invention.
  • FIG. 4 is a flowchart of a method for acquiring a weight value according to an embodiment of the present invention
  • FIG. 5 is a second flowchart of a virtual resource allocation method according to an embodiment of the present invention.
  • FIG. 6 is a third flowchart of a virtual resource allocation method according to an embodiment of the present invention.
  • FIG. 7 is a flowchart of a method for releasing a virtual resource according to an embodiment of the present invention.
  • FIG. 8 is a block diagram of a virtual resource allocation apparatus according to an embodiment of the present invention.
  • FIG. 9 is a structural diagram of a virtual resource allocation apparatus in an example provided by an embodiment of the present invention.
  • FIG. 10 is a diagram showing an information model included in an application behavior performance information database according to an embodiment of the present invention.
  • a virtual resource allocation method As shown in FIG. 1, the method includes:
  • the virtual resource allocation method provided by the embodiment of the present invention obtains the efficiency score of the VNF by analyzing the efficiency of the VNF in the past, and obtains the weighting of the VNF virtual resource request according to the performance score of the VNF.
  • the method solves the problem of virtual resource allocation in the environment of multi-VNF competition, avoids the blindness caused by artificial allocation of virtual resources through application priority, and improves the rationality and efficiency of global virtual resource allocation.
  • the system can automatically adjust the allocation scheme by this method, which greatly reduces the virtual resource allocation management cost.
  • the step S101 includes: acquiring, respectively, a usage efficiency value of the virtual CPU of each VNF currently having the virtual resource request, a usage efficiency value of the virtual memory, and a usage efficiency value of the virtual hard disk.
  • step S101 uses a formula: Calculate the usage efficiency value of the virtual CPU.
  • the VCPU is the virtual CPU.
  • the sum of the allocated VCPUs is the sum of the VCPUs allocated by the virtual resource system.
  • the unit is Hz.
  • the VCPU usage is the usage rate of the VCPU allocated to each VNF. %; virtual resource return time is the total time from the VNF being assigned VCPU to returning VCPU, in seconds.
  • the unit of the allocated VCPU sum and the virtual resource return time can also be set according to actual needs.
  • step S101 uses a formula: Calculate the usage efficiency value of the virtual memory, where VMEM is the virtual memory, the sum of the allocated VMEMs is the sum of the VMEMs allocated by the virtual resource system, and the unit is GB; the VMEM occupancy rate is the usage rate of the VMEM allocated to each VNF pair, and the unit is: %; virtual resource return time is the total time from the VNF being allocated VMEM to returning VMEM, in seconds.
  • the unit of the allocated VCPU sum and the virtual resource return time can also be set according to actual needs.
  • step S101 uses a formula: Calculate the usage efficiency of the virtual hard disk.
  • the VHD is the virtual hard disk.
  • the sum of the allocated VHDs is the sum of the VHDs allocated by the virtual resource system.
  • the unit is GB.
  • the VHD occupancy rate is the usage rate of the VHD allocated to each VNF pair. :%; virtual resource return time is the total time from the VNF being assigned VHD to returning VHD, in seconds.
  • the unit of the allocated VCPU sum and the virtual resource return time can also be set according to actual needs.
  • the obtaining step S102 includes: determining the performance scores of the virtual CPU, the virtual memory, and the virtual hard disk according to the usage efficiency values of the virtual CPU, the virtual memory, and the virtual hard disk.
  • the performance scores of the virtual CPU, the virtual memory, and the virtual hard disk can be obtained by analyzing the usage efficiency values.
  • the specific implementation method is as follows:
  • S203 Determine a trend of the analysis sequence, where the trend of the analysis sequence includes: a gradual trend, an up and down volatility trend, and a gradual trend.
  • the median of the use efficiency values in the analysis sequence is selected as the performance score.
  • the first predetermined time period in the foregoing method may be a period of time near the time when the system determines that the VNF is requested to perform the virtual resource, or may be other time, and may be selected according to actual needs.
  • the performance evaluation process of the virtual CPU is taken as an example.
  • the usage efficiency values of multiple virtual CPUs collected over a period of time are sequenced in chronological order and trend analysis is performed. It can be seen from FIG. 3 that the analysis sequence is a gradual trend, and the largest value is selected as the performance score of the virtual CPU in the later use efficiency values, for example, the last three after the selection time. The largest value in the value is used as the performance score of the virtual CPU, that is, point A in FIG.
  • the performance scoring process for virtual memory and virtual hard disk is similar to the virtual CPU performance scoring process, and will not be described here.
  • the virtual CPU, the virtual memory, and the virtual hard disk of all the VNFs having the virtual resource request are respectively obtained.
  • the performance scores are summed, and the weight values of the virtual CPU, virtual memory, and virtual hard disk of each VNF having the virtual resource request are obtained according to the total performance score.
  • the specific implementation process is as follows:
  • the ratio of the scores is sequentially used as the weight value of the virtual CPU given to each VNF, the weight value of the virtual memory, and the weight value of the virtual hard disk.
  • the performance scores of all virtual CPUs of VNFs with virtual resource requests are summed up.
  • the value of the virtual CPU of the VNF with a virtual resource request is 20, and the virtual CPU of the VNF with the virtual resource request has a weight value of 20/2000, that is, 0.01, and the percentage is 1 %.
  • the process of determining the weight value for the virtual memory and the virtual hard disk is similar to the process of determining the weight value of the virtual CPU.
  • the method further includes:
  • the system When there is a virtual resource request by the VNF, the system first determines whether the current idle virtual resource can satisfy the virtual resource application of all the VNFs having the virtual resource request according to the virtual resource application information and the current idle virtual resource status, and according to the satisfaction situation, Determine the allocation scheme for virtual resources. As shown in FIG. 6, the specific implementation method is as follows:
  • the VNF pair resource can be analyzed.
  • the periodicity of occupancy is used to estimate the activity level of the VNF in a certain period of time.
  • the system preferentially releases the virtual resources occupied by some VNFs with low activity.
  • the specific implementation method is as follows:
  • a virtual resource of a VNF whose activity level is lower than a preset value For example, release active Virtual resources of VNF below 20%.
  • the preset value can be set according to actual needs.
  • DA represents the average activity of the second predetermined time period of the VNF every day
  • WA represents the average activity of the second preset time period of the VNF every day
  • MA represents the second of the VNF monthly.
  • D is the number of statistical days
  • W is the number of statistical weeks
  • M is the number of statistical months.
  • the second preset time end may be a period of time near the time when the system determines that the VNF is used for the virtual resource request, and the specific duration may be obtained by analyzing the average duration of the VNF used by the virtual resource, and obtaining the analysis period, that is, the second preset.
  • the time period can also be set according to actual needs.
  • step S701 using the formula: ( ⁇ DAT)/(T ⁇ D), calculating an average activity of the second preset time period of the VNF every day, wherein the DAT is VNF every day within the second preset time period.
  • the duration of the resource occupation, ⁇ DAT is the sum of the resource occupation time of the VNF in the second preset time period within the statistical days, and T is the length of the second preset time period.
  • step S701 using the formula: ( ⁇ WAT)/(T ⁇ W), calculating an average activity of the second preset time period of the VNF on the day of the week, wherein the WAT is VNF every day of the week.
  • the duration of the resource occupation in the second preset time period where WAT is the sum of the resource occupation durations of the VNF in the second preset time period in the statistical week, and T is the second The length of the preset time period.
  • the day of the week means that if the time when the current system detects that there is a VNF for the virtual resource request is Monday, then the day of the week is the weekly Monday, that is, the second preset time period described in the weekly Monday is calculated.
  • the average level of activity of the VNF is calculated.
  • step S701 uses the formula: ( ⁇ MAT)/(T ⁇ M) to calculate the VNF monthly.
  • the average active degree of the second preset time period wherein MAT is the resource occupation time of the VNF monthly in the second preset time period, and ⁇ MAT is the monthly VNF for the VNF.
  • the sum of the resource occupation durations in the second preset time period, and T is the length of the second preset time period.
  • the meaning of the day of the month is that if the time when the current system detects that there is a VNF for the virtual resource request is No. 1, then the monthly number is the first day of the month, that is, the second preset described in the monthly No. 1 is calculated.
  • the average activity of the VNF over time is that if the time when the current system detects that there is a VNF for the virtual resource request is No. 1, then the monthly number is the first day of the month, that is, the second preset described in the monthly No. 1 is calculated.
  • the average activity of the VNF over time is
  • the method further includes:
  • the current idle virtual resource is allocated the idle virtual resource according to the virtual resource application requirement of the VNF.
  • the virtual resource allocation method provided by the embodiment of the present invention allocates the virtual resource to the VNF having the virtual resource request by reasonably analyzing the virtual resource request of the VNF and the virtual resource allocated by the system, thereby providing an overall comparison for the virtual resource control system. Excellent virtual resource allocation method.
  • a virtual resource allocation apparatus as shown in FIG. 8, includes:
  • the first obtaining module 801 is configured to obtain a usage efficiency value of the virtual network function VNF currently having the virtual resource request;
  • the second obtaining module 802 is configured to obtain, according to the usage efficiency value, a performance score of the virtual network function VNF currently having the virtual resource request;
  • the assigning module 803 is configured to assign a weight value to the VNF currently having the virtual resource request according to the performance score of the VNF;
  • the first allocating module 804 is configured to allocate an idle virtual according to the weight value assigned by the VNF Resources to VNFs with virtual resource requests.
  • the first acquiring module is configured to obtain a usage efficiency value of the virtual CPU, a usage efficiency value of the virtual memory, and a usage efficiency value of the virtual hard disk, respectively, of each VNF currently having the virtual resource request.
