JP2017151993A - Virtual machine starting method and device - Google Patents

Abstract

PROBLEM TO BE SOLVED: To properly select a virtual machine to be started, secure a normal operation of a service, and sufficiently use a VNFC operating on a virtual machine.SOLUTION: A starting method for a virtual machine comprises: a step for determining N types of virtual network function component VNFCs according to a first service; a step for acquiring a distribution relationship between the N types of VNFCs and the virtual machine, the distribution relationship indicating an amount of each type of VNFC distributed in each virtual machine; and a step for selecting at least one virtual machine which should be started from virtual machines which are not started according to the distribution relationship so that a total amount of each of types of VNFCs included in the started virtual machine and the at least one virtual machine which should be started satisfies a corresponding preset amount.SELECTED DRAWING: Figure 4

Description

  Embodiments of the present invention relate to the communication field, and more specifically, to a virtual machine startup method and apparatus.

  The deployment and application of NFV plays an increasingly important role for operators, data centers, cloud computing service providers, and large network companies. In NFV (Network Function Virtualization), general-purpose hardware and virtualization technologies such as x86 processors are used to reduce device costs in the network with software processing to implement many functions. In NFV, network device functions are no longer dependent on special purpose hardware, and resources can be shared sufficiently and flexibly by software / hardware separation and function abstraction. In this way, rapid development and deployment of new services is carried out, and automatic deployment, automatic scaling, based on actual service (network referred to in this field is network services) requirements. Fault isolation, self-healing, etc. are implemented.

  In accordance with NFV technology principles, a network service may include at least one VNF (VNF, Virtualized Network Function Component), and each VNF may be decomposed into a group of VNFCs (VNFC, Virtualized Network Function Component). That is, the NFV network includes one or more VNFs, and the VNF includes one or more VNFCs. A VNFC is a group of processes or software modules of software packages that provide specific network service functions. Correspondingly, from a physical point of view, the VNFC includes an interface board, software package, switch board, service processing board, main control board, and the like. Correspondingly, from a software perspective, VNFC includes an interface process, a switch process, a service processing process, a main control process, and the like. The same type of VNFC can serve the same function, and each VNFC is ultimately mapped to a virtual machine (VM). In general, the same type of VNFC may be launched on multiple virtual machines (to provide stronger service processing capacity or higher service reliability), and multiple different virtual network function components may be run on the same virtual machine. Sometimes it moves. When the virtual machine is started, the virtual network function components running on the virtual machine are started and started simultaneously.

  In general, one VNF may have several types of VNFC, and each type of VNFC implements different functions. In a complete NFV network, the same type of X0 VNF exists, and correspondingly, each type of X0 VNFC also exists in this type of VNF. There may be multiple VNFCs of the same type, and only the startup time of the VNFC or the startup time of the virtual machine on which the VNFC is running is different. It should be noted that VNFC is randomly distributed within the virtual machine.

  Referring to FIG. 1, FIG. 1 is a schematic distribution diagram of VNFC in a virtual machine. VFN0 and VNF1 are used as examples. VNFCs included in VNF0 are VNFC000, VNFC001, and VNFC002, and VNFCs included in VNF1 are VNFC010 and VNFC011.

  In FIG. 1, VM0 and VM4 indicate that all VNFCs of the same VNF can run on the same virtual machine, and VNFs running on different virtual machines may be different.

  In FIG. 1, VM1 indicates that only some VNFCs of one VNF can run on the virtual machine.

  In FIG. 1, VM2 and VM5 indicate that some VNFCs of one VNF can run on a virtual machine and that there can be one or more VNFCs of each type.

  In FIG. 1, VM3 shows that VNFCs with different VNFs can run on one virtual machine.

  At least one specific VNFC set needs to function normally in order for the NFV network to be able to service normally, and the VNFC set generally includes VNFCs of different VNFs. The VNFC sets for different types of services or services with different priorities may be different.

  A large number of virtual machines may run on a large NFV network. These virtual machines share physical resources such as storage, network, and host resources. During startup, these virtual machines need to read data such as OS (Operating System) and VNFC processes from the shared storage and may generate a large amount of storage I / O operations. If too many virtual machines are started at the same time, a virtual machine boot storm can result, overloading resources such as shared storage I / O, and each virtual machine can run for a very long time. May need to be started over, or may become abnormal, for example reset or even hang up. Thus, to avoid overloading resources such as shared storage I / O when the virtual machine is started, the virtual machine management system on the NFV network may limit the amount of virtual machines that are started simultaneously.

  When the virtual machine management system limits the amount of virtual machines that are started at the same time, a corresponding unreliability problem is caused. In the prior art, the virtual machine management system starts virtual machines according to commands of the host or VNF (virtual machines deployed on the host are started one by one or VNF virtual machines are started one by one) . Because the amount of virtual machines that start at the same time is limited and unplanned, not all virtual machines can be started quickly to start all VNFCs required by a particular network service. Since not all VNFCs required by a specific network service are activated, the network service cannot be normally provided during the startup period, and the efficiency of providing the network service is low.

  Embodiments of the present invention provide a virtual machine starting method and a virtual machine starting device in order to quickly provide a service during service failure recovery and during initial system construction of a virtual machine and to quickly improve service processing capability.

  According to a first aspect, an embodiment of the present invention provides a virtual machine startup method, the method is applied to a network function virtualized NFV network where a first network service is deployed, the NFV network being a computer Implemented by multiple virtual machines running on the device, the first network service is performed jointly by the N types of virtual network function components VNFC, the method consists of N types of VNFC and multiple virtual machines Obtaining a distribution relationship between the machines, wherein the distribution relationship indicates the amount of each of the N types of VNFCs in each of the plurality of virtual machines; and a first of the virtual machines At least one start according to the distribution relationship, so that the total amount of each type of VNFC in the group of groups meets the preset amount corresponding to each type of VNFC Selecting a virtual machine that is not to be started as a virtual machine set to be started from a plurality of virtual machines, wherein the first set of virtual machines is a virtual machine set to be started and a virtual machine set to be started The started virtual machines belong to multiple virtual machines when the amount of started virtual machines in the started virtual machine set is greater than or equal to 0 and the amount is greater than 0, and Starting one virtual machine in the set of virtual machines to be started.

  The method allows appropriate and selective selection of as few virtual machines as possible according to the type and amount of VNFC in each virtual machine, so that the first network service on the NFV network can be implemented as quickly as possible .

  According to a second aspect, an embodiment of the present invention provides a virtual machine starter, wherein the device is applied to a network function virtualized NFV network on which a first network service is deployed, the NFV network being a computer Implemented by multiple virtual machines running on the device, the first network service is performed jointly by N types of virtual network function components VNFC, N is a positive integer, and N devices An acquisition module configured to acquire a distribution relationship between a plurality of types of VNFCs and a plurality of virtual machines, wherein the distribution relationship of N types of VNFCs in each of the plurality of virtual machines The total amount of each type of VNFC in the acquisition module and the first group of virtual machines showing the respective amount meets the preset amount corresponding to each type of VNFC A selection module configured to select from a plurality of virtual machines as a set of virtual machines to be started, according to the distribution relationship acquired by the acquisition module, The first set of groups includes a started virtual machine set and a virtual machine set to be started, and the amount of started virtual machines in the started virtual machine set is greater than or equal to 0 and the amount is 0 Configured to start a selection module and one virtual machine in a set of virtual machines to be started selected by the selection module, when the started virtual machine belongs to multiple virtual machines Including a starting module.

  The device can appropriately and selectively select as few virtual machines as possible according to the type and amount of VNFC in each virtual machine, so that network services on the NFV network can be performed as quickly as possible.

  The technical solutions in the first aspect and the second aspect can be applied to a scenario in which a single service type is required. In this case, the scenario in which a single service type is required includes two cases. That is, in one case, there is only one type of service with a service priority of SLA (Service Level Agreement) availability level on the NFV network, and another with a different service priority on the network. This is a case where there is no service and all virtual machines on the network support service-type services with that priority. The other case is that there are multiple services on the NFV network, and the virtual machines on the network include VNFC corresponding to different services, but the purpose of the currently started virtual machine is simply to perform a specific service This is just a case.

According to the first aspect or the second aspect, in a possible form, the step of selecting at least one virtual machine to be started from the virtual machines that are not started according to the distribution relationship comprises:
step 1:

Determine the current quantity weight of each type of VNFC in the first group of virtual machines according to: α _i is the current quantity weight of the i-th type of VNFC and Qi is the first group of virtual machines Is the current total amount of the i-th type of VNFC, and Qj is the current total amount of the j-th type of VNFC in the first group of virtual machines, i-th type of VNFC and j-th A type of VNFC belongs to N types of VNFCs, and the value range of i and j includes a positive integer not greater than N,
Step 2: According to the determined current weight of each type of VNFC and the preset weight of each type of VNFC, the current weight is maximally smaller than the preset weight Find the second type of VNFC, the kth type of VNFC is one of the N types of VNFC,
Step 3: Select an unstarted virtual machine containing the kth type of VNFC from multiple virtual machines to be added to the set of virtual machines to be started, and
Step 1, Step 2, and Step until the total amount of each type of VNFC that is in the N types and in the first group of virtual machines meets the preset amount corresponding to each type of VNFC Step 3 in order,
including.

  According to the technical solution, the algorithm can be used to selectively select the virtual machine in which the most scarce type of VNFC is located according to the type and amount of VNFC in each virtual machine, and the first Assuming fast implementation of network services, the maximum efficiency of VNFC in the currently started virtual machine can be fully realized, thereby quickly improving the service capacity of the first network service.

  According to the first aspect or the second aspect, in a possible form, an unstarted virtual machine comprising the kth type of VNFC is selected from a plurality of virtual machines to be added to the set of virtual machines to be started Selecting a non-started virtual machine containing the maximum amount of the kth type of VNFC to be added to the set of virtual machines to be started, or a plurality of unstarted virtual machines Select the unstarted virtual machine that contains the maximum amount of VNFC types among the N types to be added to the set of virtual machines to be started, Add the unstarted virtual machine containing the maximum total amount of VNFC used to implement the first network service to the virtual machine set to be started Selected to be, including the step, any one of.

  According to the technical solution, a preferential selection solution can be provided when multiple unstarted virtual machines meet the same filtering conditions when the system selects one virtual machine.

  According to the first aspect or the second aspect, in a possible form, in the process of repeating step 1, step 2, and step 3 in order, a selection order of at least one unstarted virtual machine is obtained, Also, starting one virtual machine in the virtual machine set to be started includes starting at least one unstarted virtual machine according to a selection order.

  The technical solution clearly provides an orchestration order for starting the virtual machines in the process of selecting at least one unstarted virtual machine. Compared to the prior art, the startup sequence is no longer based solely on the virtual machine number, and the maximum efficiency of the VNFC in the started virtual machine can be achieved according to the sequence.

  According to the first aspect or the second aspect, in a possible form, the step of selecting at least one unstarted virtual machine from a plurality of virtual machines as a set of virtual machines to be started is K pieces according to a distribution relationship. Determining the type of K, where K types are a subset of N types, and K types of VNFCs are not present in the started virtual machine, and a first preset Selecting L non-started virtual machines from a plurality of virtual machines to be added to the set of virtual machines to be started according to the rules, wherein L non-started virtual machines have K types of VNFCs L is not greater than K according to the first preset rule, L and H unstarted virtual machines are started according to the second preset rule Selecting from a plurality of virtual machines to be added to the set of virtual machines, the total amount of each type of VNFC in the first group of virtual machines according to a second preset rule Satisfying a preset amount corresponding to the VNFC. The first preset rule may preferentially select virtual machines that are not started with most of the K types. For example, K = 3, and the types are k1, k2, and k3 separately. A virtual machine including all three types of VNFCs k1, k2, and k3 is preferentially selected. If there is no virtual machine that includes all three types of VNFC k1, k2, and k3, a virtual machine that includes two types of VNFCs k1, k2, and k3 is selected. If there are still no virtual machines containing two types of VNFC k1, k2 and k3, then a virtual machine containing one type of VNFC k1, k2 and k3 is selected. The second preset rule may be in a random order, so only a simple addition principle needs to be followed. Alternatively, the second pre-configured rule may be in quantity priority order, ie the virtual machine containing the maximum amount of VNFC corresponding to the first network service (just considering the quantity, not the type) takes precedence Selected. Alternatively, the second pre-configured rule may be in type priority order, i.e., the virtual machine containing the largest type of VNFC corresponding to the first network service is preferentially selected. In a particular implementation of any one of the aforementioned pre-set rules, a random aspect may be used for selection if multiple virtual machines have the same priority.

  According to the technical solution, as few virtual machines as possible can be started, thereby ensuring basic implementation of the first network service, and the virtual machine has failed or a new virtual machine system At least the function of the first network service can be restored or set up when is built.

  According to the first aspect or the second aspect, in a possible form, in the process of selecting L non-started virtual machines from a plurality of virtual machines according to the first preset rule, L start-ups A first selection order of virtual machines not to be acquired and a second of the H unstarted virtual machines in the process of selecting H unstarted virtual machines from the plurality of virtual machines according to a second preset rule And starting one virtual machine in the set of virtual machines to be started first starts L non-started virtual machines according to the first selection order and then Starting H non-started virtual machines according to a selection order of 2;

  The virtual machines are orchestrated to start according to the selection order described above, so that the started virtual machine can contain as many different VNFCs as possible, and the first network service can be quickly Recovered or set up.

  According to the first aspect or the second aspect, in a possible form, the attribute information table of any virtual machine includes a VNFC field, and the i-th type of VNFC running on any virtual machine The amount of data is recorded in the VNFC field, and the step of obtaining the distribution relationship between N types of VNFC and multiple virtual machines is running on any virtual machine by reading the VNFC field including obtaining an amount of an i-th type of VNFC.

  This is a method for obtaining the distribution relationship between VNFC and virtual machine, and according to the distribution relationship, the system learns the VNFC activation state in each virtual machine and the distribution state of each type of VNFC in the virtual machine can do.

  According to the first aspect or the second aspect, in a possible form, any virtual machine includes a software package, and the amount of the i-th type of VNFC is recorded in the software package, and N The step of acquiring the distribution relationship between the type of VNFC and the plurality of virtual machines includes the step of acquiring the amount of the i-th type of VNFC included in any virtual machine by taking in the software package.

  This is another way to get the distribution relationship between VNFC and virtual machine, and according to the distribution relationship, the system will determine the VNFC activation state in each virtual machine and the distribution state of each type of VNFC in the virtual machine. Can learn.

  According to the first aspect or the second aspect, in a conceivable form, after the at least one virtual machine to be started is started, there are no remaining virtual machines to be started that have not been started. It may be started in a random manner. In general, every time a virtual machine is started, the remaining virtual machines exist, assuming that the target VNFC usage is met, thereby ensuring the reliability of the VNFC. According to the method, after the target VNFC usage has been met, order restrictions are no longer imposed on virtual machines to be started later, thus reducing the processing power consumption of the system.