  • the first obtaining module acquires the usage efficiency value of the virtual CPU of each VNF currently having the virtual resource request, the usage efficiency value of the virtual memory, and the usage efficiency value of the virtual hard disk, and the specific process is:
  • the formula Calculate the usage efficiency value of the virtual CPU, where the VCPU is the virtual CPU, the sum of the allocated VCPUs is the sum of the VCPUs allocated by the virtual resource system, and the VCPU occupancy rate is the usage rate of the VCPU allocated to each VNF pair, and the resource return time is one time.
  • the VNF is assigned the total time that the VCPU starts to return the VCPU.
  • VMEM virtual memory
  • the sum of the allocated VMEMs is the sum of the VMEMs allocated by the virtual resource system
  • the VMEM occupancy rate is the usage rate of the VMEM allocated to each VNF pair
  • the resource return time is one time.
  • the VNF is assigned the total time from the start of VMEM to the return of VMEM.
  • the VHD is the virtual hard disk.
  • the sum of the allocated VHDs is the sum of the VHDs allocated by the virtual resource system.
  • the VHD occupancy rate is the usage rate of the VHD allocated to each VNF pair.
  • the resource return time is once. The total time from when the VNF is assigned a VHD to when the VHD is returned.
  • the second obtaining module 802 is further configured to determine performance metrics of the virtual CPU, the virtual memory, and the virtual hard disk according to the usage efficiency values of the virtual CPU, the virtual memory, and the virtual hard disk.
  • the second acquiring module 802 includes:
  • a first acquisition order configured to acquire each virtual resource request within a first predetermined time period Multiple usage efficiency values of the virtual CPU, virtual memory, and virtual hard disk of the VNF;
  • a generating unit configured to generate, in a chronological order, a plurality of usage efficiency values of the virtual CPU, the virtual memory, and the virtual hard disk of each VNF having the virtual resource request;
  • a determining unit configured to determine a trend of the analysis sequence, wherein the trend of the analysis sequence includes: a gradual trend, an up and down trend, and a gradual trend;
  • the first performance score determining unit is configured to: if the trend of the analysis sequence is a gradual trend, select the number of the predetermined use efficiency values after the time, and use the value with the highest efficiency value as the performance score;
  • the second performance score determining unit is configured to: if the trend trend of the analysis sequence is a gradual trend, select the number of the predetermined use efficiency values after the time, and use the value with the smallest efficiency value as the performance score;
  • the third performance score determining unit is configured to select a median of the use efficiency values in the analysis sequence as the performance score if the trend of the analysis sequence is an up and down fluctuation trend.
  • the giving module includes:
  • the total performance score obtaining unit is configured to respectively sum the performance scores of the virtual CPU, the virtual memory, and the virtual hard disk of the VNF having the virtual resource request, and obtain the virtual CPU total performance score, the virtual memory total performance score, and the virtual hard disk total performance score;
  • the weight value assigning unit is configured to compare the ratio of the performance score of the virtual CPU of each VNF having the virtual resource request to the virtual CPU total performance score, the ratio of the virtual memory performance score to the virtual memory total performance score, and the performance rating of the virtual hard disk.
  • the ratio of the total performance score to the virtual hard disk is sequentially used as the weight value of the virtual CPU given to each VNF, the weight value of the virtual memory, and the weight value of the virtual hard disk.
  • the device further includes:
  • the third obtaining module 805 is configured to acquire the virtual resource application information of the VNF currently having the virtual resource request and the current idle virtual resource status.
  • the second allocation module 806 is configured to allocate the current idle virtual resource to the VNF having the virtual resource request according to the relationship between the current idle virtual resource and the virtual resource request requirement of the VNF having the virtual resource request.
  • the second allocating module 806 includes:
  • the determining unit is configured to determine whether the current idle virtual resource satisfies the virtual resource application requirement of all the VNFs with the virtual resource request.
  • the first allocation unit is configured to allocate the current idle virtual resource to the idle virtual resource according to the virtual resource application requirement of the VNF when the current idle virtual resource satisfies the virtual resource application requirement of the VNF having the virtual resource request.
  • the release unit is configured to release the virtual resource occupied by the idle VNF when the current idle virtual resource cannot meet the virtual resource application requirement of all the VNFs with the virtual resource request.
  • the second allocation unit is configured to allocate the current idle virtual resource according to the virtual resource application requirement of the VNF when the current idle resource is added to the virtual resource occupied by the released idle VNF and the virtual resource request requirement of the VNF with the virtual resource request is met. Free virtual resources.
  • the second obtaining unit is configured to obtain the usage efficiency value of the VNF currently having the virtual resource request when the virtual resource resource occupied by the virtual resource request is not satisfied by the current idle resource and the virtual resource occupied by the released idle VNF. .
  • the releasing unit includes:
  • the acquisition sub-unit is configured to acquire the activity level of each VNF of the currently allocated virtual resource.
  • the subunit is released and configured to release the virtual resource of the VNF whose activity is lower than a preset value.
  • the specific process of obtaining the activity level of each VNF of the currently allocated virtual resource by the obtaining subunit is:
  • DA represents the average activity of the second preset time period of the VNF every day
  • WA table The average activity level of the second preset time period indicated by the VNF on the day of the week
  • MA represents the average activity level of the second preset time period of the VNF on the day of the month
  • D is the number of statistical days
  • W is the number of statistical weeks
  • M is the number of months of statistics.
  • the average activity of the second preset time period of the VNF is calculated by using the formula: ( ⁇ DAT)/(T ⁇ D), wherein the DAT is the resource occupancy of the VNF in the second preset time period every day.
  • the duration is ⁇ DAT is the sum of the resource occupation time of the VNF in the second preset time period in the statistical days, and T is the length of the second preset time period.
  • FIG. 9 is a structural diagram of a virtual resource allocation apparatus according to an embodiment of the present invention.
  • the resource usage performance analysis module is responsible for calculating the performance score and the distribution weight value of the VNF for the virtual resource usage.
  • the module corresponds to the second acquisition module and the assignment module in the above device.
  • the application behavior analysis module is configured to analyze the periodicity of the VNF on the virtual resource occupation, calculate the activity level of the VNF, and provide a virtual resource release recommendation to the virtual resource allocation module.
  • the module corresponds to a release unit in the second of the devices described above.
  • the virtual resource allocation module is configured to receive the VNF virtual resource application forwarded by the NFVO, collect the idle virtual resource condition of the NFVI, and combine the analysis result of the performance analysis module and the behavior analysis module. If so, propose a resource allocation plan to NFVO.
  • the module corresponds to the first distribution module, the second distribution, and the third acquisition module in the above device.
  • the resource usage collection module is responsible for collecting the actual virtual resource usage information of the VNF, including the VCPU load, the VMEM occupancy rate, and the VHD occupancy rate.
  • the module runs independently in the VNF virtual machine, periodically reports usage information, and stores it in the performance information database.
  • the module corresponds to the first acquisition module of the above device.
  • the application behavior performance information database stores the virtual resource requirement information of the VNF and the actual use of the virtual resources.
  • the information model it contains is shown in Figure 10.
  • the left part of the virtual machine is the virtual resource requirement information of the VNF
  • the right part of the virtual machine is the actual use of the virtual resource of the VNF.
  • VCPU_E is the performance of the virtual CPU, that is, the use efficiency of the virtual CPU
  • VMEM_E is the use efficiency of the virtual memory, that is, the use efficiency of the virtual memory
  • VHD_E is the use efficiency of the virtual hard disk, that is, the use efficiency of the virtual hard disk; rate.
  • the virtual resource allocation device is a device corresponding to the virtual resource allocation method. All the implementation manners in the foregoing method embodiments are applicable to the device embodiment, and the same technical effects can be achieved.
  • each unit in the virtual resource allocation device may be implemented by a central processing unit (CPU) or a microprocessor (Micro Processor Unit, MPU) located in the virtual resource allocation device. ), or a digital signal processor (DSP), or a Field Programmable Gate Array (FPGA).
  • CPU central processing unit
  • MPU Micro Processor Unit
  • DSP digital signal processor
  • FPGA Field Programmable Gate Array
  • the apparatus for tracking the service signaling may also be stored in a computer readable storage medium if it is implemented in the form of a software function module and sold or used as a separate product.
  • the technical solution of the embodiments of the present invention may be embodied in the form of a software product in essence or in the form of a software product stored in a storage medium, including a plurality of instructions.
  • Make a computer device can be A personal computer, server, or network device, etc.) performs all or part of the methods described in various embodiments of the present invention.
  • the foregoing storage medium includes various media that can store program codes, such as a USB flash drive, a mobile hard disk, a read only memory (ROM), a magnetic disk, or an optical disk.
  • program codes such as a USB flash drive, a mobile hard disk, a read only memory (ROM), a magnetic disk, or an optical disk.
  • an embodiment of the present invention further provides a computer storage medium, wherein a computer program is stored, and the computer program is used to execute a virtual resource allocation method according to an embodiment of the present invention.
  • the technical solution of the embodiment of the present invention obtains the efficiency score of the VNF by analyzing the previous use efficiency of the VNF, and assigns the weight of the virtual resource request of the VNF according to the performance score of the VNF, and allocates the virtual resource in the environment of multi-VNF competition. It avoids the artificial blindness caused by the application of priority virtual resource allocation, and improves the rationality and efficiency of the system global virtual resource allocation. At the same time, the system can automatically adjust the allocation scheme by this method, which greatly reduces the virtual resource allocation management cost.

Landscapes

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

Abstract

一种虚拟资源分配方法及装置、计算机存储介质,所述虚拟资源分配方法包括:获取当前有虚拟资源请求的虚拟网络功能VNF的使用效率值(S101);根据所述使用效率值,获取当前有虚拟资源请求的虚拟网络功能VNF的效能评分(S102);根据VNF的效能评分,对当前有虚拟资源请求的VNF赋予权重值(S103);根据VNF被赋予的权重值,分配空闲虚拟资源至有虚拟资源请求的VNF(S104)。所述虚拟资源分配方法通过分析VNF以往对虚拟资源的使用效率,获得VNF的效能评分,根据VNF的效能评分对VNF的虚拟资源请求进行权重分配,提升了系统全局资源分配的合理性与效率。

Description

一种虚拟资源分配方法及装置、计算机存储介质 技术领域
本发明涉及网络功能虚拟化技术,尤其涉及一种虚拟资源分配方法及装置、计算机存储介质。
背景技术
越来越多的运营商希望业务运行在价格相对低廉的IT服务器上,而不是传统的电信设备上,这样系统整体更加节能环保,同时业务系统的部署能够透明化,运维自动化,监控、容灾等更易实现。为了满足上述需求,欧洲电信标准化协会ETSI组织提出了网络功能虚拟化(NFV,Network Function Virtualization)的概念。其中虚拟网络功能编排器(NFVO,Network Function Virtualization Orchestrator)负责虚拟网络功能(VNF,Virtual Network Function)的规划、部署与整网虚拟资源的调度。
目前大多数NFVO提供的虚拟资源分配方法就是基于虚拟资源请求的简单多要素匹配。此方法根据VNF虚拟资源申请需要的虚拟CPU核数、虚拟内存、虚拟硬盘等,与当前虚拟资源管理池(NFVI,Network Function Virtualization Infrastructure)的空闲虚拟资源逐个匹配,所有要素满足时才确认分配。如果某一项资源不满足,则失败。由于NFVO要针对多VNF进行虚拟资源编排与分配,该方法不能保证虚拟资源使用的公平性和整体效率。例如,当A、B两VNF同时发起虚拟资源请求,传统做法会将A、B的请求按到达顺序排队,若A先到达,则先满足A请求再满足B请求。如果此时满足A的请求后空余虚拟资源较少,则B的资源申请失败。再比如某些应用在使用虚拟资源后,不向系统及时归还,造成系统可用的空闲虚拟资源越来越少。
发明内容
为解决上述技术问题,本发明实施例提供了一种虚拟资源分配方法及装置、计算机存储介质,对于多VNF竞争的环境下虚拟资源分配问题,避免人为的通过应用优先级进行虚拟资源分配带来的盲目性,提升了系统全局虚拟资源分配的合理性与效率。
为了解决上述技术问题,本发明实施例采用如下技术方案:
依据本发明实施例的一个方面提供了一种虚拟资源分配方法,包括:
获取当前有虚拟资源请求的虚拟网络功能VNF的使用效率值;
根据所述使用效率值,获取当前有虚拟资源请求的虚拟网络功能VNF的效能评分;
根据VNF的效能评分,对当前有虚拟资源请求的VNF赋予权重值;
根据VNF被赋予的权重值,分配空闲虚拟资源至有虚拟资源请求的VNF。
本发明实施例中,所述获取当前有虚拟资源请求的VNF的使用效率值的步骤包括:
分别获取每个当前有虚拟资源请求的VNF的虚拟CPU的使用效率值、虚拟内存的使用效率值和虚拟硬盘的使用效率值。
本发明实施例中,利用公式:
Figure PCTCN2016078610-appb-000001
计算虚拟CPU的使用效率值,其中VCPU为虚拟CPU,分配的VCPU总和为虚拟资源系统已分配的VCPU总和,VCPU占用率为每个VNF对被分配的VCPU的使用率,资源交还时间为一次从VNF被分配VCPU开始到归还VCPU的总时间。
本发明实施例中,利用公式:
Figure PCTCN2016078610-appb-000002
计算虚拟内存的使用效率值,其中VMEM为虚拟内存,分配的VMEM总和为虚 拟资源系统已分配的VMEM总和,VMEM占用率为每个VNF对被分配的VMEM的使用率,资源交还时间为一次从VNF被分配VMEM开始到归还VMEM的总时间。
本发明实施例中,利用公式:
Figure PCTCN2016078610-appb-000003
计算虚拟硬盘的使用效率值,其中,VHD为虚拟硬盘,分配的VHD总和为虚拟资源系统已分配的VHD总和,VHD占用率为每个VNF对被分配的VHD的使用率,资源交还时间为一次从VNF被分配VHD开始到归还VHD的总时间。
本发明实施例中,所述根据所述使用效率值,获取当前有虚拟资源请求的虚拟网络功能VNF的效能评分的步骤包括:
根据虚拟CPU、虚拟内存和虚拟硬盘的使用效率值,确定虚拟CPU、虚拟内存和虚拟硬盘的效能评分。
本发明实施例中,所述根据虚拟CPU、虚拟内存和虚拟硬盘的使用效率值,确定虚拟CPU、虚拟内存和虚拟硬盘的效能评分的步骤包括:
获取在第一预定时间段内每个有虚拟资源请求的VNF的虚拟CPU、虚拟内存和虚拟硬盘的多个使用效率值;
将每个有虚拟资源请求的VNF的虚拟CPU、虚拟内存和虚拟硬盘的多个使用效率值按时间先后顺序生成分析序列;
确定所述分析序列的走向趋势,其中所述分析序列的走向趋势包括:渐好趋势、上下波动趋势和渐差趋势;
若所述分析序列的走向趋势为渐好趋势,则选取时间靠后的预定使用效率值的个数中,使用效率值最大的数值作为效能评分;
若所述分析序列的走向趋势为渐差趋势,则选取时间靠后的预定使用效率值的个数中,使用效率值最小的数值作为效能评分;
若所述分析序列的趋势为上下波动趋势,则选取分析序列中使用效率值的的中位数作为效能评分。
本发明实施例中,所述根据VNF的效能评分,对当前有虚拟资源请求的VNF赋予权重值的步骤包括:
分别对有虚拟资源请求的VNF的虚拟CPU、虚拟内存和虚拟硬盘的效能评分求和,得到虚拟CPU总效能评分、虚拟内存总效能评分和虚拟硬盘总效能评分;
将每个有虚拟资源请求的VNF的虚拟CPU的效能评分与虚拟CPU总效能评分的比值、虚拟内存的效能评分与虚拟内存总效能评分的比值和虚拟硬盘的效能评分与虚拟硬盘总效能评分的比值,依次作为赋予每个VNF的虚拟CPU的权重值、虚拟内存的权重值和虚拟硬盘的权重值。
本发明实施例中,在所述获取当前有虚拟资源请求的VNF的使用效率值的步骤之前,所述方法还包括:
获取当前有虚拟资源请求的VNF的虚拟资源申请信息以及当前空闲虚拟资源状况;
根据当前空闲虚拟资源与有虚拟资源请求的VNF的虚拟资源申请需求之间的关系,将当前空闲虚拟资源分配给有虚拟资源请求的VNF。
本发明实施例中,所述根据当前空闲虚拟资源与有虚拟资源请求的VNF的虚拟资源申请需求之间的关系,将当前空闲虚拟资源分配给有虚拟资源请求的VNF的步骤包括:
判断当前空闲虚拟资源是否满足所有有虚拟资源请求的VNF的虚拟资源申请需求;
若满足,则将当前空闲虚拟资源按VNF的虚拟资源申请需求分配空闲虚拟资源;
若不满足,则释放空闲VNF占用的虚拟资源。
本发明实施例中,所述释放当前空闲VNF占用的虚拟资源的步骤之后,还包括:
若当前空闲资源加上释放的空闲VNF占用的虚拟资源后,满足有虚拟资源请求的VNF的虚拟资源申请需求,则将当前空闲虚拟资源按VNF的虚拟资源申请需求分配空闲虚拟资源;
若当前空闲资源加上释放的空闲VNF占用的虚拟资源后,无法满足有虚拟资源请求的VNF的虚拟资源申请需求,则获取当前有虚拟资源请求的VNF的使用效率值。
本发明实施例中,所述释放空闲VNF占用的虚拟资源的步骤包括:
获取当前已分配虚拟资源的每个VNF的活跃程度;
释放活跃程度低于一预设值的VNF的虚拟资源。
本发明实施例中,所述获取当前已分配虚拟资源的每个VNF的活跃程度的步骤包括:
利用公式:(DA×D+WA×W+MA×M)/(D+W+M),计算当前已分配虚拟资源VNF在第二预设时间段内的活跃程度;