According to a third aspect, an embodiment of the present invention provides a virtual machine startup method, the method is applied to a network function virtualized NFV network in which M network services are deployed, the NFV network being a computer Implemented by multiple virtual machines running on the device, the xth service is one network service in M network services, and the xth service is collaborated by N (x) types of VNFCs Where N is a positive integer, the value range of x includes all positive integers not greater than M, M is a natural number greater than 1, and the method has N (x) types Obtaining a distribution relationship between the VNFC of the virtual machine and the plurality of virtual machines, wherein the distribution relationship indicates an amount of VNFC of the xith type distributed in each of the plurality of virtual machines, Type of Is one of N (x) types corresponding to the xth service, and is a positive integer whose value range of i is not greater than N (x);
At least one startup according to the distribution relationship so that the total amount of each type of VNFC in the first group of virtual machines and in the M network services meets a preset amount corresponding to each type of VNFC Selecting a virtual machine that is not to be started from a plurality of virtual machines as a virtual machine set to be started, wherein the first set of virtual machines includes a started virtual machine set and a virtual machine set to be started The started virtual machine belongs to multiple virtual machines when the amount of started virtual machines in the started virtual machine set is greater than or equal to 0 and the amount is greater than 0, and Starting one virtual machine in the virtual machine set to be performed.

  According to the method, as few virtual machines as possible can be appropriately and selectively selected according to the type and amount of VNFC in each virtual machine, so that M network services can be implemented as quickly as possible.

  According to a fourth aspect, an embodiment of the present invention provides a virtual machine starter, the device is applied to a network function virtualized NFV network in which M network services are deployed, and the NFV network is a computer Implemented by multiple virtual machines running on the device, the xth service is one network service in M network services, and the xth service is collaborated by N (x) types of VNFCs N is a positive integer, includes all positive integers whose value range of x is not greater than M, M is a natural number greater than 1, and the device has N (x) types An acquisition module configured to acquire a distribution relationship between a VNFC and a plurality of virtual machines, wherein the distribution relationship is distributed within each of the plurality of virtual machines The xith type is one of the N (x) types corresponding to the xth service, and any positive integer whose value range of i is not greater than N (x) Acquired so that the total amount of each type of VNFC in the first group of virtual machines and M network services in the first group of virtual machines meets the preset amount corresponding to each type of VNFC A selection module that selects at least one unstarted virtual machine from a plurality of virtual machines as a set of virtual machines to be started according to the distribution relationship obtained by the module, wherein the first set of virtual machines is started The amount of started virtual machines in the started virtual machine set is greater than or equal to 0 and the amount is greater than 0 A selection module, wherein the started virtual machine belongs to a plurality of virtual machines, and a startup module configured to start one virtual machine in the set of virtual machines to be started selected by the selection module, Including.

  The device can appropriately and selectively select as few virtual machines as possible according to the type and amount of VNFC in each virtual machine, so that network services on the NFV network can be performed as quickly as possible.

According to the third aspect or the fourth aspect, in a possible form, selecting at least one unstarted virtual machine from a plurality of virtual machines as a set of virtual machines to be started according to the distribution relationship comprises:
step 1:

Determine the current quantity weight of the xith type VNFC in the first group of virtual machines according to: α _xi is the current quantity weight of the xith type VNFC and Qxi is the first quantity of the virtual machine Is the current total amount of xith type VNFC in the group of, Qxj is the current total amount of xjth type VNFC in the first group of virtual machines, and the xjth type is xth Is one of N (x) types corresponding to the service of and includes all positive integers whose j value range is not greater than N (x),
Step 2: According to the determined current weight α α _xi of the xith type VNFC and the preset quantity weight of the xith type VNFC, the current quantity weight is greater than the preset quantity weight Find the maximally small kth type of VNFC, where the kth type of VNFC is one of N (x) types of VNFCs,
Step 3: Select an unstarted virtual machine containing the kth type of VNFC from multiple virtual machines to be added to the set of virtual machines to be started, and
Step 1, Step 2, and so on until the total amount of each type of VNFC in the first group of virtual machines in the M network services meets the preset amount corresponding to each type of VNFC, and , Repeat step 3 in order,
including.

  According to the technical solution, the algorithm can be used to selectively select the virtual machine in which the most scarce type of VNFC is located according to the type and amount of VNFC in each virtual machine, and M It is possible to achieve the maximum efficiency of VNFC in the virtual machine that is currently started on the premise of ensuring quick implementation of the service.

According to the third aspect or the fourth aspect, in a possible form, selecting at least one unstarted virtual machine from a plurality of virtual machines as a set of virtual machines to be started according to the distribution relationship comprises:
Determining a descending order of priority of M services, and continuously performing a service authorization operation for all services in the M services according to the descending order of priority;
Service authorization work
Step 1: alpha current amount weights of each type of VNFC in the first group of virtual machine determined according _xi, alpha _xi is the current amount weights xi th type VNFC, Qxi the virtual machine Is the current total amount of the xi-th type VNFC in the first group of, and Qxj is the current total amount of the xj-th type VNFC in the first group of virtual machines, and the value range of xj is Contains all positive integers not greater than N (x) corresponding to the current service,
Step 2: The current quantity weight is preset according to the determined current quantity weight α _xi of each type of VNFC in the current service and the preset quantity weight of each type of VNFC in the current service Find the kth type of VNFC that is maximally smaller than the given quantity weight, where the kth type is one of N (x) types in the current service,
Step 3: Select an unstarted virtual machine containing the kth type of VNFC from multiple virtual machines to be added to the set of virtual machines to be started, and
Step 1, Step 2, until the total amount of each type of VNFC in the first group of virtual machines in the current service meets the preset amount corresponding to each type of VNFC in the current service, And repeating step 3 in order,
including.

  According to the technical solution, the algorithm can be used to selectively select the virtual machine in which the most scarce type of VNFC is located according to the type and amount of VNFC in each virtual machine, and M It is possible to achieve the maximum efficiency of VNFC in the virtual machine that is currently started on the premise of ensuring quick implementation of the service. In addition, service priority is distinguished. Therefore, the maximum efficiency of VNFC in the high priority service can be preferentially realized and then VNFC in the low priority service can be taken into account, in particular, ensuring fast recovery of high priority. The

According to the third aspect or the fourth aspect, in a possible form, an unstarted virtual machine including the kth type of VNFC is selected from a plurality of virtual machines to be added to the set of virtual machines to be started The step of performing is as follows:
Selecting an unstarted virtual machine containing the maximum amount of the kth type of VNFC to be added to the set of virtual machines to be started, or
If multiple unstarted virtual machines contain the maximum amount of the kth type of VNFC, select the unstarted virtual machine containing the maximum amount of VNFC to be added to the set of virtual machines to be started, or Includes any one of the steps of selecting a virtual machine that will not be started containing the maximum amount of VNFCs to be added to the set of virtual machines to be started.

  According to the technical solution, a preferential selection solution is given when a plurality of unstarted virtual machines meet the same filtering condition when the system selects at least one unstarted virtual machine.

  According to the third aspect or the fourth aspect, in a possible form, the selection order of at least one selected unstarted virtual machine is obtained in the process of repeating Step 1, Step 2, and Step 3. Also, starting one virtual machine in the virtual machine set to be started includes starting at least one selected unstarted virtual machine according to a selection order.

  An orchestration order for starting the virtual machines in the process of selecting at least one unstarted virtual machine is clearly provided. Compared to the prior art, the startup sequence is no longer based solely on the virtual machine number, and the maximum efficiency of the VNFC in the started virtual machine can be achieved according to the sequence.

  According to the third aspect or the fourth aspect, in a possible form, at least one selected in the process of continuously performing service authorization work for all services in the M services in descending order of priority. A selection order of unstarted virtual machines is obtained and starting one virtual machine in the set of virtual machines to be started includes starting at least one selected unstarted virtual machine according to the selection order including.

  An orchestration order for starting the virtual machines in the process of selecting at least one unstarted virtual machine is clearly provided. Compared to the prior art, the startup sequence is no longer based solely on the virtual machine number, and the maximum efficiency of the VNFC in the started virtual machine can be achieved according to the sequence.

  According to the third aspect or the fourth aspect, in a possible form, the step of selecting at least one unstarted virtual machine from a plurality of virtual machines as a set of virtual machines to be started is K pieces according to the distribution relation. Determining the types of K, where K types are a subset of all VNFC types corresponding to M services, and K types of VNFCs are not present in the started virtual machine, and Selecting L non-started virtual machines from a plurality of virtual machines to be added to the set of virtual machines to be started according to a first preset rule, wherein L unstarted virtual machines are K With 5 types of VNFC, L is not greater than K according to the first preset rule, and H starts according to the second preset rule Selecting from a plurality of virtual machines to be added to a set of virtual machines to be started and included in a first group of virtual machines according to a second preset rule And the total amount of each type of VNFC in the M network services satisfies a preset amount corresponding to each type of VNFC. For example, K = 3, and the types are k1, k2, and k3 separately. A virtual machine including all three types of VNFCs k1, k2, and k3 is preferentially selected. If there is no virtual machine that includes all three types of VNFC k1, k2, and k3, a virtual machine that includes two types of VNFCs k1, k2, and k3 is selected. If there are still no virtual machines containing two types of VNFC k1, k2 and k3, then a virtual machine containing one type of VNFC k1, k2 and k3 is selected. The second preset rule may be in a random order, so only a simple addition principle needs to be followed. Alternatively, the second pre-configured rule may be in quantity priority order, i.e. the virtual machine preferentially contains the maximum amount of VNFC corresponding to the corresponding service (just considering the amount, not the type) Selected. Alternatively, the second pre-configured rule may be in type priority order, i.e. the virtual machine containing the largest amount of type of VNFC corresponding to the first service is preferentially selected. In a particular implementation of any one of the aforementioned pre-set rules, a random aspect may be used for selection if multiple virtual machines have the same priority.

  According to the new orchestration method for starting a virtual machine given in the technical solution, as few virtual machines as possible can be started, thereby ensuring the basic implementation of the service and the virtual machine Can be restored or set up when M is out of service or a new virtual machine system is built.

  According to the third aspect or the fourth aspect, in an envisaged form, in the process of selecting L unstarted virtual machines from a plurality of virtual machines according to the first preset rule, L start In the process of obtaining a first selection order of virtual machines to be taken and selecting H non-started virtual machines from a plurality of virtual machines according to a second preset rule, H virtuals to be started The second selection order of machines is obtained and starting one virtual machine in the set of virtual machines to be started first starts L non-started virtual machines according to the first selection order and then , Starting H unstarted virtual machines according to a second selection order.

  According to the new orchestration method for starting a virtual machine given in the technical solution, as few virtual machines as possible can be started, thereby ensuring the basic implementation of the service and the virtual machine Can be restored or set up when M is out of service or a new virtual machine system is built.

  According to the third aspect or the fourth aspect, in a possible form, the step of selecting at least one unstarted virtual machine from a plurality of virtual machines as a set of virtual machines to be started comprises: Determine the descending order of priority, perform service assurance work continuously for all services in M services according to descending priority order, and make H1 unstarted virtual machines into virtual machine sets to be started After selecting the multiple virtual machines to be added and performing the service assurance work for all services in M services, start H2 unstarted virtual machines according to the fourth preset rule Selecting from a plurality of virtual machines to be added to a set of virtual machines to be performed, a fourth preconfigured In accordance with the rules, the total amount of each type of VNFC contained in the first group of virtual machines and in the M network services satisfies a preset amount corresponding to each type of VNFC. , Service assurance work is the step of determining K (x) types according to the distribution relation, where K (x) types are a subset of N (x) types in the current service, and K ( x) There are no types of VNFCs in the started virtual machine, steps and L (x) unstarted virtual machines are added to the virtual machine set to be started according to the third preset rule To select from multiple virtual machines, where L (x) unstarted virtual machines contain K (x) types of VNFCs and L (x) is the third preset rule Greater than K (x) according to Not including steps.

  The technical solution provides a virtual machine orchestration selection method. According to this orchestration order, as few virtual machines as possible are started, and basic implementation of the service can be ensured. In particular, when a plurality of services exist, the high priority service can be preferentially implemented. Therefore, at least a service function can be secured or set up when a virtual machine is out of order or when a new virtual machine system is constructed.

  According to the third aspect or the fourth aspect, in a possible form, the third selection order of H1 unstarted virtual machines in the process of performing service assurance work for all services in M services In the process of selecting H2 unstarted virtual machines from multiple virtual machines according to the fourth preset rule, the fourth selection order of H2 unstarted virtual machines is acquired and started. The step of starting one virtual machine in the set of virtual machines to be started first starts H1 unstarted virtual machines according to the third selection order and then H2 unstarted virtuals according to the fourth selection order Including starting the machine.

  According to the technical solution, as few virtual machines as possible can be started, ensuring basic implementation of the service. In particular, when a plurality of services exist, the high priority service can be preferentially implemented. Therefore, at least a service function can be secured or set up when a virtual machine is out of order or when a new virtual machine system is constructed.

  According to the third aspect or the fourth aspect, in a possible form, the attribute information table of any virtual machine includes a VNFC field, and the xith type of VNFC running on any virtual machine. The amount of N is recorded in the VNFC field, and also the step of obtaining the distribution relationship between N (x) types of VNFC and the virtual machine works on any virtual machine by reading the VNFC field Including obtaining the amount of VNFC of the xith type.

  This is a method for obtaining the distribution relationship between VNFC and virtual machine, and according to the distribution relationship, the system learns the VNFC activation state in each virtual machine and the distribution state of each type of VNFC in the virtual machine can do.

  According to the third aspect or the fourth aspect, in a possible form, any virtual machine includes a software package, and the amount of the xi-th type of VNFC is recorded in the software package, and N ( x) The step of obtaining the distribution relationship between the VNFCs of the number of types and the virtual machine includes the step of obtaining the amount of the NFth type of VNFC included in any virtual machine by incorporating the software package. .

  This is another way to get the distribution relationship between VNFC and virtual machine, and according to the distribution relationship, the system will determine the VNFC activation state in each virtual machine and the distribution state of each type of VNFC in the virtual machine. Can learn.

  According to the third aspect or the fourth aspect, in a possible form, after at least one virtual machine to be started is started, the remaining virtual machines to be started that have not been started are random aspects. And after the subject VNFC usage has been met, there are no longer any order restrictions on the virtual machines to be started later.

  In general, every time a virtual machine is started, the remaining virtual machines exist, assuming that the target VNFC usage is met, thereby ensuring the reliability of the VNFC. If there is no order restriction, the processing power consumption of the system can be reduced.

  According to a fifth aspect, the present invention provides a virtual machine startup device, the device including at least one processor, a memory, and a bus, wherein the processor and the memory communicate by using the bus. Connected, the memory stores the program and instructions, and the processor is configured to recall the program and instructions to the memory and implement any one of the methods of the foregoing forms or equivalent methods .