其中,DA表示VNF每天所述第二预设时间段的平均活跃度,WA表示VNF每周该天所述第二预设时间段的平均活跃度,MA表示VNF每月该天所述第二预设时间段的平均活跃度,D为统计天数,W为统计周数,M为统计月数。
本发明实施例中,利用公式:(∑DAT)/(T×D),计算VNF每天所述第二预设时间段的平均活跃度,其中,DAT为VNF每天在第二预设时间段内的资源占用时长,∑DAT为在统计天数内对VNF每天在第二预设时间段内的资源占用时长的求和,T为所述第二预设时间段的长度。
本发明实施例中,利用公式:(∑WAT)/(T×W),计算VNF每周该天所述第二预设时间段的平均活跃度,其中,WAT为VNF每周该天在所述第二预设时间段内的资源占用时长,∑WAT为在统计周数内对VNF每周该天在所述第二预设时间段内的资源占用时长的求和,T为所述第二预设时间段 的长度。
本发明实施例中,利用公式:(∑MAT)/(T×M),计算VNF每月该天所述第二预设时间段的平均活跃度,其中,MAT为VNF每月该天在所述第二预设时间段内的资源占用时长,∑MAT在统计月数内对VNF每月该天在所述第二预设时间段内的资源占用时长的求和,T为所述第二预设时间段的长度。
依据本发明实施例的另一个方面提供了一种虚拟资源分配装置,包括:
第一获取模块,配置为获取当前有虚拟资源请求的虚拟网络功能VNF的使用效率值;
第二获取模块,配置为根据所述使用效率值,获取当前有虚拟资源请求的虚拟网络功能VNF的效能评分;
赋予模块,配置为根据VNF的效能评分,对当前有虚拟资源请求的VNF赋予权重值;
分配模块,配置为根据VNF被赋予的权重值,分配空闲虚拟资源至有虚拟资源请求的VNF。
本发明实施例中,所述第一获取模块还配置为分别获取每个当前有虚拟资源请求的VNF的虚拟CPU的使用效率值、虚拟内存的使用效率值和虚拟硬盘的使用效率值。
依据本发明实施例的另一个方面提供了一种计算机存储介质,本发明实施例提供的计算机存储介质存储有计算机程序,该计算机程序用于执行上述虚拟资源分配方法。
本发明的有益效果是:本发明实施例提供的虚拟资源分配方法通过分析VNF以往对资源的使用效率,获得VNF的效能评分,根据VNF的效能评分对VNF的虚拟资源请求进行权重分配,对于多VNF竞争的环境下虚拟资源分配问题,避免了人为的通过应用优先级进行虚拟资源分配带来的 盲目性,提升了系统全局虚拟资源分配的合理性与效率。同时,通过该方法系统可自动调整分配方案,大大降低了虚拟资源分配管理成本。
附图说明
图1表示本发明实施例提供的虚拟资源分配方法的流程图之一;
图2表示本发明实施例提供的效能评分获取方法的流程图;
图3表示本发明实施例提供的分析序列走向趋势示意图;
图4表示本发明实施例提供的权重值获取方法的流程图;
图5表示本发明实施例提供的虚拟资源分配方法的流程图之二;
图6表示本发明实施例提供的虚拟资源分配方法的流程图之三;
图7表示本发明实施例提供的虚拟资源释放方法的流程图;
图8表示本发明实施例提供的虚拟资源分配装置框图;
图9表示本发明实施例提供的示例中虚拟资源分配装置构架图;
图10表示本发明实施例提供的应用行为效能信息数据库包含的信息模型图。
具体实施方式
下面将参照附图更详细地描述本发明的示例性实施例。虽然附图中显示了本发明的示例性实施例,然而应当理解,可以以各种形式实现本发明而不应被这里阐述的实施例所限制。相反,提供这些实施例是为了能够更透彻地理解本发明,并且能够将本发明的范围完整的传达给本领域的技术人员。
依据本发明实施例的一个方面,提供了一种虚拟资源分配方法,如图1所示,所述方法包括:
S101、获取当前有虚拟资源请求的虚拟网络功能VNF的使用效率值。
S102、根据所述使用效率值,获取当前有虚拟资源请求的虚拟网络功 能VNF的效能评分。
S103、根据VNF的效能评分,对当前有虚拟资源请求的VNF赋予权重值。
S104、根据VNF被赋予的权重值,分配空闲虚拟资源至有虚拟资源请求的VNF。
本发明实施例提供的虚拟资源分配方法通过分析VNF以往对虚拟资源的使用效率,获得VNF的效能评分,根据VNF的效能评分对VNF的虚拟资源请求进行权重分配。该方法对于多VNF竞争的环境下虚拟资源分配问题,避免了人为的通过应用优先级进行虚拟资源分配带来的盲目性,提升了系统全局虚拟资源分配的合理性与效率。同时,通过该方法系统可自动调整分配方案,大大降低了虚拟资源分配管理成本。
本发明实施例中,步骤S101包括:分别获取每个当前有虚拟资源请求的VNF的虚拟CPU的使用效率值、虚拟内存的使用效率值和虚拟硬盘的使用效率值。
具体的,步骤S101利用公式:
Figure PCTCN2016078610-appb-000004
计算虚拟CPU的使用效率值,其中VCPU为虚拟CPU,分配的VCPU总和为虚拟资源系统已分配的VCPU总和,单位:Hz;VCPU占用率为每个VNF对被分配的VCPU的使用率,单位:%;虚拟资源交还时间为一次从VNF被分配VCPU开始到归还VCPU的总时间,单位:秒。其中,分配的VCPU总和的单位和虚拟资源交还时间也可根据实际需要进行设定。
具体的,步骤S101利用公式:
Figure PCTCN2016078610-appb-000005
计算虚拟内存的使用效率值,其中VMEM为虚拟内存,分配的VMEM总和为虚拟资源系统已分配的VMEM总和,单位:GB;VMEM占用率为每个VNF对被分配的VMEM的使用率,单位:%;虚拟资源交还时间为一次从VNF 被分配VMEM开始到归还VMEM的总时间,单位:秒。其中,分配的VCPU总和的单位和虚拟资源交还时间也可根据实际需要进行设定。
具体的,步骤S101利用公式:
Figure PCTCN2016078610-appb-000006
计算虚拟硬盘的使用效率值,其中,VHD为虚拟硬盘,分配的VHD总和为虚拟资源系统已分配的VHD总和,单位:GB;VHD占用率为每个VNF对被分配的VHD的使用率,单位:%;虚拟资源交还时间为一次从VNF被分配VHD开始到归还VHD的总时间,单位:秒。其中,分配的VCPU总和的单位和虚拟资源交还时间也可根据实际需要进行设定。
本发明实施例中,在获得步骤S102包括:根据虚拟CPU、虚拟内存和虚拟硬盘的使用效率值,确定虚拟CPU、虚拟内存和虚拟硬盘的效能评分。
在获得虚拟CPU、虚拟内存和虚拟硬盘的使用效率值后,可通过分析使用效率值,得到虚拟CPU、虚拟内存和虚拟硬盘的效能评分。如图2所示,其具体实现方法为:
S201、获取在第一预定时间段内每个有虚拟资源请求的VNF的虚拟CPU、虚拟内存和虚拟硬盘的多个使用效率值。
S202、将每个有虚拟资源请求的VNF的虚拟CPU、虚拟内存和虚拟硬盘的多个使用效率值按时间先后顺序生成分析序列。
S203、确定所述分析序列的走向趋势,其中所述分析序列的走向趋势包括:渐好趋势、上下波动趋势和渐差趋势。
S204、若所述分析序列的走向趋势为渐好趋势,则选取时间靠后的预定使用效率值的个数中,使用效率值最大的数值作为效能评分。
S205、若所述分析序列的走向趋势为渐差趋势,则选取时间靠后的预定使用效率值的个数中,使用效率值最小的数值作为效能评分。
S206、若所述分析序列的趋势为上下波动趋势,则选取分析序列中使用效率值的的中位数作为效能评分。
需要说明的是,上述方法中的第一预定时间段可以为系统判断有VNF进行虚拟资源请求时刻附近的一段时间,也可为其它时间,可根据实际需要进行选取。
为进一步理解本发明实施例提供的上述方法,下面举例说明:
如图3所示,以虚拟CPU的效能评分确定过程为例。将在一段时间内采集的多个虚拟CPU的使用效率值按照时间先后顺序组成分析序列,并做趋势分析。其中,由图3可看出,该分析序列为渐好趋势,则在时间靠后的几个使用效率值中选取最大的数值作为该虚拟CPU的效能评分,例如选取时间靠后的最后三个数值中最大的数值作为该虚拟CPU的效能评分,即图3中的A点。对于虚拟内存和虚拟硬盘的效能评分确定过程同虚拟CPU的效能评分确定过程类似,这里不再进行赘述。