  In the prior art, to avoid virtual machine boot storms, it is necessary to limit the amount of virtual machines that are started simultaneously. Thus, when there is a restriction on the startup amount, the virtual machine is generally started according to the host instruction or the VNF instruction. In this case, not all types of VNFCs required to perform a specific service (target service) can be started quickly, and a relatively low service startup efficiency is provided. However, in embodiments of the present invention, the distribution of the type and amount of VNFC within a virtual machine can be obtained, a virtual machine to be started according to service requirements can be selected, and the order for the virtual machines to be started is It is determined. According to this method, a virtual machine used to start a type of VNFC corresponding to a specific service or a high priority service, instead of starting the virtual machines randomly in a batch state as in the prior art. Can be preferentially started so that service is provided quickly and service throughput is improved quickly.

It is a schematic distribution map of VNFC in a virtual machine. 1 is a schematic diagram of a system architecture of an NFV system 000 that may be applied according to an embodiment of the present invention. It is the schematic which shares a physical resource with a virtual machine. FIG. 4 is a schematic diagram of a correspondence between a service on an NFV network and VNFC. It is a schematic flowchart of the virtual machine starting method which concerns on one Embodiment of this invention. It is a figure of a correspondence between VNF and VNFC in the session control service concerning one embodiment of the present invention. 4 is a schematic flowchart of a virtual machine starting method in a multi-service scenario according to an embodiment of the present invention. The comparative example which orchestrates a virtual machine by using a prior art is shown. The comparative example which orchestrates a virtual machine by using the method in this invention is shown. It is a schematic structure figure of the virtual machine starting device concerning one embodiment of the present invention. It is the schematic of the selection module 303 in the virtual machine starting device which concerns on one Embodiment of this invention. It is the schematic of the selection module 303 in the other virtual machine starting device which concerns on one Embodiment of this invention. It is a computer schematic structure figure of the virtual machine starting device concerning one embodiment of the present invention.

  The following clearly and completely describes the technical solutions in the embodiments of the present invention with reference to the accompanying drawings in the embodiments of the present invention. Apparently, the described embodiments are some but not all of the embodiments of the present invention. All other embodiments obtained by a person of ordinary skill in the art based on the embodiments of the present invention without creative efforts shall fall within the protection scope of the present invention.

  Referring to FIG. 2a, FIG. 2a is a schematic diagram of a system architecture of an NFV system 000 that may be applied according to an embodiment of the present invention. It should be noted that the NFV system is merely used as an illustrative example, and the scope of the present invention is not limited thereto. The NFV system shown in FIG. 2a may be implemented by using multiple networks, for example, a data center network, a service provider network, or a local area network (LAN). As shown in Figure 2a, the NFV system consists of the NFV Management and Orchestration System 028, NFV Infrastructure (NFVI) 030, multiple virtual network functions 008, multiple element management systems ( element management system (EMS) 022, service, VNF and infrastructure description system 026, and one or more operation support systems / business support systems (operation support systems / business support) System, OSS / BSS) 024 may be included.

  The NFV management and orchestration system 028 may include an orchestrator 002, one or more VNF managers 004, and one or more virtualized infrastructure managers 006. NFVI030 includes a hardware resource layer including computing hardware 012, storage hardware 014, and network hardware 016, a virtualization layer, virtual computing 010 (eg, virtual machine), virtual storage 018, and A virtual resource layer including the virtual network 020 may be included. The computing hardware 012 may be a special purpose processor or a general purpose processor configured to provide processing and computing functions. Storage hardware 014 is configured to provide storage capability. The storage capability may be provided by storage hardware 014 (eg, the server's local memory) or may be provided by a network (eg, the server is connected to a network storage device by using the network). Good. The network hardware 016 may be a switch, router, and / or other network device. The network hardware 016 is configured to perform communication between a plurality of devices. The plurality of devices are connected in a wired manner or a wireless manner. The virtualization layer in the NFVI 030 is configured to abstract the hardware resource in the hardware resource layer, separate the physical layer to which the hardware resource belongs and the VNF 008, and give the virtual resource to the VNF 008. As shown in FIG. 2a, the virtual resource layer may include a virtual computing 010, a virtual storage 018, and a virtual network 020. The virtual computing 010 and the virtual storage 018 may be provided in the VNF 008 in the form of a virtual machine, or may be provided in the virtual container in another form. For example, one or more VNFs 008 may be deployed in one virtual machine. The virtualization layer abstracts the network hardware 016 to form the virtual network 020. A virtual network 020, eg, virtual switches (eg, Vswitches), is used to perform communications between multiple virtual machines or other types of virtual containers with VNFs. Network hardware virtualization can be virtual LAN (Virtual LAN, VLAN), virtual private local area network service (Virtual Private LAN Service, VPLS), virtual eXtensible Local Area Network (VxLAN), or It may be implemented by using techniques such as Network Virtualization using Generic Routing Encapsulation (NVGRE). OSS / BSS 024 is mainly intended for telecommunications service operators and has comprehensive network management functions and business operation functions including network management (fault monitoring and network information collection, etc.), billing management, customer service management, etc. give. The Service / VNF / Infrastructure Description System 026 is described in detail in the ETSI GS NFV 002 v1.1.1 standard. Accordingly, details are not described in this embodiment of the invention.

  The NFV management and orchestration system 028 may be configured to monitor and manage the VNF 008 and the NFVI 030. Orchestrator 002 may perform resource-related requests and send configuration information to VNF manager 004 and may communicate with one or more VNF managers 004 to collect VNF 008 status information. The orchestrator 002 may also communicate further with the virtualization infrastructure manager 006 to perform resource allocation and / or to reserve and exchange virtualization hardware resource state information and configuration information. The VNF manager 004 may be configured to manage one or more VNFs 008 and perform various management functions such as initializing, updating, querying and / or stopping the VNFs 008. The virtualization infrastructure manager 006 controls the interaction between the VNF 008 and the computing hardware 012, the storage hardware 014, the network hardware 016, the virtual computing 010, the virtual storage 018, or the virtual network 020. It may be configured to manage. For example, the virtualization infrastructure manager 006 may be configured to perform work to allocate resources to the VNF 008. The VNF manager 004 and the virtual infrastructure manager 006 may communicate with each other to exchange configuration information and state information of virtual hardware resources.

  NFVI030 includes hardware and software. Both hardware and software jointly build a virtualized environment to deploy, manage and execute VNF008. In other words, the hardware resource and virtual resource layer are configured to provide virtual resources to VNF 008, eg, to virtual machines and / or virtual containers in other forms.

  As shown in FIG. 2a, VNF manager 004 may communicate with VNF 008 and EMS 022 to perform VNF lifecycle management and exchange configuration / status information. VNF 008 is a virtualization of at least one network function, and the network function is given in advance by a physical network device. In one embodiment, VNF 008 may be a virtualized mobility management entity (MME) node and is configured to provide network functionality provided by a typical non-virtualized MME device. In other embodiments, the VNF 008 may be configured to perform some function of all components provided by the non-virtualized MME device. One or more VNFs 008 may be deployed on one virtual machine (or one virtual container in other forms). The EMS022 may be configured to manage one or more VNFs.

  The MME is an important control node in a 3GPP (3rd Generation Partnership Project) protocol LTE (Long Term Evolution) access network. The MME is involved in positioning, paging and relaying UE (User Equipment) in idle mode. That is, MME is involved in signal transmission processing. When the UE is initialized and connected to the MME, the MME is involved in selecting an SGW (Serving Gateway) for the UE. The MME is further configured to authenticate a user by interacting with a Home Subscriber Server (HSS) and assign a temporary ID (identifier) to the user. MME will also support interception and listening to the extent permitted by law.

  The present invention relates to orchestration and management in a virtual network function. Therefore, the method and apparatus involved in the present invention mainly corresponds to the virtual network function management / orchestration system 028.

  In the present invention, it is necessary to first consider the VNFC distribution state in the NFV network. A large number of virtual machines run on a large NFV network. These virtual machines may share some physical resources. Physical resources include a host, shared storage, and the like. For example, as shown in FIG. 2b, VM0 and VM1 share host 1, VM2 and VM3 share host 2, VM4 and VM5 share host 3, VM6 and VM7 share host 4, In addition, VM0, VM1, VM2, VM3, VM4, VM5, VM6, and VM7 share storage resources. VNFC components run on virtual machines. In general, the same VNFC runs on multiple virtual machines to provide stronger service processing capability or higher service reliability. Some virtual machines also activate a plurality of different virtual network function components. As shown in Figure 2b, VNFC000 and VNFC001 move on both VM0 and VM2, VNFC002 moves on both VM1 and VM3, VNFC010 moves on both VM4 and VM5, and VNFC001 on both VM6 and VM7 Move.

  In practical applications, an NFV network needs to successfully provide one or more network services (services for short), and each network service corresponds to a specific VNFC set. Thus, the aforementioned NFV network requires one or more specific VNFC sets to work properly. In general, the VNFC sets for different types of services are different. For example, as shown in FIG. 3, service 0 needs to be jointly performed by VNFC in set {VNFC000, VNFC001, VNFC010, VNFC011}, and service 1 is set {VNFC000, VNFC002, VNFC020 , VNFC021} needs to be implemented jointly by VNFC. VNFC000, VNFC001, and VNFC002 belong to VNF00, VNFC010 and VNFC011 belong to VNF01, and VNFC020 and VNFC021 belong to VNF02. It should be noted that generally different VNFs do not correspond to the same VNFC.

  In a specific implementation process, a large NFV network can provide many network services such as voice services and Rich Communication Suite RCS (Short Message Service Message and Picture Sharing) services. For any particular service, the NFV network requires a specific VNFC set. The set includes at least one type of VNFC, and the VNFC types in the VNFC set are different from each other. All VNFCs in the set cooperate with each other to implement a particular network service, and none are unnecessary.

  In the prior art, to avoid virtual machine boot storms, it is necessary to limit the amount of virtual machines that are started simultaneously. Therefore, when there is a restriction on the starting amount of the virtual machine, the virtual machine is generally started according to the host instruction or the VNF instruction. In this case, it is not always possible to quickly start all of the multiple types of VNFCs required to implement a specific service (target service). Thus, a relatively low service startup efficiency is provided. This problem needs to be solved especially in the process of starting virtual machines in a centralized manner or in scenarios where a large number of faulty virtual machines need to be recovered. Therefore, to solve this problem, this specification describes a virtual machine that can determine and plan the virtual machine startup order in an appropriate manner according to the VNFC distribution in the virtual machine so that the target service can be restored as soon as possible. A starting method and apparatus are provided. Meeting the target service requirements mainly includes the following two aspects.

  (1) The target service can be implemented as quickly as possible, and all VNFCs that provide the target service can be started. For example, if the target service is service 0, it is necessary to start all four VNFCs, namely VNFC000, VNFC001, VNFC010, and VNFC011, in order to perform or restore the function of service 0.

  (2) To meet the quantity requirements for the target service, the amount of each type of VNFC in the target service reaches the preset amount as quickly as possible. For example, if the target service is service 1 (VNFC corresponding to service 1 includes VNFC000, VNFC002, VNFC020, and VNFC021), the service provider can collect VNFC000, VNFC002, The estimated activation amounts of VNFC020 and VNFC021 are acquired, and generally, the estimated amount of each type of VNFC is different. The user may use VNFC000 for most of the time and only use VNFC002 for a short time. It is assumed that the amounts of VNFC000, VNFC002, VNFC020, and VNFC021 expected by the service provider are 100, 30, 20, and 10, respectively. In this case, the service provider has selected some virtual machines so that the amount of VNFC000, VNFC002, VNFC020, and VNFC021 that will eventually run on the started virtual machine reaches 100, 30, 20, and 10, respectively. Must be started as quickly as possible. In a process where not all VNFCs are started, the service provider can use VNFC000, VNFC002 to fully utilize the limited amount of VNFC000, VNFC002, VNFC020, and VNFC021 in the currently started virtual machine. We expect that the ratio of the amounts of VNFC020 and VNFC021 can approach 100: 30: 20: 10.

  The application scenario of the present invention may be related to single service type and multiple service type according to different requirements, so the two scenarios will be described separately below.

  1． Single service scenario

  Scenarios that require a single service type include two cases. That is, in one case, there is only one type of service with a service priority of SLA (Service Level Agreement) availability level on the NFV network, and another with a different service priority on the network. This is the case where there is no service and all virtual machines on the network support this type of service at that priority. The other case is that there are multiple services on the NFV network, and the virtual machines on the network include VNFC corresponding to different services, but the purpose of the currently started virtual machine is simply to perform a specific service This is just a case.

  Referring to FIG. 4, FIG. 4 is a schematic flowchart of a virtual machine starting method according to an embodiment of the present invention. The method 100 includes the following steps.

  Step 101: Determine N types of virtual network functional components VNFC according to a first network service, in which case the first network service can be implemented by a virtual machine, eg voice transmission or multimedia transmission Any network service within a network service, where N types of VNFCs form a first VNFC set, and N is a positive integer.

  Step 102: Obtain a distribution relationship between N types of VNFCs and virtual machines, where the distribution relationship is the amount of i-th type VNFCs distributed within each virtual machine within the target search range Each virtual machine includes a virtual machine that is started or not started, the i-th type is one of the N types described above, and the value range of i is less than or equal to N Contains all positive integers.

  A virtual machine that is not started is a virtual machine that was not powered on for startup, and may include a virtual machine that is automatically detected by the system and needs to be restarted due to a failure, or a system It may include virtual machines that need to be started in the initialization phase. The started virtual machine may include a virtual machine that has already completed the start-up operation, and a virtual machine that has been started and worked before performing the start-up operation. Note that the virtual machine to be started mentioned in this application document is a virtual machine that is selected from virtual machines that are not started and that is orchestrated for starting. Thus, before the method is performed, the entire virtual machine system includes a started virtual machine (which may be zero in quantity) and a virtual machine that is not started. In the process of performing the method, the entire virtual machine system includes a started virtual machine, a virtual machine to be started, and a virtual machine that is not started.

  Step 103: At least one virtual machine that is not started according to the distribution relationship obtained at 102 so that the total amount of the i-th type of VNFC included in the first group of virtual machines satisfies a corresponding preset amount Select a virtual machine to be started, in which case the first set of virtual machines includes a started virtual machine and at least one selected unstarted virtual machine, and at least one A virtual machine that is not started may be used as a virtual machine set to be started.

  Step 104: Start at least one virtual machine to be started as determined in 103.

  The following uses the first service as an example for explanation. The VNFC set corresponding to the first service is {VNFC1, VNFC2, VNFC3, VNFC4}. The first service is not a specific service, is not intended to limit the order, and is a service in a general sense. Any service whose implementation principle is equivalent can be made within the protection scope of the present invention.

  Optionally, step 102 may specifically be as follows: That is, the attribute information table of an arbitrary virtual machine includes a VNFC field, and the amount of the i-th type VNFC running on the arbitrary virtual machine is recorded in the VNFC field and moved on the arbitrary virtual machine. The amount of i-th type VNFC is obtained by reading the VNFC field. Generally, the attribute information table is generated when a virtual machine is formed. Special purpose VNFC fields may be a set, or existing fields may be replaced to record the correspondence between virtual machines and VNFCs, as shown in the table below.