本发明实施例中,在获得每个有虚拟资源请求的VNF的虚拟CPU、虚拟内存和虚拟硬盘的效能评分后,再分别对所有有虚拟资源请求的VNF的虚拟CPU、虚拟内存和虚拟硬盘的效能评分求和,根据总的效能评分得到每个有虚拟资源请求的VNF的虚拟CPU、虚拟内存和虚拟硬盘的权重值。如图4所示,其具体实现过程为:
S401、分别对有虚拟资源请求的VNF的虚拟CPU、虚拟内存和虚拟硬盘的效能评分求和,得到虚拟CPU总效能评分、虚拟内存总效能评分和虚拟硬盘总效能评分。
S402、将每个有虚拟资源请求的VNF的虚拟CPU的效能评分与虚拟CPU总效能评分的比值、虚拟内存的效能评分与虚拟内存总效能评分的比值和虚拟硬盘的效能评分与虚拟硬盘总效能评分的比值,依次作为赋予每个VNF的虚拟CPU的权重值、虚拟内存的权重值和虚拟硬盘的权重值。
为进一步理解本发明实施例提供的上述方法,下面举例说明:
假设将所有有虚拟资源请求的VNF的虚拟CPU的效能评分求和后得 到的数值为2000,而其中一个有虚拟资源请求的VNF的虚拟CPU的效能评分为20,则该有虚拟资源请求的VNF的虚拟CPU的权重值为20/2000,即0.01,化成百分数为1%。对于虚拟内存和虚拟硬盘的权重值确定过程同虚拟CPU的权重值的确定过程类似。
本发明实施例中,如图5所示,在步骤S101之前,所述方法还包括:
S501、获取当前有虚拟资源请求的VNF的虚拟资源申请信息以及当前空闲虚拟资源状况。
S502、根据当前空闲虚拟资源与有虚拟资源请求的VNF的虚拟资源申请需求之间的关系,将当前空闲虚拟资源分配给有虚拟资源请求的VNF。
当有VNF有虚拟资源请求时,系统首先会根据虚拟资源申请信息和当前空闲虚拟资源状况,判断当前空闲虚拟资源是否能满足当前所有有虚拟资源请求的VNF的虚拟资源申请,并根据满足情况,确定虚拟资源的分配方案。如图6所示,其具体实现方法为:
S601、判断当前空闲虚拟资源是否满足所有有虚拟资源请求的VNF的虚拟资源申请需求。
S602、若满足,则将当前空闲虚拟资源按VNF的虚拟资源申请需求分配空闲虚拟资源。
S603、若不满足,则释放空闲VNF占用的虚拟资源。
其中,若当前空闲虚拟资源无法满足所有有虚拟资源请求的VNF的虚拟资源申请需求时,会选择性的释放一些VNF占用的虚拟资源,对于如何选择虚拟资源的释放对象,可通过分析VNF对资源占用的周期性规律,估算VNF的未来某时间段的活跃程度,系统优先选择释放一些活跃程度低的VNF占用的虚拟资源,如图7所示,其具体实现方法为:
S701、获取当前已分配虚拟资源的每个VNF的活跃程度。
S702、释放活跃程度低于一预设值的VNF的虚拟资源。例如释放活跃 程度低于20%的VNF的虚拟资源。所述预设值可根据实际需要设定。
其中,获取当前已分配虚拟资源的每个VNF的活跃程度的具体实现方式为:
利用公式:(DA×D+WA×W+MA×M)/(D+W+M),计算当前已分配虚拟资源VNF在第二预设时间段内的活跃程度;
其中,DA表示VNF每天所述第二预设时间段的平均活跃度,WA表示VNF每周该天所述第二预设时间段的平均活跃度,MA表示VNF每月该天所述第二预设时间段的平均活跃度,D为统计天数,W为统计周数,M为统计月数。
需要说明的是,第二预设时间端可以为系统判断有VNF进行虚拟资源请求时刻附近的一段时间,具体时长可以通过分析VNF以往虚拟资源使用的平均时长,获得分析时段,即第二预设时间段,也可根据实际需要进行设定。
具体的,步骤S701利用公式:(∑DAT)/(T×D),计算VNF每天所述第二预设时间段的平均活跃度,其中,DAT为VNF每天在第二预设时间段内的资源占用时长,∑DAT为在统计天数内对VNF每天在第二预设时间段内的资源占用时长的求和,T为所述第二预设时间段的长度。
具体的,步骤S701利用公式:(∑WAT)/(T×W),计算VNF每周该天所述第二预设时间段的平均活跃度,其中,WAT为VNF每周该天在所述第二预设时间段内的资源占用时长,∑WAT为在统计周数内对VNF每周该天在所述第二预设时间段内的资源占用时长的求和,T为所述第二预设时间段的长度。其中,每周该天的意思为若当前系统检测到有VNF进行虚拟资源请求的时间为周一,则每周该天即每周周一,也就是计算每周周一所述第二预设时间段的VNF的平均活跃程度。
具体的,步骤S701利用公式:(∑MAT)/(T×M),计算VNF每月该 天所述第二预设时间段的平均活跃度,其中,MAT为VNF每月该天在所述第二预设时间段内的资源占用时长,∑MAT在统计月数内对VNF每月该天在所述第二预设时间段内的资源占用时长的求和,T为所述第二预设时间段的长度。其中,每月该天的意思为若当前系统检测到有VNF进行虚拟资源请求的时间为1号,则每月该天即每月1号,也就是计算每月1号所述第二预设时间段的VNF的平均活跃程度。
本发明实施例中,如图6所示,在步骤S603之后,所述方法还包括:
S604、若当前空闲资源加上释放的空闲VNF占用的虚拟资源后,满足有虚拟资源请求的VNF的虚拟资源申请需求,则将当前空闲虚拟资源按VNF的虚拟资源申请需求分配空闲虚拟资源。
S605、若当前空闲资源加上释放的空闲VNF占用的虚拟资源后,无法满足有虚拟资源请求的VNF的虚拟资源申请需求,则获取当前有虚拟资源请求的VNF的使用效率值,即进入步骤S101。
本发明实施例提供的虚拟资源分配方法,通过合理分析VNF的虚拟资源请求和系统已分配出去的虚拟资源,将虚拟资源分配至有虚拟资源请求的VNF,为虚拟资源控制系统提供一种整体较优的虚拟资源分配方法。
依据本发明实施例的另一个方面提供了一种虚拟资源分配装置,如图8所示,包括:
第一获取模块801,配置为获取当前有虚拟资源请求的虚拟网络功能VNF的使用效率值;
第二获取模块802,配置为根据所述使用效率值,获取当前有虚拟资源请求的虚拟网络功能VNF的效能评分;
赋予模块803,配置为根据VNF的效能评分,对当前有虚拟资源请求的VNF赋予权重值;
第一分配模块804,配置为根据VNF被赋予的权重值,分配空闲虚拟 资源至有虚拟资源请求的VNF。
本发明实施例中,所述第一获取模块具体配置为分别获取每个当前有虚拟资源请求的VNF的虚拟CPU的使用效率值、虚拟内存的使用效率值和虚拟硬盘的使用效率值。
其中,所述第一获取模块获取每个当前有虚拟资源请求的VNF的虚拟CPU的使用效率值、虚拟内存的使用效率值和虚拟硬盘的使用效率值,具体过程为:
利用公式:
Figure PCTCN2016078610-appb-000007
计算虚拟CPU的使用效率值,其中VCPU为虚拟CPU,分配的VCPU总和为虚拟资源系统已分配的VCPU总和,VCPU占用率为每个VNF对被分配的VCPU的使用率,资源交还时间为一次从VNF被分配VCPU开始到归还VCPU的总时间。
利用公式:
Figure PCTCN2016078610-appb-000008
计算虚拟内存的使用效率值,其中VMEM为虚拟内存,分配的VMEM总和为虚拟资源系统已分配的VMEM总和,VMEM占用率为每个VNF对被分配的VMEM的使用率,资源交还时间为一次从VNF被分配VMEM开始到归还VMEM的总时间。
利用公式:
Figure PCTCN2016078610-appb-000009
计算虚拟硬盘的使用效率值,其中,VHD为虚拟硬盘,分配的VHD总和为虚拟资源系统已分配的VHD总和,VHD占用率为每个VNF对被分配的VHD的使用率,资源交还时间为一次从VNF被分配VHD开始到归还VHD的总时间。
本发明实施例中,所述第二获取模块802,还配置为根据虚拟CPU、虚拟内存和虚拟硬盘的使用效率值,确定虚拟CPU、虚拟内存和虚拟硬盘的效能评分。
其中,所述第二获取模802,包括:
第一获取单,配置为获取在第一预定时间段内每个有虚拟资源请求的 VNF的虚拟CPU、虚拟内存和虚拟硬盘的多个使用效率值;
生成单元,配置为将每个有虚拟资源请求的VNF的虚拟CPU、虚拟内存和虚拟硬盘的多个使用效率值按时间先后顺序生成分析序列;
确定单元,配置为确定所述分析序列的走向趋势,其中所述分析序列的走向趋势包括:渐好趋势、上下波动趋势和渐差趋势;
第一效能评分确定单元,配置为若所述分析序列的走向趋势为渐好趋势,则选取时间靠后的预定使用效率值的个数中,使用效率值最大的数值作为效能评分;
第二效能评分确定单元,配置为若所述分析序列的走向趋势为渐差趋势,则选取时间靠后的预定使用效率值的个数中,使用效率值最小的数值作为效能评分;
第三效能评分确定单元,配置为若所述分析序列的趋势为上下波动趋势,则选取分析序列中使用效率值的的中位数作为效能评分。
本发明实施例中,所述赋予模块包括:
总效能评分获取单元,配置为分别对有虚拟资源请求的VNF的虚拟CPU、虚拟内存和虚拟硬盘的效能评分求和,得到虚拟CPU总效能评分、虚拟内存总效能评分和虚拟硬盘总效能评分;
权重值赋予单元,配置为将每个有虚拟资源请求的VNF的虚拟CPU的效能评分与虚拟CPU总效能评分的比值、虚拟内存的效能评分与虚拟内存总效能评分的比值和虚拟硬盘的效能评分与虚拟硬盘总效能评分的比值,依次作为赋予每个VNF的虚拟CPU的权重值、虚拟内存的权重值和虚拟硬盘的权重值。
本发明实施例中,如图8所示,所述装置还包括:
第三获取模块805,配置为获取当前有虚拟资源请求的VNF的虚拟资源申请信息以及当前空闲虚拟资源状况;
第二分配模块806,配置为根据当前空闲虚拟资源与有虚拟资源请求的VNF的虚拟资源申请需求之间的关系,将当前空闲虚拟资源分配给有虚拟资源请求的VNF。
本发明实施例中,所述第二分配模块806,包括:
判断单元,配置为判断当前空闲虚拟资源是否满足所有有虚拟资源请求的VNF的虚拟资源申请需求。
第一分配单元,配置为当当前空闲虚拟资源满足所有有虚拟资源请求的VNF的虚拟资源申请需求时,将当前空闲虚拟资源按VNF的虚拟资源申请需求分配空闲虚拟资源。
释放单元,配置为当当前空闲虚拟资源无法满足所有有虚拟资源请求的VNF的虚拟资源申请需求时,释放空闲VNF占用的虚拟资源。
第二分配单元,配置为当当前空闲资源加上释放的空闲VNF占用的虚拟资源后,满足有虚拟资源请求的VNF的虚拟资源申请需求时,将当前空闲虚拟资源按VNF的虚拟资源申请需求分配空闲虚拟资源。
第二获取单元,配置为当当前空闲资源加上释放的空闲VNF占用的虚拟资源后,无法满足有虚拟资源请求的VNF的虚拟资源申请需求时,获取当前有虚拟资源请求的VNF的使用效率值。
本发明实施例中,所述释放单元包括:
获取子单元,配置为获取当前已分配虚拟资源的每个VNF的活跃程度。
释放子单元,配置为释放活跃程度低于一预设值的VNF的虚拟资源。
本发明实施例中,所述获取子单元获取当前已分配虚拟资源的每个VNF的活跃程度具体过程为:
利用公式:(DA×D+WA×W+MA×M)/(D+W+M),计算当前已分配虚拟资源VNF在第二预设时间段内的活跃程度;
其中,DA表示VNF每天所述第二预设时间段的平均活跃度,WA表 示VNF每周该天所述第二预设时间段的平均活跃度,MA表示VNF每月该天所述第二预设时间段的平均活跃度,D为统计天数,W为统计周数,M为统计月数。
具体的,利用公式:(∑DAT)/(T×D),计算VNF每天所述第二预设时间段的平均活跃度,其中,DAT为VNF每天在第二预设时间段内的资源占用时长,∑DAT为在统计天数内对VNF每天在第二预设时间段内的资源占用时长的求和,T为所述第二预设时间段的长度。
具体的,利用公式:(∑WAT)/(T×W),计算VNF每周该天所述第二预设时间段的平均活跃度,其中,WAT为VNF每周该天在所述第二预设时间段内的资源占用时长,∑WAT为在统计周数内对VNF每周该天在所述第二预设时间段内的资源占用时长的求和,T为所述第二预设时间段的长度。
具体的,利用公式:(∑MAT)/(T×M),计算VNF每月该天所述第二预设时间段的平均活跃度,其中,MAT为VNF每月该天在所述第二预设时间段内的资源占用时长,∑MAT在统计月数内对VNF每月该天在所述第二预设时间段内的资源占用时长的求和,T为所述第二预设时间段的长度。
为了进一步理解本发明实施例提供的虚拟资源分配装置,下面举例说明:
如图9所示,为本发明实施例提供的一种虚拟资源分配装置构架图。
其中,资源使用效能分析模块负责计算VNF对虚拟资源使用情况的效能评分和分配权重值。该模块对应上述装置中的第二获取模块和赋予模块。
应用行为分析模块配置为分析VNF对虚拟资源占用的周期性规律,计算VNF的活跃程度,向虚拟资源分配模块提供虚拟资源释放建议。该模块对应上述装置中的第二分配模块中的释放单元。
虚拟资源分配模块配置为接收NFVO转发的VNF虚拟资源申请,采集NFVI的空闲虚拟资源情况,结合效能分析模块与行为分析模块的分析结 果,向NFVO提出资源分配方案。该模块对应上述装置中的第一分配模块、第二分配和第三获取模块。
资源使用情况采集模块负责收集VNF实际的虚拟资源使用信息,包括VCPU负载、VMEM占用率、VHD占用率等。该模块独立运行在VNF虚拟机中,定期上报使用信息,存入效能信息数据库。该模块对应上述装置的第一获取模块。
应用行为效能信息数据库存储VNF的虚拟资源需求信息与虚拟资源实际使用的情况。其包含的信息模型如图10所示。图中虚拟机左侧部分为VNF的虚拟资源需求信息,虚拟机右侧部分为VNF的虚拟资源实际使用的情况。其中,VCPU_E为虚拟CPU的使用效能,即虚拟CPU的使用效率;VMEM_E为虚拟内存的使用效能,即虚拟内存的使用效率;VHD_E为虚拟硬盘的使用效能,即虚拟硬盘的使用效率;usage为占用率。
需要说明的是,该虚拟资源分配装置是与上述虚拟资源分配方法对应的装置,上述方法实施例中所有实现方式均适用于该装置的实施例中,也能达到相同的技术效果。
在实际应用中,所述虚拟资源分配装置中的各个单元所实现的功能,均可由位于虚拟资源分配装置中的中央处理器(Central Processing Unit,CPU)、或微处理器(Micro Processor Unit,MPU)、或数字信号处理器(Digital Signal Processor,DSP)、或现场可编程门阵列(Field Programmable Gate Array,FPGA)等实现。
本发明实施例上述业务信令跟踪的装置如果以软件功能模块的形式实现并作为独立的产品销售或使用时,也可以存储在一个计算机可读取存储介质中。基于这样的理解,本发明实施例的技术方案本质上或者说对现有技术做出贡献的部分可以以软件产品的形式体现出来,该计算机软件产品存储在一个存储介质中,包括若干指令用以使得一台计算机设备(可以是 个人计算机、服务器、或者网络设备等)执行本发明各个实施例所述方法的全部或部分。而前述的存储介质包括:U盘、移动硬盘、只读存储器(ROM,Read Only Memory)、磁碟或者光盘等各种可以存储程序代码的介质。这样,本发明实施例不限制于任何特定的硬件和软件结合。
相应地,本发明实施例还提供一种计算机存储介质,其中存储有计算机程序,该计算机程序用于执行本发明实施例的虚拟资源分配方法。
以上所述的是本发明的优选实施方式,应当指出对于本技术领域的普通人员来说,在不脱离本发明所述的原理前提下还可以作出若干改进和润饰,这些改进和润饰也在本发明的保护范围内。
工业实用性
本发明实施例的技术方案,通过分析VNF以往对资源的使用效率,获得VNF的效能评分,根据VNF的效能评分对VNF的虚拟资源请求进行权重分配,对于多VNF竞争的环境下虚拟资源分配问题,避免了人为的通过应用优先级进行虚拟资源分配带来的盲目性,提升了系统全局虚拟资源分配的合理性与效率。同时,通过该方法系统可自动调整分配方案,大大降低了虚拟资源分配管理成本。