  Optionally, step 102 may specifically be as follows: That is, any virtual machine includes a software package, and the amount of the i-th type of VNFC is recorded in the software package, and any virtual machine learns the correspondence between the virtual machine and the VNFC. By importing a software package, the amount of the i-th type VNFC contained in any virtual machine is obtained. The description file in the software package is shown in the following table.

  The correspondence between the virtual machine and the VNFC may include the virtual machine identifier and type identifier of all VNFCs running on the virtual machine. For example, VM0 {10-VNFC000} indicates that 10 VNFC000s run on the virtual machine VM0, and VM1 {10-VNFC000, 20-VNFC001, 30-VNFC100} indicates 10 VNFC000, 20 VNFC001 In addition to acquiring the VNFC activation state on the virtual machine, such as indicating that 30 VNFCs 100 move on the virtual machine VM1, the VNFC distribution in the virtual machine may be acquired, for example, VNFC000 {10-VM0, 20-VM1, 30-VM3} indicates that the amount of VNFC000 distributed in VM0, VM1, and VM3 is 10, 20, and 30, respectively, and VNFC001 { 15-VM2, 25-VM3} indicates that the amounts of VNFC001 distributed in VM2 and VM3 are 15 and 25, respectively. It should be noted that the technical solution in the present invention is not limited to a specific configuration mode or a storage mode of corresponding configuration data between the virtual machine and the VNFC, and is not limited to the field representation mode.

  Optionally, step 103 may specifically be to perform a “service assurance operation” according to the distribution relationship acquired at 102, in which case the “service assurance operation” is specifically as follows: It may be.

  Determine whether the started virtual machine already contains four VNFC types corresponding to the first service.

  If the result of the decision that the started virtual machine does not include the four VNFC types corresponding to the first service, then one of the four VNFC types that is not included in the started virtual machine, i.e. Determine the type of VNFC that still needs to be started to perform the first service, and then select L virtual machines to be started (L is 4 or less) from the virtual machines that are not started, in this case , L virtual machines to be started include k VNFC types (k is 4 or less and L or more) that were in the first VNFC set and were not included in the started virtual machines. For example, the first service needs to be jointly implemented by VNFCs in the set {VNFC1, VNFC2, VNFC3, VNFC4}. If only VNFC1 and VNFC3 run on a started virtual machine, the first service cannot be implemented. This is because VNFC2 and VNFC4 are also required. In this case, it is necessary to select an unstarted virtual machine that contains VNFC2 or VNFC4, an unstarted virtual machine on which both VNFC2 and VNFC4 run may be preferentially selected, and a larger amount Virtual machines that are not started, including VNFC2 and VNFC4, are better virtual machines. If there are no unstarted virtual machines containing both VNFC2 and VNFC4, the unstarted virtual machine containing the maximum amount of VNFC2 may be selected first, followed by the unstarted virtual machine containing the maximum amount of VNFC4 Is done. The specific selection order is not limited. In short, an unstarted virtual machine with the maximum amount of VNFC type is selected, in which case the VNFC type is set to start virtual machine and L unstarted virtual machines in the first service. Until all corresponding types of VNFC are included, they are not present in the started virtual machine and in the virtual machine that is already orchestrated for startup. In an optional embodiment, after the L virtual machines to be started are orchestrated, the H virtual machines to be started are selected from unstarted virtual machines, in which case the started virtual machines; The amount of LNF virtual machines to be started and all types of VNFCs included in H virtual machines to be started and corresponding to the first service are the corresponding pre-set quantities Meet.

  If the decision is that the started virtual machine contains four VNFC types corresponding to the first service, select L virtual machines to be started from the virtual machines that are not started, in this case started The amount of all types of VNFCs included in the virtual machine and the L virtual machines to be started and corresponding to the first service satisfy the corresponding preset amount. Implementation ideas may be implemented using a single additional principle. For example, the amounts of VNFC1, VNFC2, VNFC3, and VNFC4 reach 400, 300, 200, and 100 or more, respectively.

  In another possible form, step 103 may specifically be to perform “service high efficiency work” according to the distribution relationship acquired in step 102, in which case “service high efficiency work”. Specifically, the method may include the following steps.

  Step 1031: Actual quantity weight of i-th type VNFC

Where Qi is the total amount of the i th type VNFC in the second group of virtual machines and Qj is the total amount of the j th type VNFC in the second group of virtual machines And the value range of j includes all positive integers not greater than N, and the second group of virtual machines referred to in the present invention is currently selected as the currently started virtual machine Including a virtual machine to be started.

Step 1032: According to the actual quantity weight α _i of the i-th type VNFC and the preset quantity weight of the i-th type VNFC, the actual quantity weight is a preset quantity of the k-th type Find the kth type maximally smaller than the weight, where the kth type is one of N types.

  Step 1033: Select a virtual machine containing the kth type of VNFC as a new virtual machine to be started, and if multiple virtual machines contain the maximum amount of the kth type of VNFC, then N types Select a virtual machine with the largest amount of VNFC type from multiple virtual machines, or select a virtual machine with the largest total amount of VNFC corresponding to the first service, in this case, corresponding to the first service VNFC is the VNFC of all types of N VNFC types.

  Steps 1031, 1032 and 1033 are repeated until the total amount of the i-th type VNFC included in the first group of virtual machines satisfies the corresponding preset amount. As a supplementary explanation, the second group of virtual machines in this application document is performed each time, i.e. in a new series of calculations performed each time a new virtual machine to be started is selected by orchestration. , May be further understood as the current first group of virtual machines, where the first group of virtual machines is updated in real time.

Hereinafter, an example will be described. In a particular implementation process, in the simplest case, there is one unstarted virtual machine and the amount of all components contained in the unstarted virtual machine is sufficient to meet a predetermined amount. However, that case is rare or rarely exists. In most practical cases, it is necessary to start many unstarted virtual machines in order to start all types of VNFCs that reach that predetermined amount, and generally at least several hundred unstarted virtual machines Need to be started. In this case, there is a need to solve the problem of determining which unstarted virtual machine needs to be started first among a plurality of unstarted virtual machines. As described above, a service provider obtains a basic quantity relationship of all components that meet user requirements by collecting a large amount of user usage habits. For example, the predetermined amounts of VNFC1, VNFC2, VNFC3, and VNFC4 required by the first service are 400, 300, 200, and 100, respectively. However, 400, 300, 200, and 100 are merely final values, and the started virtual machine cannot immediately reach the quantity relationship. The virtual machine to be started is gradually started and the amount gradually approaches a predetermined amount. When the amount ratio of the four types of components, namely VNFC1, VNFC2, VNFC3, and VNFC4, satisfies the condition of 400: 300: 200: 100 (ie, 4: 3: 2: 1) It is easy to learn from the final amount that the implementation efficiency is the highest. Thus, each type of component has a preset quantity weight, and the predetermined quantity weights for VNFC1, VNFC2, VNFC3, and VNFC4 are as follows:
R _VNFC1 = 4 / (1 + 2 + 3 + 4) = 0.4;
R _VNFC2 = 3 / (1 + 2 + 3 + 4) = 0.3;
R _VNFC3 = 2 / (1 + 2 + 3 + 4) = 0.2; and
R _VNFC4 = 1 / (1 + 2 + 3 + 4) = 0.1

In the actual case, the virtual machine to be started that is currently orchestrated for startup (ie, the currently selected virtual machine to be started) and each type of VNFC included in the currently started virtual machine The amount does not always meet the ratio but varies dynamically most of the time. Therefore, the service implementation efficiency is determined by the VNFC of one of these VNFCs whose actual weight is smaller than the preset weight. The current actual quantity weight of VNFC1, VNFC2, VNFC3, and VNFC4 needs to be determined first. For example, at a given moment, VNFC1, VNFC2, VNFC3 running on a second group of virtual machines (ie, currently started virtual machines and virtual machines to be started that are currently orchestrated for startup) , And the actual amount of VNFC4 is 50, 10, 10, and 30, and correspondingly, the actual amount weight is as follows.
R ′ _VNFC1 = 50 / (50 + 10 + 10 + 30) = 0.5;
R ′ _VNFC2 = 10 / (50 + 10 + 10 + 30) = 0.1;
R ′ _VNFC3 = 10 / (50 + 10 + 10 + 30) = 0.1; and
R ' _VNFC4 = 30 / (50 + 10 + 10 + 30) = 0.3

The deviation of the actual quantity weight from these preset quantity weights of these VNFCs in the second group of virtual machines is evaluated by using the formula S = (R′−R) / R Also good. At this point, the actual amount weight deviation from the preset amount weights of VNFC1, VNFC2, VNFC3, and VNFC4 in the second group of virtual machines is correspondingly as follows: is there.
S _VNFC1 = (0.5−0.4) /0.4=+25%;
S _VNFC2 = (0.1-0.3) /0.3=-66.7%;
S _VNFC3 = (0.1-0.2) /0.2=-50%; and
S _VNFC4 = (0.3-0.1) /0.1 = + 200%

  A positive excursion result indicates that this type of VNFC is excessive, a larger positive result indicates a higher degree of excess, and a negative excursion result indicates that this type of VNFC is low. It is also easy to learn that smaller negative results indicate a higher degree of deficiency. In this example, in the current quantity distribution of VNFC, VNFC1 and VNFC4 in the second group of virtual machines are relatively excessive, especially VNFC4 is relatively high, and VNFC2 and VNFC3 are relative In particular, it can be learned that the degree of deficiency is the maximum in VNFC2. Therefore, in order to allow the first service to use the current VNFC resources as quickly and fully as possible, the method for improving the current VNFC utilization efficiency as quickly as possible is to first reduce the amount of VNFC2 It is supposed to increase. In this case, it is necessary to select an unstarted virtual machine including VNFC2, and a virtual machine including the maximum amount of VNFC2 is preferentially selected. If more than one unstarted virtual machine contains the maximum amount of VNFC2, it is one of these unstarted virtual machines that has a larger amount of other VNFC types or a larger amount of other There is a higher probability of being able to preferentially select virtual machines that contain a type of VNFC. For example, a virtual machine that contains the maximum amount of VNFC2 and that does not start with VNFC3 is preferentially selected, or a virtual machine that contains the maximum amount of VNFC2 and contains VNFC1, VNFC3, and VNFC4 is preferentially selected Is done. In short, virtual machines that can make up for the missing VNFC resources as quickly as possible, especially virtual machines that can make up for all the missing VNFC resources, have a high selection priority. Each time a new virtual machine to be started is selected, the VNFC distribution in the second group of virtual machines is updated, and the system is preconfigured with VNFC in the second group of virtual machines Calculate and update the deviation of the actual weight from the weight. The results after each update are different. Thus, in the process of selecting virtual machines one by one for orchestration, the magnitude of both the positive and negative excursions of each VNFC component varies dynamically. The preferred selection mechanism used after each update and calculation is as described above. Thus, the current amount of VNFC can achieve its maximum efficiency in the selection process until all types of VNFC in the first group of virtual machines reach a predetermined amount. For example, after several virtual machines are started continuously, the required amount of VNFC1, VNFC2, VNFC3, and VNFC4 reaches 400, 300, 200, and 100, respectively. The first group can meet the user requirements for the first service, and subsequent selection of virtual machines to be started may no longer be limited by order constraints and may be done in a random manner It is suggested.

  Correspondingly, the virtual machines to be started are selected according to the selection order, and the order for starting these virtual machines to be started is the same as the selection order. For example, in the process of repeating step 1031 to step 1033, each time step 1031 to step 1033 is performed, the most suitable virtual machine to be started may be selected to be included in the startup sequence for orchestration. Good. After step 1031 to step 1033 are performed a plurality of times, there is a selection order of virtual machines to be started. Correspondingly, the virtual machine to be started earlier selected is ranked higher in the startup sequence and started earlier. In the virtual machine start mechanism, the selected virtual machines to be started are started in order.

  Optionally, the virtual machine startup mode is as follows. That is, first, an orchestration order for starting a virtual machine to be started is obtained. Next, a startup operation is performed for the virtual machine to be started according to the startup sequence. If there are too many virtual machines to be started, for example if the virtual machines to be started exceed the amount limited by the boot storm, assuming that the limited amount is c, c virtual machines will quickly It is started. Each virtual machine has a different activation state, and therefore its startup time is different. If c virtual machines to be started are started simultaneously in an instant, the virtual machines to be started ranked after c virtual machines cannot be started immediately. Similar to the stack overflow aspect, the next virtual machine in the sequence to be started can be started only after the start of the currently started virtual machine is completed. Obviously, in the method, the maximum efficiency of VNFC in the started virtual machine can still be implemented under virtual machine boot storm constraints.

  The above deviation degree calculation algorithm is merely used as an example for explanation. Other computational aspects may be further derived and transformed from the algorithm according to mathematical principles, and in this embodiment of the invention, the other computational aspects are not listed one by one. In addition, in the process of selecting virtual machines, when multiple unstarted virtual machines meet one condition, for example, the same type and amount of VNFC running on two unstarted virtual machines that meet one condition , One of a plurality of unstarted virtual machines may be selected in a random manner. Any simple equivalent solution without creative effort shall fall within the protection scope of the present invention.

  In a particular implementation process, step 103 may be further performed in multiple aspects, and multiple implementations may be derived by derivation due to different service policies that exist in the system. For example, in the process of selecting at least one virtual machine to be started (not limited to the whole process, including part of the process), the virtual machine that was started and the start to be orchestrated for startup Until the i-th type of VNFC contained in the virtual machine reaches a predetermined amount, polling is repeated to select a virtual machine that is not currently started according to the order of the largest amount of VNFC1, VNFC2, VNFC3, and VNFC4. Alternatively, polling may be performed to select virtual machines that are not currently started according to the descending order of the total amount of VNFC included in the virtual machine, or current start according to the descending order of the amount of VNFC types included in the virtual machine. Polling may be performed to select virtual machines that will not be performed. Note that in any one of the aforementioned cases, when a plurality of virtual machines satisfy the same filtering condition, one of the plurality of virtual machines may be selected in a random manner or according to an algorithm rule.

  According to the virtual machine startup method provided in this embodiment of the invention, the virtual machine to be started is appropriately selected according to the distribution relationship between the VNFC and the virtual machine and according to the rules, and correspondingly , Started according to the selection sequence. Therefore, the service performance can be improved, the VNFC resource in the currently started virtual machine can be used sufficiently in a relatively short time, and the maximum utilization efficiency can be realized under the condition that the VNFC resource is low.

  In order to facilitate a clearer understanding of the present invention, the following provides a detailed description by using actual applications. Referring to FIG. 5, FIG. 5 is a diagram illustrating a correspondence between VNF and VNFC in the session control service according to an embodiment of the present invention.