Claims (19)

  1. 一种虚拟资源分配方法,包括:
    获取当前有虚拟资源请求的虚拟网络功能VNF的使用效率值;
    根据所述使用效率值,获取当前有虚拟资源请求的虚拟网络功能VNF的效能评分;
    根据VNF的效能评分,对当前有虚拟资源请求的VNF赋予权重值;
    根据VNF被赋予的权重值,分配空闲虚拟资源至有虚拟资源请求的VNF。
  2. 如权利要求1所述的虚拟资源分配方法,其中,所述获取当前有虚拟资源请求的VNF的使用效率值的步骤包括:
    分别获取每个当前有虚拟资源请求的VNF的虚拟CPU的使用效率值、虚拟内存的使用效率值和虚拟硬盘的使用效率值。
  3. 如权利要求2所述的虚拟资源分配方法,其中,
    利用公式:
    Figure PCTCN2016078610-appb-100001
    计算虚拟CPU的使用效率值,其中VCPU为虚拟CPU,分配的VCPU总和为虚拟资源系统已分配的VCPU总和,VCPU占用率为每个VNF对被分配的VCPU的使用率,资源交还时间为一次从VNF被分配VCPU开始到归还VCPU的总时间。
  4. 如权利要求2所述的虚拟资源分配方法,其中,
    利用公式:
    Figure PCTCN2016078610-appb-100002
    计算虚拟内存的使用效率值,其中VMEM为虚拟内存,分配的VMEM总和为虚拟资源系统已分配的VMEM总和,VMEM占用率为每个VNF对被分配的VMEM的使用率,资源交还时间为一次从VNF被分配VMEM开始到归还VMEM的总时间。
  5. 如权利要求2所述的虚拟资源分配方法,其中,
    利用公式:
    Figure PCTCN2016078610-appb-100003
    计算虚拟硬盘的使用效率值,其中,VHD为虚拟硬盘,分配的VHD总和为虚拟资源系统已分配的VHD总和,VHD占用率为每个VNF对被分配的VHD的使用率,资源交还时间为一次从VNF被分配VHD开始到归还VHD的总时间。
  6. 如权利要求2所述的虚拟资源分配方法,其中,所述根据所述使用效率值,获取当前有虚拟资源请求的虚拟网络功能VNF的效能评分的步骤包括:
    根据虚拟CPU、虚拟内存和虚拟硬盘的使用效率值,确定虚拟CPU、虚拟内存和虚拟硬盘的效能评分。
  7. 如权利要求6所述的虚拟资源分配方法,其中,所述根据虚拟CPU、虚拟内存和虚拟硬盘的使用效率值,确定虚拟CPU、虚拟内存和虚拟硬盘的效能评分的步骤包括:
    获取在第一预定时间段内每个有虚拟资源请求的VNF的虚拟CPU、虚拟内存和虚拟硬盘的多个使用效率值;
    将每个有虚拟资源请求的VNF的虚拟CPU、虚拟内存和虚拟硬盘的多个使用效率值按时间先后顺序生成分析序列;
    确定所述分析序列的走向趋势,其中所述分析序列的走向趋势包括:渐好趋势、上下波动趋势和渐差趋势;
    若所述分析序列的走向趋势为渐好趋势,则选取时间靠后的预定使用效率值的个数中,使用效率值最大的数值作为效能评分;
    若所述分析序列的走向趋势为渐差趋势,则选取时间靠后的预定使用效率值的个数中,使用效率值最小的数值作为效能评分;
    若所述分析序列的趋势为上下波动趋势,则选取分析序列中使用效率值的的中位数作为效能评分。
  8. 如权利要求1所述的虚拟资源分配方法,其中,所述根据VNF的效 能评分,对当前有虚拟资源请求的VNF赋予权重值的步骤包括:
    分别对有虚拟资源请求的VNF的虚拟CPU、虚拟内存和虚拟硬盘的效能评分求和,得到虚拟CPU总效能评分、虚拟内存总效能评分和虚拟硬盘总效能评分;
    将每个有虚拟资源请求的VNF的虚拟CPU的效能评分与虚拟CPU总效能评分的比值、虚拟内存的效能评分与虚拟内存总效能评分的比值和虚拟硬盘的效能评分与虚拟硬盘总效能评分的比值,依次作为赋予每个VNF的虚拟CPU的权重值、虚拟内存的权重值和虚拟硬盘的权重值。
  9. 如权利要求1所述的虚拟资源分配方法,其中,在所述获取当前有虚拟资源请求的VNF的使用效率值的步骤之前,所述方法还包括:
    获取当前有虚拟资源请求的VNF的虚拟资源申请信息以及当前空闲虚拟资源状况;
    根据当前空闲虚拟资源与有虚拟资源请求的VNF的虚拟资源申请需求之间的关系,将当前空闲虚拟资源分配给有虚拟资源请求的VNF。
  10. 如权利要求9所述的虚拟资源分配方法,其中,所述根据当前空闲虚拟资源与有虚拟资源请求的VNF的虚拟资源申请需求之间的关系,将当前空闲虚拟资源分配给有虚拟资源请求的VNF的步骤包括:
    判断当前空闲虚拟资源是否满足所有有虚拟资源请求的VNF的虚拟资源申请需求;
    若满足,则将当前空闲虚拟资源按VNF的虚拟资源申请需求分配空闲虚拟资源;
    若不满足,则释放空闲VNF占用的虚拟资源。
  11. 如权利要求10所述的虚拟资源分配方法,其中,所述释放当前空闲VNF占用的虚拟资源的步骤之后,还包括:
    若当前空闲资源加上释放的空闲VNF占用的虚拟资源后,满足有虚拟 资源请求的VNF的虚拟资源申请需求,则将当前空闲虚拟资源按VNF的虚拟资源申请需求分配空闲虚拟资源;
    若当前空闲资源加上释放的空闲VNF占用的虚拟资源后,无法满足有虚拟资源请求的VNF的虚拟资源申请需求,则获取当前有虚拟资源请求的VNF的使用效率值。
  12. 如权利要求10所述的虚拟资源分配方法,其中,所述释放空闲VNF占用的虚拟资源的步骤包括:
    获取当前已分配虚拟资源的每个VNF的活跃程度;
    释放活跃程度低于一预设值的VNF的虚拟资源。
  13. 如权利要求12所述的虚拟资源分配方法,其中,所述获取当前已分配虚拟资源的每个VNF的活跃程度的步骤包括:
    利用公式:(DA×D+WA×W+MA×M)/(D+W+M),计算当前已分配虚拟资源VNF在第二预设时间段内的活跃程度;
    其中,DA表示VNF每天所述第二预设时间段的平均活跃度,WA表示VNF每周该天所述第二预设时间段的平均活跃度,MA表示VNF每月该天所述第二预设时间段的平均活跃度,D为统计天数,W为统计周数,M为统计月数。
  14. 如权利要求13所述的虚拟资源分配方法,其中,
    利用公式:(∑DAT)/(T×D),计算VNF每天所述第二预设时间段的平均活跃度,其中,DAT为VNF每天在第二预设时间段内的资源占用时长,∑DAT为在统计天数内对VNF每天在第二预设时间段内的资源占用时长的求和,T为所述第二预设时间段的长度。
  15. 如权利要求13所述的虚拟资源分配方法,其中,
    利用公式:(∑WAT)/(T×W),计算VNF每周该天所述第二预设时间段的平均活跃度,其中,WAT为VNF每周该天在所述第二预设时间段内的 资源占用时长,∑WAT为在统计周数内对VNF每周该天在所述第二预设时间段内的资源占用时长的求和,T为所述第二预设时间段的长度。
  16. 如权利要求13所述的虚拟资源分配方法,其中,
    利用公式:(∑MAT)/(T×M),计算VNF每月该天所述第二预设时间段的平均活跃度,其中,MAT为VNF每月该天在所述第二预设时间段内的资源占用时长,∑MAT在统计月数内对VNF每月该天在所述第二预设时间段内的资源占用时长的求和,T为所述第二预设时间段的长度。
  17. 一种虚拟资源分配装置,包括:
    第一获取模块,配置为获取当前有虚拟资源请求的虚拟网络功能VNF的使用效率值;
    第二获取模块,配置为根据所述使用效率值,获取当前有虚拟资源请求的虚拟网络功能VNF的效能评分;
    赋予模块,配置为根据VNF的效能评分,对当前有虚拟资源请求的VNF赋予权重值;
    分配模块,配置为根据VNF被赋予的权重值,分配空闲虚拟资源至有虚拟资源请求的VNF。
  18. 如权利要求17所述的虚拟资源分配装置,其中,所述第一获取模块还配置为分别获取每个当前有虚拟资源请求的VNF的虚拟CPU的使用效率值、虚拟内存的使用效率值和虚拟硬盘的使用效率值。
  19. 一种计算机存储介质,所述计算机存储介质中存储有计算机可执行指令,该计算机可执行指令配置为执行权利要求1-16任一项所述的虚拟资源分配方法。
PCT/CN2016/078610 2015-06-26 2016-04-06 一种虚拟资源分配方法及装置、计算机存储介质 WO2016206441A1 (zh)