  In a simple IMS session control service, the terminal UE1 starts a voice call, and a voice signal passes through an IMS (IP Multimedia Subsystem) IP multimedia subsystem and reaches the terminal UE2. IMS includes P-CSCF (Proxy-Call Session Control Function), I-CSCF (Interrogating Call Session Control Function), and S-CSCF (Serving Call Session Control Function).

  (1) Proxy CSCF (P-CSCF)

  P-CSCF is a proxy call session control function. The P-CSCF is an entrance node that connects an IMS terminal and an IMS network. A session initiation protocol SIP (Session Initiation Protocol) signal that is started by all IMS terminals and ended by one IMS terminal needs to pass through the P-CSCF. Actually, the P-CSCF acts as a proxy server. The P-CSCF receives the SIP request or response and transfers the SIP request or response to the IMS network or IMS server. The P-CSCF may be located in the visited network or may be located in the home network.

  (2) Interrogating CSCF (I-CSCF)

  I-CSCF is an inquiry call / session control function. The I-CSCF is an ingress node, and by using this ingress node, an IMS terminal located on the visited network enters the local IMS network domain. The I-CSCF routes the SIP request or response of the overseas terminal to the domestic S-CSCF. The I-CSCF may be located in the home network or in certain cases in the visited network.

  (3) Serving CSCF (S-CSCF): S-CSCF is a serving call session control function. The S-CSCF is a central node on the SIP signal plane. The function of the S-CSCF is the same as the function of the soft switch device in the soft switch system. The S-CSCF performs a session control service for the IMS terminal and maintains the session state. SIP signals transmitted and received by all IMS terminals must pass through the S-CSCF. The S-CSCF checks the SIP signal to determine if it needs to access the application server and forwards the signal to the final destination. The S-CSCF is always located on the home network.

  The P-CSCF, I-CSCF, and S-CSCF have a DPU, SCU, and RDB necessary to complete the voice call service. A Dispatching Unit (DPU) that dispatches units is an external exit and entrance of the CSCF system and is configured to complete the dispatching of SIP messages. The session control unit (SCU) session control unit is configured to complete the service processing function of each logical network element. A remote distributed database (RDB) remote distributed storage unit is configured to complete distributed storage of user data and session data in order to perform separation between data and programs.

  In the voice call service, the corresponding VNFs are P-CSCF, I-CSCF, and S-CSCF. P-CSCF, I-CSCF, and S-CSCF have their own corresponding process / software modules. Process / software modules are DPU00, SCU01 and RDB02 in P-CSCF, DPU10, SCU11 and RDB12 in I-CSCF, and VNFC such as DPU20, SCU21 and RDB22 in S-CSCF. is there. Certainly, P-CSCF, I-CSCF, and S-CSCF have additional functions and some other VNFCs that are not needed for voice calls. Therefore, this embodiment does not give details.

  According to the solution of the present invention, the VNFC types determined according to the voice call service are DPU00, SCU01, RDB02, DPU10, SCU11, RDB12, DPU20, SCU21, and RDB22. For example, the operator needs to complete the deployment of the voice call service, and the amount of DPU00, SCU01, RDB02, DPU10, SCU11, RDB12, DPU20, SCU21, and RDB22 is 1000, 2500, 1500, 600, 1500, respectively. 900, 400, 1000, and 600 are expected. That is, the corresponding preset weights of DPU00, SCU01, RDB02, DPU10, SCU11, RDB12, DPU20, SCU21, and RDB22 are 10%, 25%, 15%, 6%, 15%, 9%, 4%, respectively. 10% and 6%.

  The virtual network function management and orchestration system learns the distribution relationship of the above 9 types of VNFC in each virtual machine by reading the attribute information of each virtual machine, that is, distributed within the started virtual machine Learn the amount of the above 9 types of VNFC and the amount of the above 9 types of VNFC distributed within the virtual machine that is not started. In addition, the first actual weights of DPU00, SCU01, RDB02, DPU10, SCU11, RDB12, DPU20, SCU21, and RDB22 in the started virtual machine are calculated.

  A VNFC, for example, SCU01 whose initial actual weight is smaller (relatively least deficient) than the preset quantity weight to the maximum excursion is selected. Thereafter, an unstarted virtual machine including SCU01 is determined according to the distribution relationship, and an unstarted virtual machine including the maximum amount of SCU01 is selected from virtual machines that are not started as new virtual machines to be started. If multiple unstarted virtual machines contain the same maximum amount of SCU01, an unstarted virtual machine containing the maximum total amount of nine types of VNFCs may be selected from multiple unstarted virtual machines, or 9 An unstarted virtual machine that includes the largest amount of types of VNFCs among the types may be selected from a plurality of unstarted virtual machines. If multiple unstarted virtual machines contain the same type of VNFC and the amount of each type of VNFC is the same, one of the multiple unstarted virtual machines may be selected in a random manner. After the selection of a new virtual machine to be started is completed, the new virtual machine is added to the second group of virtual machines. Therefore, the amount of DPU00, SCU01, RDB02, DPU10, SCU11, RDB12, DPU20, SCU21, and RDB22 in the second group of virtual machines changes and the actual weight is updated. A new virtual machine to be started is selected based on the new actual weight and according to the rules corresponding to the method described above. Each time a virtual machine is selected, a new virtual machine to be started is selected based on the new actual weight and according to the same rules, and DPU00, SCU01, RDB02, in the first group of virtual machines Polling is repeated a plurality of times until the amounts of DPU10, SCU11, RDB12, DPU20, SCU21, and RDB22 reach 1000, 2500, 1500, 600, 1500, 900, 400, 1000, and 600, respectively. If a waiting unstarted virtual machine still needs to be orchestrated for subsequent startup, the virtual machine may be selected according to rules equivalent to the method described above, or the startup sequence may be random May be determined.

  2． Multi-service scenario

  Multi-service is that the NFV network supports services with different service priorities. VNFC sets for services of different service types are different. In some scenarios, when a large number of VMs with VNFC supporting high priority services running on them and a large number of VMs with VNFC running low priority services running on them at the same time, the virtual machine is started first. The factors considered for starting up are more complex, so the present invention provides a corresponding technical solution according to the actual request policy of the serviced user. The multi-service type may mean that an A-type service with a service priority of an SLA (Service Level Agreement) availability level exists on the NFV network. The purpose of starting the virtual machine in this case is to implement a specific service of type B out of service types, where B is less than or equal to A and both A and B are positive integers It is.

  For example, starting virtual machines batchwise at a time can meet user requirements for the second service, the third service, and the fourth service. The second service needs to be jointly performed by VNFC in the set {VNFC21, VNFC22, VNFC23, VNFC24}, and the third service is jointly performed by VNFC in the set {VNFC31, VNFC32, VNFC33} And the fourth service needs to be jointly implemented by VNFCs in the set {VNFC41, VNFC42}. It is learned that multi-service requirements involve a wide variety of VNFCs and that, according to existing virtual machine startup aspects, the distribution of VNFCs is not determined, so that the startup VNFC resources cannot be fully used in the startup process. obtain. Thus, the present invention also envisages starting virtual machines in an environment with multi-service requirements. Specific solutions are as follows.

  Referring to FIG. 6, FIG. 6 is a schematic flowchart of a virtual machine startup method in a multi-service scenario according to an embodiment of the present invention. The method is applied to a network function virtualized NFV network with M network service requirements, where M is a natural number greater than one. The method 200 includes the following steps.

  Step 201: Determine N (x) types of virtual network function components VNFC corresponding to the xth service according to the xth service, where the xth service is one of the M network services. N (x) types of VNFCs can jointly implement the xth service, N is a positive integer, and the value range of x is any positive integer less than or equal to M including.

  Step 202: Obtain a distribution relationship between N (x) types of VNFCs and a virtual machine, where the distribution relationship indicates the amount of the xith type of VNFC distributed within each virtual machine. , Each virtual machine includes a virtual machine that is started or not started, the xi-th type is one of N (x) types, and the value range of xi is N (x ) Includes all of the following positive integers:

  Step 203: At least one virtual machine that is not started according to the distribution relationship obtained in 202 such that the total amount of xith type VNFCs included in the first group of virtual machines satisfies a corresponding preset amount Select a virtual machine to be started, in which case the first set of virtual machines includes a started virtual machine and at least one virtual machine to be started, and at least one start The virtual machine to be performed may be used as a virtual machine set to be started.

  Step 204: Start at least one virtual machine to be started as determined in 203.

  Optionally, in a particular implementation process, if the service does not have absolute priority within the policy, for example, the service with absolute priority is from the second service, the third service, and the fourth service. If not distinguished, no attention is paid to whether the expected performance standards of the service must be preferentially met, instead the goal is that each type of VNFC in each service reaches an estimator. is there. In this case, specifically, step 203 may be to perform “service highly efficient work” according to the distribution relationship acquired in step 202. In this case, “service highly efficient work” is specifically May include the following steps.

  Step 2031a: Actual quantity weight of xith type VNFC

Where Qxi is the total amount of VNFC of type xi in the second group of virtual machines and Qxj is the total amount of VNFC of type xj in the second group of virtual machines And the value range of xj includes all positive integers not greater than N (x), and the second group of virtual machines is started with the currently started virtual machines and the currently selected Including virtual machines.

Step 2032a: According to the actual quantity weight α _xi of the xith type VNFC and the preset quantity weight of the xith type VNFC, the actual quantity weight is a preset quantity of the kth type Find the kth type maximally smaller than the weight.

  Step 2033a: Select an unstarted virtual machine containing the kth type of VNFC as a new virtual machine to be started.

  Steps 2031a, 2032a, and 2033a are repeated until the total amount of the xith type of VNFCs included in the first group of virtual machines satisfies the corresponding preset amount. As a supplementary explanation, the second group of virtual machines in this application document is performed each time, i.e. in a new series of calculations performed each time a new virtual machine to be started is selected by orchestration. , May be further understood as the current first group of virtual machines, where the first group of virtual machines is updated in real time.

  In other specific implementation processes, if the service has absolute priority within the policy, for example, the priority of the second service is higher than the priority of the third service and the third service If the priority is higher than the priority of the fourth service, the expected service capability of the second service needs to be met first. That is, after the amount of all types of VNFC in the second service reaches the corresponding preset amount, the third service is involved and the amount of all types of VNFC in the third service corresponds After reaching the preset amount, the fourth service is involved, and the amount of all types of VNFC in the fourth service reaches the corresponding preset amount. In this case, the step 203 specifically determines the descending order of the priority of the M services, and continuously performs “service authorization work” for all the services in the M services according to the descending order of the priority. In this case, the “service authorization operation” includes the following steps.

  Step 2031b: Actual quantity weight of xith type VNFC in current service

Where Qxi is the total amount of VNFC of type xi in the second group of virtual machines and in the current service, and Qxj is in the second group of virtual machines The total amount of xjth type VNFCs in the current service, the value range of xj includes all positive integers not greater than N (x), and the second group of virtual machines Includes the currently started virtual machine and the currently selected virtual machine to be started.

Step 2032b: According to the actual quantity weight α _xi of the xi-th type VNFC in the current service and the preset quantity weight of the xi-th type VNFC in the current service, its actual quantity weight is k-th Find the kth type maximally smaller than the preset quantity weight of the type.

  Step 2033b: Select an unstarted virtual machine including the kth type of VNFC as a new virtual machine to be started.

  Steps 2031b, 2032b, and 2033b are repeated until the total amount of the xith type of VNFCs included in the first group of virtual machines in the current service satisfies the corresponding preset amount.

  In the specific implementation process of the above method, when a virtual machine containing the kth type of VNFC is selected from virtual machines that will not be started as new virtual machines to be started, the maximum amount of kth type of VNFC is A virtual machine that will not be started may be selected. If multiple unstarted virtual machines contain the maximum amount of the kth type of VNFC, the unstarted virtual machine containing the maximum amount of VNFC type is selected from multiple unstarted virtual machines or the maximum amount of VNFC A non-started virtual machine (non-category) containing is selected from a plurality of unstarted virtual machines. In addition, in the process of selecting virtual machines, when multiple unstarted virtual machines have the same VNFC activation state, for example, the type and amount of VNFC running on two unstarted virtual machines that meet one condition When the same, one of a plurality of unstarted virtual machines may be selected in a random manner.

  In the process of repeating step 2031a to step 2033a or step 2031b to step 2033b, each time step 2031a to step 2033a or step 2031b to step 2033b is performed, the relatively best unstarted virtual machine is selected and started. Added to the power sequence. After step 2031a to step 2033a or step 2031b to step 2033b are performed a plurality of times, a selection order for the selected virtual machine to be started exists in the sequence. An orchestration mechanism for starting virtual machines allows these selected virtual machines to be started to be subsequently started in order.

  Certain implementation processes require further reference to policies predetermined by the system in the particular process performing step 203. Multiple implementations may be derived from step 203 due to different preset multi-service policies. For example, in the process of selecting at least one virtual machine to be started, for each type of VNFC corresponding to multiple services, select a virtual machine that will not be started according to the order of the maximum amount of currently included types of VNFC Polling may be repeated to select a virtual machine that is not started according to the descending order of the total amount of VNFC contained in the virtual machine, or the amount of VNFC type contained in the virtual machine Polling may be repeated to select virtual machines that are not started in descending order. Note that in any one of the above cases, when a plurality of unstarted virtual machines satisfy the same filtering condition, one of the plurality of unstarted virtual machines may be selected in a random manner.

  This embodiment of the present invention may be applied at the network initialization stage, that is, at the stage when not all virtual machines have been started. In this case, the virtual machine may be started according to one service. This embodiment of the invention may be further applied at the network startup stage. For example, after a virtual machine is started in accordance with a virtual machine instruction or a host instruction in the prior art, if a plurality of virtual machines are newly added at the network startup stage, or if a plurality of virtual machines are Once recovered, newly added or suspended virtual machines need to be started according to service.

  In this embodiment of the invention, N VNFC types corresponding to the first service are determined, and of the virtual machines that need to be started, there are N types of virtual machines to be started. The virtual machine to be started is started, determined according to the distribution relationship between the VNFC and the virtual machine. In the present invention, a virtual machine on which a VNFC type corresponding to a particular service runs may be started as quickly as possible, thereby limiting the amount of virtual machine startup to avoid a virtual machine boot storm The problem caused in the middle, that is, the problem that the service cannot be provided for a long period of time or the service processing capacity cannot be improved for a long period of time can be solved, the service is provided quickly, and the service processing capacity is Improve quickly. In addition, according to the single / multiple service policy mentioned in the previous method, according to the VNFC amount status in the current virtual machine, it is possible to select the unstarted virtual machine where the most scarce type of VNFC is located, thereby It can run on the virtual machine that is started now and can fully use the VNFC in the current service at each time, improving service performance.