Applications Claiming Priority (2)

Application Number Priority Date Filing Date Title
CN201510362175.1 2015-06-26
CN201510362175.1A CN106325997B (zh) 2015-06-26 2015-06-26 一种虚拟资源分配方法及装置

Publications (1)

Publication Number Publication Date
WO2016206441A1 true WO2016206441A1 (zh) 2016-12-29

Family

ID=57586579

Family Applications (1)

Application Number Title Priority Date Filing Date
PCT/CN2016/078610 WO2016206441A1 (zh) 2015-06-26 2016-04-06 一种虚拟资源分配方法及装置、计算机存储介质

Country Status (2)

Country Link
CN (1) CN106325997B (zh)
WO (1) WO2016206441A1 (zh)

Families Citing this family (3)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN107147517A (zh) 2017-03-24 2017-09-08 上海交通大学 一种针对虚拟网络功能的自适应计算资源分配方法
CN110865871A (zh) * 2019-11-14 2020-03-06 北京京航计算通讯研究所 基于资源合理化应用的虚拟化集群资源调度方法
CN113032101B (zh) * 2021-03-31 2023-12-29 深信服科技股份有限公司 虚拟机的资源分配方法、服务器及计算机可读存储介质

Citations (6)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN102103518A (zh) * 2011-02-23 2011-06-22 运软网络科技(上海)有限公司 一种在虚拟化环境中管理资源的系统及其实现方法
CN102223419A (zh) * 2011-07-05 2011-10-19 北京邮电大学 面向网络化操作系统的虚拟资源动态反馈均衡分配机制
CN103118124A (zh) * 2013-02-22 2013-05-22 桂林电子科技大学 一种基于分层多代理的云计算负载均衡方法
CN103309719A (zh) * 2012-03-07 2013-09-18 上海聚耘信息科技有限公司 一种应用于云计算的虚拟机管理系统
CN104065745A (zh) * 2014-07-07 2014-09-24 电子科技大学 云计算动态资源调度系统和方法
WO2014168225A2 (en) * 2013-04-10 2014-10-16 Nec Corporation Communication system

Family Cites Families (4)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
US8667399B1 (en) * 2010-12-29 2014-03-04 Amazon Technologies, Inc. Cost tracking for virtual control planes
US8954978B1 (en) * 2010-12-29 2015-02-10 Amazon Technologies, Inc. Reputation-based mediation of virtual control planes
CN103067476B (zh) * 2012-12-25 2015-08-19 浙江大学 一种基于虚拟机的动态网络重构方法
CN103713956B (zh) * 2014-01-06 2017-01-25 山东大学 应用于云计算虚拟化管理环境中的智能加权负载均衡方法

Patent Citations (6)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN102103518A (zh) * 2011-02-23 2011-06-22 运软网络科技(上海)有限公司 一种在虚拟化环境中管理资源的系统及其实现方法
CN102223419A (zh) * 2011-07-05 2011-10-19 北京邮电大学 面向网络化操作系统的虚拟资源动态反馈均衡分配机制
CN103309719A (zh) * 2012-03-07 2013-09-18 上海聚耘信息科技有限公司 一种应用于云计算的虚拟机管理系统
CN103118124A (zh) * 2013-02-22 2013-05-22 桂林电子科技大学 一种基于分层多代理的云计算负载均衡方法
WO2014168225A2 (en) * 2013-04-10 2014-10-16 Nec Corporation Communication system
CN104065745A (zh) * 2014-07-07 2014-09-24 电子科技大学 云计算动态资源调度系统和方法

Also Published As

Publication number Publication date
CN106325997A (zh) 2017-01-11
CN106325997B (zh) 2020-10-02

Similar Documents

Publication Publication Date Title
CN109783224B (zh) 基于负载调配的任务分配方法、装置及终端设备
CN107295090B (zh) 一种资源调度的方法和装置
US20180198855A1 (en) Method and apparatus for scheduling calculation tasks among clusters
CN106817499B (zh) 一种话务资源调度方法及预测调度装置
Niu et al. Cost-effective cloud HPC resource provisioning by building semi-elastic virtual clusters
EP3320433B1 (en) Method and device for computing resource scheduling
US20180247265A1 (en) Task grouping method and apparatus, electronic device, and computer storage medium
WO2016119412A1 (zh) 一种云平台上的资源伸缩方法和一种云平台
CN108776934A (zh) 分布式数据计算方法、装置、计算机设备及可读存储介质
US11496413B2 (en) Allocating cloud computing resources in a cloud computing environment based on user predictability
CN108984301A (zh) 自适应云资源调配方法和装置
US20140040895A1 (en) Electronic device and method for allocating resources for virtual machines
CN106161485B (zh) 一种基础服务集群的资源调度方法、装置和系统
CN111258745B (zh) 一种任务处理方法及设备
Bi et al. SLA-based optimisation of virtualised resource for multi-tier web applications in cloud data centres
Hacker et al. Flexible resource allocation for reliable virtual cluster computing systems
CN112988390A (zh) 一种算力资源分配方法及装置
CN105022668B (zh) 一种作业调度方法及系统
WO2019062000A1 (zh) 业务员呼入分配方法、电子装置及计算机可读存储介质
WO2017107456A1 (zh) 确定任务消耗资源的方法及装置
CN108241535B (zh) 资源管理的方法、装置及服务器设备
WO2016206441A1 (zh) 一种虚拟资源分配方法及装置、计算机存储介质
Delavar et al. A synthetic heuristic algorithm for independent task scheduling in cloud systems
CN108536525B (zh) 一种宿主机调度方法及装置
CN113886086A (zh) 云平台计算资源分配方法、系统、终端及存储介质

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: 16813557

Country of ref document: EP

Kind code of ref document: A1

NENP Non-entry into the national phase

Ref country code: DE

122 Ep: pct application non-entry in european phase

Ref document number: 16813557

Country of ref document: EP

Kind code of ref document: A1