  In a simple example, the VNFC types corresponding to service 0 are VNFC000, VNFC001, VNFC010, and VNFC011, and the state of VNFC included in the eight virtual machines VM0 to VM7 that are not started first is shown in FIG. (VNFC not related to service 0 is not shown in this figure). In the prior art, if virtual machines are orchestrated to start according to the order of virtual machine numbers (shown in FIG. 7a), service 0 only after seven virtual machines VM0-VM6 are orchestrated Can start normally. If a selection is made according to the method in this embodiment of the invention, i.e. the virtual machine with the least deficient type of VNFC corresponding to service 0 is selected, four virtual machines, i.e. VM0 ( Or, as long as VM1), VM2 (or VM3), VM4 (or VM5), and VM6 (or VM7) are started, service 0 can be started normally. A possible embodiment is shown in FIG. 7b. Thus, in the present invention, as long as a minimal amount of unstarted virtual machines are started, service can be provided as quickly as possible, thereby reducing the time to provide the service.

  Referring to FIG. 8, FIG. 8 is a schematic structural diagram of a virtual machine starting device according to an embodiment of the present invention. The apparatus 300 includes a determination module 301, an acquisition module 302, a selection module 303, and a startup module 304. There are several possible embodiments. Therefore, the following will continue to describe the apparatus by way of example, and any technical solution that can make an equivalent replacement shall fall within the protection scope of the present invention.

  Example 1

A virtual machine starter is provided, and the device is applied to a network function virtualized NFV network with a single network service requirement (a scenario where a single service type is required is described above, and In Example 1, the first service continues to be used as an example)
A decision module 301 configured to determine N types of virtual network function components VNFC according to a first service, wherein the first service is one of many network services that can be implemented by a virtual machine. A decision module 301, which is a network service, in which N types of VNFCs can jointly implement the first service, and N is a positive integer;
An acquisition module 302 configured to acquire a distribution relationship between N types of VNFCs and virtual machines determined by a determination module 301, wherein the i th distribution relationship is distributed within each virtual machine Type of VNFC, each virtual machine includes a virtual machine that is started or not started, the i-th type is one of N types, and the value range of i is greater than N An acquisition module 302 that is all positive integers not too large,
At least one virtual machine to be started according to the distribution relationship obtained by the acquisition module 302 such that the total amount of the i-th type of VNFC included in the first group of virtual machines satisfies a corresponding preset amount A selection module 303 configured to select from unstarted virtual machines, wherein the first set of virtual machines includes a started virtual machine and at least one virtual machine to be started Module 303;
A startup module 304 configured to start at least one virtual machine to be started selected by the selection module 303;
including.

  In a particular implementation process, the acquisition module 302 acquires a distribution relationship between N types of VNFCs and virtual machines in multiple ways. In two typical aspects, reference may be made to implementation methods corresponding to Table 1 and Table 2 in the above-described embodiment.

  In a specific implementation process different from a conventional virtual machine starter, the selection module 303 in the present invention plays a very important role and selects a virtual machine that is not started according to the distribution relationship acquired by the acquisition module 302, A virtual machine that is not started may be orchestrated as a virtual machine to be started. The operation mode is as follows. That is, y unstarted virtual machines are initially added so that the total amount of i-th type VNFC running on y virtual machines and on started virtual machines reaches a corresponding predetermined amount. May be selected by using a simple theory. The selection module 303 has multiple implementations, and multiple implementations may be derived by derivation due to different service policies stored in or invoked by the selection module 303. The first service in the previous embodiment is still used as an example. For example, polling is repeated to select according to the order of VNFC1, VNFC2, VNFC3, and VNFC4. (Step 1: Select a virtual machine that will not be started including VNFC1; Step 2: Select a virtual machine that will not be started including VNFC2; Step 3: Select a virtual machine that will not be started including VNFC3; Step 4: Select VNFC4 Select the virtual machines that will not be started to include; and all the amounts of VNFC of all types included in the virtual machines that are started and the virtual machines to be started that are orchestrated for startup correspond to the first service Repeat steps 1 to 4 until a predetermined amount of type VNFC can be met). In the polling process, the amount of virtual machines that contain VNFC1 is greater than 1, so the unstarted virtual machine that contains the maximum amount of VNFC1 may be preferentially selected, and a similar aspect applies to other types of VNFCs. Applicable. In other examples, a series of selections may be made according to a descending order of the total amount of VNFC included in the virtual machine, or a series of selections may be made according to a descending order of the amount of VNFC types included in the virtual machine. Note that in any one of the above cases, when a plurality of unstarted virtual machines satisfy the same filtering condition, one of the plurality of unstarted virtual machines may be selected in a random manner. Correspondingly, the startup module 304 may continuously start selected virtual machines to be started according to the selection order of virtual machines to be started. Compared to the conventional way to start virtual machines according to sequence numbers, the selection of virtual machines according to distributed imagination is faster and more appropriate. Therefore, the goal that the total amount of the i-th type VNFC in the first service reaches the predetermined amount can be achieved more quickly, and the expected effect of the service capability of the first service can be achieved as quickly as possible.

  In the form that can be assumed, the selection module 303 is capable of making the “service assurance work”, the case corresponding to the method in the above-described embodiment, and the equivalent replacement according to the distribution relationship acquired by the acquisition module 302. The embodiment may be performed.

  In another possible form, the selection module 303 may select the “service high efficiency work” in the above-described embodiment according to the distribution relationship acquired by the acquisition module 302, that is, the method and method described in step 1031 to step 1033. All embodiments that can make equivalent cases and equivalent replacements may be made.

  In a possible form, reference may be made to FIG. 9 regarding the form of the selection module 303. FIG. 9 is a schematic diagram of the selection module 303 in the virtual machine starting device according to the embodiment of the present invention. The device is applied to a network function virtualized NFV network with a single network service requirement. Specific embodiments are as follows. That is, the selection module 303 includes a calculation unit 3031a, a determination unit 3032a, and a selection unit 3033a.

  The calculation unit 3031a is the actual quantity weight of the i-th type VNFC

In this case, Qi is the total amount of i-th type VNFC in the second group of virtual machines, and Qj is the j-th type in the second group of virtual machines The value range of j includes all positive integers not greater than N, and the second group of virtual machines is the currently started virtual machine and the currently selected started Virtual machines to be included.

Determining unit 3032a in accordance with Ryoomomi alpha _i acquired by the i-th type of VNFC preset amount weights and i-th type of an actual amount weight alpha _i and computing units in VNFC 3031, the Configured to find the kth type where the actual quantity weight is maximally smaller than the kth type of preset quantity weight, where the kth type is one of the N types It is.

  The selection unit 3033a is configured to select an unstarted virtual machine including the kth type of VNFC determined by the determination unit 3032a as a new virtual machine to be started, in this case, the maximum amount of the kth An unstarted virtual machine that includes a type of VNFC may be selected, and if multiple unstarted virtual machines include the maximum amount of the kth type of VNFC, the selection unit 3033a may select N from multiple unstarted virtual machines. One of the types may select a virtual machine that does not start that includes the maximum amount of VNFC types, or may select a virtual machine that does not start that includes the maximum total amount of VNFC in the first service, and The VNFC in one service is a different type of VNFC among the N VNFC types.

  The selection module 303 continues the calculation unit 3031a, the determination unit 3032a, and the selection unit 3033a until the total amount of the i-th type VNFC included in the first group of virtual machines satisfies the corresponding preset amount. It may further include a control unit 3034a configured to control to work in order in a regular and periodic manner.

  In the process of controlling the calculation unit 3031a, the determination unit 3032a, and the selection unit 3033a by the control unit 3034 to work periodically in order, the selection order of at least one virtual machine to be started selected by the selection unit 3033a Exists. The selection module 303 may further include a sequence recording unit 3035a, in which case the sequence recording unit 3035a is configured to record the sequence of at least one virtual machine to be started selected by the selection unit 3033a. Correspondingly, the start module 304 starts at least one virtual machine to be started according to the order recorded by the order recording unit 3035a.

  In a particular implementation process, multiple implementations may be obtained by derivation due to different service policies stored in or invoked by the selection module 303. For example, in the process of selecting at least one virtual machine to be started (not limited to the whole process, including part of the process), the virtual machine that was started and the start to be orchestrated for startup Until the i-th type of VNFC contained in the virtual machine reaches a predetermined amount, polling is repeated to select a virtual machine that is not currently started according to the order of the largest amount of VNFC1, VNFC2, VNFC3, and VNFC4. Alternatively, polling may be performed to select virtual machines that are not currently started according to the descending order of the total amount of VNFC included in the virtual machine, or current start according to the descending order of the amount of VNFC types included in the virtual machine. Polling may be performed to select virtual machines that will not be performed. Note that in any one of the aforementioned cases, when a plurality of virtual machines satisfy the same filtering condition, one of the plurality of virtual machines may be selected in a random manner or according to an algorithm rule.

  Example 2

A virtual machine starter is provided, and the device is applied to a network function virtualized NFV network with M network service requirements (a multi-service scenario is described above, and in Example 2, M Network service continues to be used as an example), M is a natural number greater than 1, and the device 300 is
A decision module 301 configured to determine N (x) types of virtual network functional component VNFC corresponding to the xth service according to the xth service, wherein the xth service is M networks A network service within the service, where N (x) types of VNFCs can jointly implement the xth service, N is a positive integer, and the value range of x is not greater than M A decision module 301 comprising a positive integer of
An acquisition module 302 configured to acquire a distribution relationship between N (x) types of VNFCs determined by a determination module 301 and a virtual machine, wherein the distribution relationship is distributed within each virtual machine. Xi-th type of VNFC, including virtual machines where each virtual machine is started or not started, and xi-th type out of N (x) types corresponding to x-th service An acquisition module 302 that is any positive integer whose value range is not greater than N (x),
According to the preset multi-service policy and the distribution relationship acquired by the acquisition module 302 so that the total amount of the xith type VNFC included in the first group of virtual machines satisfies the corresponding preset amount A selection module 303 configured to select at least one virtual machine to be started from unstarted virtual machines, wherein the first set of virtual machines is started with at least one started virtual machine A selection module 303 including a virtual machine to be
A startup module 304 configured to start at least one virtual machine to be started selected by the selection module 303;
including.

  In a particular implementation process, the acquisition module 302 acquires a distribution relationship between N (x) types of VNFCs and virtual machines in a number of ways. For details, reference can be made to the description of the acquisition module 302 in Example 1, and therefore details are not described here again.

  In a possible form, if the service does not have absolute priority in the policy, for example, if the service with absolute priority is not distinguished from the second service, the third service, and the fourth service. Attention is not paid to whether the expected performance standards of the service must be met preferentially, instead the goal is that each type of VNFC in each service reaches an estimator. In this case, the selection module 303 may perform the method described in steps 2031a to 2033a, the case corresponding to the method, and all embodiments that can make equivalent replacements. In a possible form, FIG. 9 may be referred to for an implementation of the selection module 303. FIG. 9 is a schematic structural diagram of the selection module 303 in another virtual machine starting device according to an embodiment of the present invention. In this case, it is not necessary to distinguish the priorities of the M network services. Specific embodiments are as follows. That is, the selection module 303 includes a calculation unit 3031a, a determination unit 3032a, and a selection unit 3033a.

  The calculation unit 3031a is the actual quantity weight of the xith type VNFC

Where Qxi is the total amount of VNFC of the xith type in the second group of virtual machines and Qxj is the xjth type in the second group of virtual machines Xjth type is one of N (x) types corresponding to the xth service, and the value range of j is not greater than N (x) And the second group of virtual machines includes the currently started virtual machine and the currently selected virtual machine to be started.

Decision unit 3032a is, xi-th type of VNFC and amount weight that has been pre-set for the actual amount weight alpha _xi is calculated by the calculation unit 3031a and an actual amount weight alpha _xi of xi-th type of VNFC Is configured to find the kth type whose actual weight is maximally smaller than the preset weight of the kth type, where the kth type is N (x) One of the types.

  The selection unit 3033a is configured to select an unstarted virtual machine including the kth type of VNFC determined by the determination unit 3032a as a new virtual machine to be started, in this case, the maximum amount of the kth An unstarted virtual machine that includes a type of VNFC may be selected, and if multiple unstarted virtual machines contain the maximum amount of the kth type of VNFC, the selection unit 3033a may select from a plurality of unstarted virtual machines. A virtual machine that is not started including a large amount of VNFC types may be selected, or a virtual machine that is not started including the maximum total amount of VNFC may be selected.

  The selection module 303 continues the calculation unit 3031a, the determination unit 3032a, and the selection unit 3033a until the total amount of the xith type VNFC included in the first group of virtual machines satisfies the corresponding preset amount. It may further include a control unit 3034a configured to control to work in order in a regular and periodic manner.

  In the process of controlling the calculation unit 3031a, the determination unit 3032a, and the selection unit 3033a from the control unit 3034a so as to work in order continuously and periodically, at least one virtual to be started selected by the selection unit 3033a There is a machine selection order. The apparatus further includes a sequence recording unit 3036a configured to record the selection sequence for the virtual machine to be started selected by the selection unit 3033a. After obtaining the selection order recorded by the order recording unit 3036a, the startup module 304 starts the virtual machine to be started selected by the selection unit 3033a according to the selection order.

  In other possible forms, if the service has absolute priority within the policy, for example, the priority of the second service is higher than the priority of the third service and the third service If the priority is higher than the priority of the fourth service, the expected performance standard of the second service needs to be reached first. That is, after each type of VNFC corresponding to the second service reaches the expected amount, the third service is considered by using the same idea, and then the fourth service is considered. In this case, the selection module 303 may perform the methods described in steps 2031b to 2033b, examples corresponding to the methods, and all embodiments that can make equivalent replacements. In a possible form, reference may be made to FIG. 10 for an implementation of the selection module 303. FIG. 10 is a schematic diagram of the selection module 303 in another virtual machine starting device according to an embodiment of the present invention. In this case, the priority of M network services is distinguished. Specific embodiments are as follows.

  The selection module 303 includes a calculation unit 3031b, a determination unit 3032b, and a selection unit 3033b, and may further include a control unit 3034b and a traversal unit 3035b. Functions performed by the calculation unit 3031b, the determination unit 3032b, the selection unit 3033b, and the control unit 3034b in FIG. 10 correspond to the calculation unit 3031a, the determination unit 3032a, the selection unit 3033a, and the control unit 3034a in FIG. The functions performed are the same, the principle for transferring signals between units is similar, the same operating mode "calculation unit-decision unit-selection unit" is adapted for operation, and the control unit The calculation unit, the determination unit, and the selection unit are controlled to repeat each operation according to the operation mode. The difference is that the embodiment shown in FIG. 10 further includes a traversal unit 3035b. Traversal unit 3035b is configured to determine a descending order of priority for M services, so that control unit 3034b is in the first group of virtual machines for each service according to service priority. Until the amount of each type of VNFC in each service reaches a preset amount, the calculation unit 3031b, the determination unit 3032b, and the selection unit 3033b are operated according to the operation mode of “calculation unit-determination unit-selection unit”. Control to work continuously and periodically.

  The traversal unit 3035b causes the control unit 3034b to operate the calculation unit 3031b, the determination unit 3032b, and the selection unit 3033b for each service continuously and periodically according to the operation mode of “calculation unit-decision unit-selection unit”. In a process that can be controlled in this way, there is a selection order for the virtual machine to be started that is selected by the selection unit 3033b. The order recording unit 3036b is configured to record the selection order for the virtual machine to be started that is selected by the selection unit 3033b. After obtaining the selection order, the startup module 304 starts the virtual machine to be started that is selected by the selection unit 3033b.

  In a particular implementation process, multiple implementations may be obtained by derivation due to different service policies stored in or invoked by the selection module 303. For example, in the process of selecting at least one virtual machine to be started (not limited to the whole process, including part of the process), the virtual machine that was started and the start to be orchestrated for startup Polling is repeated to select virtual machines that are not currently started according to the order of the maximum amount of VNFC1, VNFC2, VNFC3, and VNFC4 until the xi-th type of VNFC included in the virtual machine reaches a predetermined amount Alternatively, polling may be performed to select virtual machines that are not currently started according to the descending order of the total amount of VNFC included in the virtual machine, or current start according to the descending order of the amount of VNFC types included in the virtual machine. Polling may be performed to select virtual machines that will not be performed. Note that in any one of the aforementioned cases, when a plurality of virtual machines satisfy the same filtering condition, one of the plurality of virtual machines may be selected in a random manner or according to an algorithm rule.

  In this embodiment of the invention, the virtual machine starter includes a decision module, an acquisition module, a selection module, and a start module. The determination module is configured to determine N (x) VNFC types corresponding to the xth service. The selection module is started from the virtual machine that needs to be started according to the single / multiple service policy and the distributed relationship between the N (x) types of VNFCs acquired by the acquisition module and the virtual machine. Decide which virtual machine should be used. The start module starts the virtual machine to be started. In the present invention, according to the VNFC required by the service, the virtual machine on which the VNFC type corresponding to the particular service runs may be found and started as quickly as possible, thereby reducing the virtual machine boot storm. To solve the problem caused while limiting the startup amount of the virtual machine to avoid, i.e., the service cannot be provided for a long period of time or the service processing capacity cannot be improved for a long period of time, Services are provided quickly and service processing capabilities are improved quickly. In addition, embodiments of the selection module described above can select the unstarted virtual machine in which the most scarce type of VNFC is located, according to the VNFC amount status in the current virtual machine, thereby allowing the current virtual machine In addition to the above, the VNFC usage efficiency in the current service can be fully realized in each time, and the service performance is improved.

  Referring to FIG. 11, FIG. 11 is a schematic structural diagram of a virtual machine start computer according to an embodiment of the present invention. The device 500 includes a processor 501, a memory 502, and a bus 503.

  The processor 501 generates a corresponding operation control signal, sends the operation control signal to the corresponding portion of the computer processing device, and reads and processes the data with software. In particular, all the functional modules in the computer processing device Is configured to read and process data and programs into the memory 502 so as to perform corresponding functions, thereby controlling the corresponding portions and performing operations according to instruction requirements.

  The memory 502 is configured to store mainly software units, such as operating systems, applications, and functional instructions, or subsets thereof, or extensions thereof, that store programs and various data. The operating system includes various system programs, such as the windows series, that are used to perform various basic services and handle hardware-based tasks. Memory 502 may further include non-volatile random access memory (NVRAM), and includes software and applications for managing hardware and software and data resources in computer processing devices to support control. You may give to the processor 501.

  All of the aforementioned hardware units may be connected for communication by using bus 503.

  In this way, the processor 501 determines N types of virtual network function components VNFC according to the first service by calling a program or instruction stored in the memory 502; N types of VNFCs and virtual machines Where the distribution relationship indicates the amount of the i th type of VNFC distributed within each virtual machine, where the i th type is one of the N types Yes, including all positive integers whose value range i is not greater than N; so that the total amount of the i-th type VNFC in the first group of virtual machines satisfies the corresponding preset amount Selecting at least one virtual machine to be started from the unstarted virtual machines according to the pre-configured service policy and distribution relationship in the memory 502, in this case a set of a first group of virtual machines Includes a starting virtual machine and at least one should be started virtual machine; starting the virtual machine to be at least one start.

  Similarly, the processor 501 may perform the methods and related cases described in steps 101 to 104 and steps 201 to 204 by calling a program or instruction stored in the memory 502.

  In addition, in a particular implementation process, the processor 501 calls a program or instruction stored in the memory 502 to invoke an envisaged implementation method, case corresponding to the method in all scenarios of the method embodiments described above, and Any method capable of making equivalent replacements may be performed.

  As will be appreciated by those skilled in the art, all or part of the method steps described above may be implemented by a program that instructs the associated hardware. The program may be stored in a computer readable storage medium. From the above description of the embodiments, those skilled in the art can clearly understand that the present invention can be implemented by hardware, firmware, or a combination thereof.

  The above descriptions are merely some embodiments of the technical solutions of the present invention, and are not intended to limit the protection scope of the present invention. Any modification, equivalent replacement, or improvement made without departing from the principle of the present invention shall fall within the protection scope of the present invention. That is, any change, equivalent replacement, or improvement shall satisfy that the corresponding virtual machine is started according to the service and fall within the protection scope of the present invention.

000 NFV system
002 Orchestrator
004 Virtual network function manager
006 Virtualization Infrastructure Manager
008 Virtual network function
010 Virtual computing
012 Computing hardware
014 storage hardware
016 Network hardware
018 Virtual storage
020 Virtual network
022 Element management system
024 Operation Support System / Business Support System
026 Service / Virtual Network Function / Infrastructure Description
028 Virtual network function management and orchestration system
030 Virtual network function infrastructure
300 devices
301 decision module
302 Acquisition module
303 selection module
304 start-up module
500 devices
501 processor
502 memory
503 bus
3031a calculation unit
3031b calculation unit
3032a decision unit
3032b decision unit
3033a selection unit
3033b Selection unit
3034a Control unit
3034b Control unit
3035b traversal unit
3036a Sequence recording unit
3036b Sequence recording unit

Claims (33)

A virtual machine starting method, wherein the method is applied to a network function virtualized NFV network where a first network service is deployed, and the NFV network is implemented by a plurality of virtual machines running on a computing device. The first network service is jointly implemented by N types of virtual network function components VNFC, where N is a positive integer;
Obtaining a distribution relationship between the N types of VNFCs and the plurality of virtual machines, wherein the distribution relationship is within each of the plurality of virtual machines. Step, indicating the amount of each
The virtual to be started at least one unstarted virtual machine according to the distribution relationship such that the total amount of each type of VNFC in the first group of virtual machines satisfies a preset amount corresponding to each type of VNFC Selecting from the plurality of virtual machines as a machine set, the set of the first group of virtual machines comprising a started virtual machine set and a virtual machine set to be started, The amount of started virtual machines in the set of virtual machines is greater than or equal to 0, and when the amount is greater than 0, the started virtual machines belong to the plurality of virtual machines; and
Starting one virtual machine in the set of virtual machines to be started;
A method comprising: Selecting at least one unstarted virtual machine from the plurality of virtual machines as a set of virtual machines to be started according to the distribution relationship,
step 1:
Determining the current quantity weight of each type of VNFC in the first group of virtual machines according to: α _i is the current quantity weight of the i-th type of VNFC, and Qi is the virtual quantity The current total amount of the i-th type VNFC in the first group of machines, Qj is the current total amount of the j-th type VNFC in the first group of virtual machines, and the i-th type VNFC and the j-th type VNFC belong to N types of VNFCs and comprise positive integers in which the value range of i and j is not greater than N; and
Step 2: According to the determined current quantity weight of each type of VNFC and the preset quantity weight of each type of VNFC, the current quantity weight is maximally smaller than the preset quantity weight finding a k-th type of VNFC, wherein the k-th type of VNFC is one of the N types of VNFCs; and
Step 3: selecting an unstarted virtual machine comprising the kth type of VNFC from the plurality of virtual machines to be added to the set of virtual machines to be started;
Step 1, Step 2 until the total amount of each type of VNFC that is in the N types and in the first group of virtual machines satisfies a preset amount corresponding to each type of VNFC And repeating step 3 in order,
The method of claim 1 comprising: Selecting the unstarted virtual machine comprising the kth type of VNFC from the plurality of virtual machines to be added to the virtual machine set to be started comprises the following aspects:
Selecting an unstarted virtual machine with the maximum amount of the k-th type of VNFC to be added to the set of virtual machines to be started, or
If multiple unstarted virtual machines have the maximum amount of the k-th type of VNFC, the unstarted virtual machines with the maximum amount of VNFC types among the N types are set as the virtual machine set to be started. Select to be added, or select an unstarted virtual machine with the maximum total amount of VNFC used to implement the first network service to be added to the set of virtual machines to be started, The method of claim 2 comprising any one of the steps. In the process of repeating step 1, step 2 and step 3 in order, further comprising obtaining a selection order of the at least one unstarted virtual machine;
The step of starting one virtual machine in the set of virtual machines to be started comprises
Starting the at least one unstarted virtual machine according to the selection order;
The method according to claim 2 or 3. Selecting at least one unstarted virtual machine from the plurality of virtual machines as a set of virtual machines to be started according to the distribution relationship,
Determining K types according to the distribution relationship, wherein the K types are a subset of the N types, and the K types of VNFCs do not exist in the started virtual machine; Steps,
Selecting L unstarted virtual machines from the plurality of virtual machines to be added to the set of virtual machines to be started according to a first preset rule, wherein the L unstarted virtual machines A machine comprising the K types of VNFCs, and L is not greater than K according to the first preset rule; and
Selecting H non-started virtual machines from the plurality of virtual machines to be added to the set of virtual machines to be started according to a second preset rule, the second preset The total amount of each type of VNFC in the first group of virtual machines satisfies a preset amount corresponding to each type of VNFC,
The method of claim 1 comprising: Obtaining a first selection order of the L unstarted virtual machines in the process of selecting the L unstarted virtual machines from the plurality of virtual machines according to the first preset rule;
Obtaining a second selection order of the H unstarted virtual machines in the process of selecting the H unstarted virtual machines from the plurality of virtual machines according to the second preset rule;
Further comprising
The step of starting one virtual machine in the set of virtual machines to be started comprises
First starting the L unstarted virtual machines according to the first selection order, and then starting the H unstarted virtual machines according to the second selection order;
6. The method according to claim 5. A virtual machine starting method, the method being applied to a network function virtualized NFV network in which M network services are deployed, the NFV network being implemented by a plurality of virtual machines running on a computing device The x th service is one network service in the M network services, and the x th service is jointly implemented by N (x) types of VNFCs, where N is a positive integer , With all positive integers whose value range of x is not greater than M, where M is a natural number greater than 1, the method comprises:
Obtaining a distribution relationship between the N (x) types of VNFCs and the plurality of virtual machines, wherein the distribution relationship is distributed in each of the plurality of virtual machines. The xi th type is one of the N (x) types corresponding to the x th service, and the value range of xi is N (x) A step comprising all positive integers not too large, and
At least one according to the distribution relationship, such that the total amount of each type of VNFC in the first group of virtual machines and in the M network services satisfies a preset amount corresponding to each type of VNFC. Selecting one unstarted virtual machine from the plurality of virtual machines as a set of virtual machines to be started, wherein the first set of virtual machines is started with the started virtual machine set And when the amount of the started virtual machines in the started virtual machine set is greater than or equal to 0 and the amount is greater than 0, the started virtual machine is the plurality of virtual machines. A step belonging to a virtual machine,
Starting one virtual machine in the set of virtual machines to be started;
A method comprising: Selecting at least one unstarted virtual machine from the plurality of virtual machines as a set of virtual machines to be started according to the distribution relationship,
step 1:
Determining the current quantity weight of the xith type VNFC in the first group of the virtual machine according to: α _xi is the current quantity weight of the xith type VNFC; Qxi is the current total amount of the xith type VNFC in the first group of virtual machines, and Qxj is the current total amount of the xjth type VNFC in the first group of virtual machines. The xj-th type is one of the N (x) types corresponding to the x-th service, and all positive values whose value range of xj is not greater than N (x) A step comprising an integer of
Step 2: according to the determined current quantity weight α _xi of the xith type VNFC and the preset quantity weight of the xith type VNFC, the current quantity weight is preset Finding a kth type of VNFC that is maximally smaller than a quantity weight, wherein the kth type of VNFC is one of the N (x) types of VNFCs; and
Step 3: selecting an unstarted virtual machine comprising the kth type of VNFC from the plurality of virtual machines to be added to the set of virtual machines to be started;
Until the total amount of each type of VNFC provided in the first group of virtual machines in the M network services satisfies the preset amount corresponding to the type of VNFC 1, step 2 and step 3 are repeated in order,
The method of claim 7 comprising: Selecting at least one unstarted virtual machine from the plurality of virtual machines as a set of virtual machines to be started according to the distribution relationship,
Determining a descending order of priority of M services, and continuously performing service authorization operations for all services in the M services according to the descending order of priority;
The service authorization work is
step 1:
Determining the current quantity weight of each type of VNFC in the first group of virtual machines according to: α _xi is the current quantity weight of the xith type of VNFC, and Qxi is the The current total amount of the xi-th type of VNFC in the first group of virtual machines, Qxj is the current total amount of the xj-th type of VNFC in the first group of virtual machines; the value range of xj comprises all positive integers not greater than N (x) corresponding to the current service; and
Step 2: its current quantity weight according to the determined current quantity weight α _xi of each type of VNFC in the current service and a preset quantity weight of each type of VNFC in the current service Finding a kth type of VNFC that is maximally smaller than a pre-set quantity weight, where the kth type is one of the N (x) types in the current service Is a step,
Step 3: selecting an unstarted virtual machine comprising the kth type of VNFC from the plurality of virtual machines to be added to the set of virtual machines to be started;
Until the total amount of each type of VNFC in the first group of virtual machines in the current service meets a preset amount corresponding to each type of VNFC in the current service, step 1, Repeating step 2 and step 3 in order;
The method of claim 7 comprising: Selecting the unstarted virtual machine comprising the kth type of VNFC from the plurality of virtual machines to be added to the virtual machine set to be started comprises the following aspects:
Selecting an unstarted virtual machine with the maximum amount of the k-th type of VNFC to be added to the set of virtual machines to be started, or
If multiple unstarted virtual machines have the maximum amount of the kth type of VNFC, select unstarted virtual machines with the maximum amount of VNFC types to be added to the set of virtual machines to be started 10. A method according to claim 8 or 9, comprising any one of the following steps: selecting an unstarted virtual machine with the maximum amount of VNFC to be added to the set of virtual machines to be started. Obtaining a selection order of the at least one selected unstarted virtual machine in the process of repeating step 1, step 2 and step 3;
The step of starting one virtual machine in the set of virtual machines to be started comprises
Starting the at least one selected unstarted virtual machine according to the selection order;
The method according to claim 8. Obtaining a selection order of the at least one selected unstarted virtual machine in a process of continuously performing service authorization operations for all services in the M services in descending order of priority;
The step of starting one virtual machine in the set of virtual machines to be started comprises
Starting the at least one selected unstarted virtual machine according to the selection order;
The method of claim 9. Selecting the at least one unstarted virtual machine from the plurality of virtual machines as a set of virtual machines to be started;
Determining K types according to the distribution relationship, wherein the K types are a subset of all VNFC types corresponding to the M services, and the K types of VNFCs are activated. Steps that do not exist in the virtual machine
Selecting L unstarted virtual machines from the plurality of virtual machines to be added to the set of virtual machines to be started according to a first preset rule, wherein the L unstarted virtual machines A machine comprising the K types of VNFCs, and L is not greater than K according to the first preset rule; and
Selecting H non-started virtual machines from the plurality of virtual machines to be added to the set of virtual machines to be started according to a second preset rule, the second preset The total amount of each type of VNFC provided in the first group of virtual machines in the first group of virtual machines satisfies a preset amount corresponding to each type of VNFC, Steps,
The method of claim 7 comprising: Obtaining a first selection order of the L virtual machines to be started in the process of selecting the L non-started virtual machines from the plurality of virtual machines according to the first preset rule. When,
Obtaining a second selection order of the H virtual machines to be started in the process of selecting the H unstarted virtual machines from the plurality of virtual machines according to the second preset rule. When,
Further comprising
The step of starting one virtual machine in the set of virtual machines to be started comprises
First starting the L unstarted virtual machines according to the first selection order, and then starting the H unstarted virtual machines according to the second selection order;
The method according to claim 13. Selecting at least one unstarted virtual machine from the plurality of virtual machines as a set of virtual machines to be started according to the distribution relationship,
Determining the descending order of priority of the M services, continuously performing service guarantee work for all the services in the M services according to the descending order of priority, and H1 unstarted virtual machines Selecting from the plurality of virtual machines to be added to the set of virtual machines to be started;
After the step of performing service assurance work for all services in the M services, H2 unstarted virtual machines are added to the set of virtual machines to be started according to a fourth preset rule. Selecting each of the plurality of virtual machines in the M network services provided in the first group of virtual machines according to the fourth preset rule. The total amount of VNFC satisfies the preset amount corresponding to each type of VNFC; and
With
The service assurance work is
Determining K (x) types according to the distribution relation, wherein the K (x) types are a subset of the N (x) types in a current service, and the K (x) No VNFC types exist in the started virtual machine, and
Selecting L (x) unstarted virtual machines from the plurality of virtual machines to be added to the set of virtual machines to be started according to a third preset rule, the L (x A) the unstarted virtual machine comprises the K (x) types of VNFCs, and L (x) is not greater than K (x) according to the third preset rule;
The method of claim 7 comprising: Obtaining a third selection order of the H1 unstarted virtual machines in a process of performing service assurance work for all services in the M services;
Obtaining a fourth selection order of the H2 unstarted virtual machines in the process of selecting H2 unstarted virtual machines from the plurality of virtual machines according to a fourth preset rule;
Further comprising
The step of starting one virtual machine in the set of virtual machines to be started comprises
First starting the H1 unstarted virtual machines according to the third selection order, and then starting the H2 unstarted virtual machines according to the fourth selection order;
The method of claim 15. A virtual machine starter, wherein the device is applied to a network function virtualized NFV network in which a first network service is deployed, and the NFV network is implemented by a plurality of virtual machines running on a computing device , The first network service is jointly implemented by N types of virtual network function components VNFC, where N is a positive integer,
An acquisition module configured to acquire a distribution relationship between the N types of VNFCs and the plurality of virtual machines, wherein the distribution relationship is in each of the plurality of virtual machines. An acquisition module indicating the amount of each of the N types of VNFCs;
At least one unstarted virtual according to the distribution relationship acquired by the acquisition module such that the total amount of each type of VNFC in the first group of virtual machines meets a preset amount corresponding to each type of VNFC A selection module configured to select from the plurality of virtual machines as a virtual machine set to be started, wherein the first set of virtual machines is started with the started virtual machine set. And when the amount of the started virtual machines in the started virtual machine set is equal to or greater than 0 and the amount is greater than 0, the started virtual machines are the plurality of virtual machines. A selection module belonging to the virtual machine
A startup module configured to start one virtual machine in the set of virtual machines to be started selected by the selection module;
A device comprising: The selection module is specifically the pre-configured, wherein the total amount of each type of VNFC in the N types and in the first group of virtual machines corresponds to each type of VNFC. It is configured to repeat the following Step 1, Step 2, and Step 3 in order until the required amount is satisfied,
step 1:
Determine the current quantity weight of each type of VNFC in the first group of virtual machines according to: α _i is the current quantity weight of the i th type of VNFC, and Qi is the first quantity weight of the virtual machine. Is the current total amount of the i-th type VNFC in the group of 1, Qj is the current total amount of the j-th type VNFC in the first group of the virtual machine, and the i-th type VNFC and the j-th type VNFC belong to N types of VNFCs, and i and j have a positive integer whose value range is not greater than N,
Step 2: According to the determined current quantity weight of each type of VNFC and the preset quantity weight of each type of VNFC, the current quantity weight is maximally smaller than the preset quantity weight find the kth type of VNFC, and the kth type of VNFC is one of the N types of VNFCs;
Step 3: Select an unstarted virtual machine comprising the kth type of VNFC from the plurality of virtual machines to be added to the set of virtual machines to be started.
The apparatus of claim 17. In step 3, the selection module specifically includes the following aspects:
Selecting an unstarted virtual machine with the maximum amount of the k th type of VNFC to be added to the set of virtual machines to be started, or
When a plurality of the unstarted virtual machines have the maximum amount of the kth type of VNFC, a virtual machine set to be started is designated as a virtual machine to be started having the maximum amount of VNFC types among the N types. Or select an unstarted virtual machine with the maximum total amount of VNFC used to implement the first network service to be added to the set of virtual machines to be started ,
The apparatus of claim 18, configured to perform any one of the following: The starting module is
In the process of repeating step 1, step 2 and step 3 in order by the selection module, obtaining a selection order of the at least one unstarted virtual machine and obtaining the at least one unstarted virtual machine according to the selection order Start,
20. An apparatus according to claim 18 or 19, further configured as follows. Specifically, the selection module includes:
Configured to determine K types according to the distribution relationship, wherein the K types are subsets of the N types, and the K types of VNFCs are not present in the started virtual machine. ,
The L unstarted virtual machines are configured to select from the plurality of virtual machines to be added to the set of virtual machines to be started according to a first preset rule. The machine comprises the K types of VNFCs, and L is not greater than K according to the first preset rule,
Configured to select from the plurality of virtual machines to be added to the set of virtual machines to be started in accordance with a second preset rule, the second preset The total amount of each type of VNFC in the first group of virtual machines satisfies a preset amount corresponding to each type of VNFC,
The apparatus of claim 17. Specifically, the starting module is
Obtaining a first selection order of the L unstarted virtual machines in the process of selecting the L unstarted virtual machines from the plurality of virtual machines according to the first preset rule by the selection module; Configured to
Obtaining a second selection order of the H unstarted virtual machines in the process of selecting the H unstarted virtual machines from the plurality of virtual machines according to the second preset rule by the selection module; Configured to
Configured to first start the L unstarted virtual machines according to the first selection order and then start the H unstarted virtual machines according to the second selection order;
The apparatus according to claim 21. A virtual machine starter, wherein the device is applied to a network function virtualized NFV network in which M network services are deployed, and the NFV network is implemented by a plurality of virtual machines running on a computing device The x th service is one network service in the M network services, and the x th service is jointly implemented by N (x) types of VNFCs, where N is a positive integer , Comprising all positive integers whose value range of x is not greater than M, wherein M is a natural number greater than 1,
An acquisition module configured to acquire a distribution relationship between the N (x) types of VNFCs and the plurality of virtual machines, wherein the distribution relationship is a medium of each of the plurality of virtual machines. Xi-th type of VNFC distributed in the xi-th type is one of the N (x) types corresponding to the x-th service, and the value range of i An acquisition module with all positive integers not greater than N (x);
Acquired by the acquisition module such that the total amount of each type of VNFC in the first group of virtual machines and in the M network services satisfies a preset amount corresponding to each type of VNFC. A selection module configured to select from the plurality of virtual machines as a set of virtual machines to be started, the at least one virtual machine not to be started according to the distribution relationship comprising: The set comprises a started virtual machine set and a virtual machine set to be started, and the amount of started virtual machines in the started virtual machine set is greater than or equal to zero, and the amount is greater than zero A selection module, wherein the started virtual machine belongs to the plurality of virtual machines;
A startup module configured to start one virtual machine in the set of virtual machines to be started selected by the selection module;
A device comprising: Specifically, the selection module is pre-configured so that the total amount of each type of VNFC in the M network services provided in the first group of virtual machines corresponds to each type of VNFC It is configured to repeat the following Step 1, Step 2, and Step 3 in order until the given amount is satisfied,
step 1:
Determining the current quantity weight of the xi-th type of VNFC in the first group of the virtual machine according to: α _xi is the current quantity weight of the xi-th type of VNFC, and Qxi is the Is the current total amount of the xi-th type of VNFC in the first group of virtual machines, and Qxj is the current total amount of the xj-th type of VNFC in the first group of virtual machines All positive integers in which the xj type is one of the N (x) types corresponding to the xth service and the value range of j is not greater than N (x) With
Step 2: according to the determined current quantity weight α _xi of the xi-th type of VNFC and a preset quantity weight of the xi-th type of VNFC, the quantity of which the current quantity weight is preset Find the kth type of VNFC that is maximally smaller than the weight, and the kth type of VNFC is one of the N (x) types of VNFCs;
Step 3: Select an unstarted virtual machine comprising the kth type of VNFC from the plurality of virtual machines to be added to the set of virtual machines to be started.
24. The apparatus of claim 23. Specifically, the selection module includes:
Determining a descending order of priority of M services, and configured to continuously perform service authorization work for all services in the M services according to the descending order of priority;
The service authorization work is
step 1:
Determining the current quantity weight of each type of VNFC in the first group of virtual machines according to: α _xi is the current quantity weight of the xith type of VNFC and Qxi is a virtual quantity Is the current total amount of the xi-th type VNFC in the first group of machines, and Qxj is the current total amount of the xj-th type VNFC in the first group of virtual machines, and xj A value range of comprises all positive integers not greater than N (x) corresponding to the current service; and
Step 2: its current quantity weight according to the determined current quantity weight α _xi of each type of VNFC in the current service and a preset quantity weight of each type of VNFC in the current service Finding a kth type of VNFC that is maximally smaller than a pre-set quantity weight, where the kth type is one of the N (x) types in the current service Is a step,
Step 3: selecting an unstarted virtual machine comprising the kth type of VNFC from the plurality of virtual machines to be added to the set of virtual machines to be started;
Until the total amount of each type of VNFC in the first group of virtual machines in the current service meets a preset amount corresponding to each type of VNFC in the current service, step 1, Repeating step 2 and step 3 in order;
24. The apparatus of claim 23, comprising: In step 3, the selection module specifically includes the following aspects:
Selecting an unstarted virtual machine with the maximum amount of the k th type of VNFC to be added to the set of virtual machines to be started, or
If multiple unstarted virtual machines have the maximum amount of the kth type of VNFC, select unstarted virtual machines with the maximum amount of VNFC types to be added to the set of virtual machines to be started Or select an unstarted virtual machine with the maximum amount of VNFC to be added to the virtual machine set to be started,
26. An apparatus according to claim 24 or 25, configured to perform any one of: Specifically, the starting module is
Obtaining a selection order of the at least one selected unstarted virtual machine in the process of repeating step 1, step 2 and step 3;
Starting the at least one selected unstarted virtual machine according to the selection order;
25. The apparatus of claim 24, configured as follows. Specifically, the starting module is
Obtaining a selection order of the at least one selected unstarted virtual machine in a process in which the service authorization operation is continuously performed by the selection module for all services in the M services in descending order of priority; ,
Starting the at least one selected unstarted virtual machine according to the selection order;
26. The apparatus of claim 25, configured as follows. Specifically, the selection module includes:
Configured to determine K types according to the distribution relationship, wherein the K types are a subset of all VNFC types corresponding to the M services, and the K types of VNFCs are started. Does not exist in the virtual machine
The L unstarted virtual machines are configured to select from the plurality of virtual machines to be added to the set of virtual machines to be started according to a first preset rule. The machine comprises K types of VNFC, L is not greater than K according to the first preset rule,
Configured to select from the plurality of virtual machines to be added to the set of virtual machines to be started in accordance with a second preset rule, the second preset The total amount of each type of VNFC provided in the first group of virtual machines and in the M network services satisfies the preset amount corresponding to each type of VNFC ,
24. The apparatus of claim 23. Specifically, the starting module is
A first selection order of the L virtual machines to be started in the process of selecting the L unstarted virtual machines from the plurality of virtual machines according to the first preset rule by the selection module; Is configured to get
A second selection order of the H virtual machines to be started in the process of selecting the H unstarted virtual machines from the plurality of virtual machines according to the second preset rule by the selection module; Is configured to get
Configured to first start the L unstarted virtual machines according to the first selection order and then start the H unstarted virtual machines according to the second selection order;
30. The device of claim 29. Specifically, the selection module includes:
Determining a descending order of priority of the M services, continuously performing a service guarantee operation on all the services in the M services according to the descending order of priority, and adding H1 unstarted virtual machines to the Configured to select from the plurality of virtual machines to be added to a set of virtual machines to be started;
After performing the service assurance work on all the services in the M services, H2 unstarted virtual machines can be added to the virtual machine set to be started according to a fourth preset rule. Each type of VNFC configured to select from the plurality of virtual machines and provided in the first group of virtual machines and in the M network services according to the fourth preset rule The total amount of meets the preset amount corresponding to each type of VNFC,
The service assurance work is
Determining K (x) types according to the distribution relation, wherein the K (x) types are a subset of the N (x) types in a current service, and the K (x) No VNFC types exist in the started virtual machine, and
Selecting L (x) unstarted virtual machines from the plurality of virtual machines to be added to the set of virtual machines to be started according to a third preset rule, the L (x A) the unstarted virtual machine comprises the K (x) types of VNFCs, and L (x) is not greater than K (x) according to the third preset rule;
24. The apparatus of claim 23, comprising: Specifically, the starting module is
Configured to obtain a third selection order of the H1 unstarted virtual machines in the process of performing the service assurance work for all services in the M services by the selection module;
Obtaining a fourth selection order of the H2 unstarted virtual machines in the process of selecting the H2 unstarted virtual machines from the plurality of virtual machines according to the fourth preset rule by the selection module; Configured to
Configured to first start the H1 unstarted virtual machines according to the third selection order and then start the H2 unstarted virtual machines according to the fourth selection order;
32. The apparatus of claim 31.   A virtual machine start device, the device comprising at least one processor, a memory, and a bus, the processor and the memory being connected for communication by using the bus, the memory being 17. A virtual machine startup device that stores programs and instructions, wherein the processor is configured to recall the programs and instructions to the memory and to implement the method of any one of claims 1-16.